Drones doi: 10.3390/drones10080639

Authors: Wei Xia Peng Chen Feifei Song Kai Li Tong Zhang Kun Zhu

Unmanned Aerial Vehicles (UAVs) have experienced rapid development due to their advantages of low cost, high efficiency, flexibility, reliability, and strong environmental adaptability. To fully leverage the potential of UAV swarms in multi-task scenarios, optimizing task allocation has become a crucial direction to enhance the efficiency of UAV swarms. While existing task allocation methods have achieved promising results, frequent inter-UAV information exchange can impose substantial communication overhead in distributed UAV swarms, particularly in resource-constrained applications such as emergency rescue and mountainous operations. In this context, to reduce inter-UAV interactions while maintaining the solution quality of task allocation, this paper first analyzes the factors affecting communication interactions between UAVs. Based on this analysis, we utilize historical bidding information to infer other UAVs’ positions and employ estimation strategies to resolve task conflicts, thereby reducing communication iterations. Furthermore, we propose a bidding-based grouping method to eliminate ineffective communication interactions. Finally, we introduce a network simplification algorithm based on reducing the number of triangular network topologies to optimize the communication network structure. Simulation results demonstrate that the proposed algorithm significantly reduces inter-UAV interactions while preserving the number of allocated tasks, with a maximum observed increase of only approximately 8% in task waiting time across the evaluated simulation settings.

Drones doi: 10.3390/drones10080638

Authors: Zishuang Pan Dazhao Yu Wei Han Xichao Su Jie Wang Shansong Song Bing Wan

Automatic carrier landing of fixed-wing UAVs remains challenging under deck motion, carrier airwake, gusts, and actuator faults. This paper proposes a robust flexible predefined-time prescribed performance control (RFPTPPC) framework with beneficial disturbance utilization (BDU). A control-oriented six-degree-of-freedom cascaded model with direct lift control is first established. To address the temporary infeasibility of fixed predefined-time PPC boundaries, direction-selective flexible boundaries restore the admissible attitude error region when the nominal envelope is threatened, while a smoothly coordinated recovery branch drives the error inward. An adaptive super-twisting extended state observer (ASTESO) reconstructs lumped disturbances in the cascaded loops. Using the ASTESO outputs, BDU evaluates each disturbance component, retains those that favor error convergence, and compensates for adverse components, thereby improving attitude tracking accuracy and maintaining PPC feasibility. Lyapunov analysis establishes practical predefined-time stability of the closed-loop system. Comparative simulations demonstrate improved trajectory tracking, attitude regulation, actuator coordination, and robustness under randomized landing conditions.

Drones doi: 10.3390/drones10080637

Authors: Aykut Sirma Angelos Plastropoulos Gilbert Tang Argyrios Zolotas

Search-and-rescue (SAR) missions following earthquakes and other disasters require aerial video perception systems that do more than detect objects in isolated frames. Operators must maintain the identities of access points, vehicles, responders, hazards, and other mission-relevant targets despite UAV ego-motion, dust, debris, occlusion, scale variation, and abrupt scene transitions. This paper presents E’CHIT (Edge-Oriented Colour Histogram Instance-Guided Tracking), a deployment-oriented, operator-centric UAV tracking framework for real-world disaster-response applications. Its primary scientific contribution is an identity-stabilised, detector-assisted tracking methodology. YOLOv8-seg proposals trained on D’RespNeT initialise and refresh tracks; a Custom-RE3 recurrent module propagates target states through short detector dropouts; and a lightweight EOMC verifier, based on edge orientation, mean colour, and shape consistency, determines whether tracks should be accepted, refreshed, or reacquired. A scene-cut watchdog that combines luminance mean absolute difference (MAD) with HSV histogram divergence prevents stale identities from carrying over after hard edits or sudden feed changes. Custom-RE3 is the continuation module implemented and evaluated in this study. The surrounding E’CHIT wrapper follows an initialise–reseed–verify–reset cycle and is tracker-adaptable at the software-interface level: another compatible SOT or MOT continuation module can be integrated through adapter modifications, state and bounding-box conversion, and method-specific retuning, followed by independent validation. All reported quantitative results therefore apply to the Custom-RE3 implementation. D’RespNeT, the optional reinforcement learning (RL) warm start, the HUD, and the deployment stack support this central tracking contribution. D’RespNeT provides 28 polygon-annotated SAR classes. An author-developed PPO/SAC script is used only during offline detector training. In the reported runs, it produces different early optimisation trajectories for selected difficult or under-represented classes, while the default supervised schedule remains the strongest final global mAP reference. No RL policy runs during deployment; the detector architecture, parameter count, and inference graph remain unchanged. Evaluation on D’RespNeT and authentic disaster-response UAV footage shows that E’CHIT increases Success@IoU ≥ 0.5 from 0.62 to 0.79, reduces identity switches by approximately 71%, and maintains real-time 1080p performance, achieving 164–330 FPS for single-target tracking and 24–100+ FPS for end-to-end multi-target operation on an RTX-class GPU using FP16. The VOT2014, NT-VOT211, and VOTS2024 figures reproduce historical result spaces reported in the literature and include a clearly labelled, non-official E’CHIT operating-point marker solely for context. This marker was not produced using the corresponding official datasets, toolkits, reset rules, or submission routes; it is excluded from the primary quantitative claims and must not be interpreted as a leaderboard rank or a protocol-identical comparison. Overall, the system demonstrates how identity-stable UAV tracks can provide actionable operator cues for target monitoring, entry-point assessment, and UAV–UGV/ground-team coordination in cluttered disaster scenes.

Drones doi: 10.3390/drones10080636

Authors: Lei Zuo Ying Wang Jialu Liu Yu Lu Ruiwen Gu

Unauthorized low-altitude UAVs can challenge conventional fixed-parameter interception algorithms through agile maneuvers. This study develops a three-dimensional one-on-one terminal-interception simulation environment that incorporates protected-zone penetration, a within-step geometric interception criterion, and kinematic constraints. The intruder, denoted as the red UAV, combines six physically interpretable maneuver templates to generate structured evasive penetration behavior. The defender, denoted as the blue UAV, augments blended generalized proportional navigation (B-GPN) with a bounded residual corrective acceleration produced by deep reinforcement learning, thereby forming a hybrid architecture that combines a geometry-based nominal guidance command with reward-driven bounded compensation. In standardized tests on 1000 unseen scenarios, the implemented residual PPO pipeline increased the interception rate from 63.8% for nominal guidance to 94.1% (95% Wilson interval: 92.46–95.40%) and maintained at least 88.0% interception under the tested control-delay, kinematic, and noise perturbations. It also achieved the highest interception rate among the evaluated residual-learning implementations under both the stable-configuration comparison and the auxiliary task-side-controlled check; this result is limited to the reported implementations and is not a general ranking of algorithm families. These findings indicate that bounded residual learning can compensate for structural limitations of conventional guidance under the evaluated conditions.

Drones doi: 10.3390/drones10080635

Authors: Watchara Ruangsang Patiyuth Pramkeaw

Wildfires need a fast response, and a small unmanned aerial vehicle (UAV) carrying its own detector is an appealing way to find them early. Almost every published UAV fire detector is validated on a single corpus, which leaves open the question that matters to an operator: how much of the reported accuracy survives a change in scene. We trained unmodified YOLOv8n, YOLOv8s, and YOLOv8m on two public benchmarks that ship official test splits, D-Fire and the UAV subset of the Flame and Smoke Detection Dataset, under one fixed recipe with three seeds on D-Fire, then evaluated every checkpoint on both test splits with frozen weights. In domain, YOLOv8m reached 79.04 ± 0.21% mAP@0.5 on D-Fire, within 0.04 points of the published value for the same architecture on the same split, and 92.71% on FASDD_UAV. Moved across corpora, the same weights fell 38 and 65 points below a locally trained model, with fire degrading about twice as far as smoke. A perceptual-hash audit then found near-duplicates of training images in 46.0% and 90.7% of the two test sets; re-evaluating on the uncontaminated remainder costs 2.8 and 12.4 points, closes two-thirds of the apparent difficulty gap between the corpora, and leaves corrected transfer shortfalls near 26 and 62 points. Onboard, an F450 carrying a Pixhawk 2.4.8 and a Jetson Nano A02 flew and detected a controlled fire, but full-resolution inference ran 4.0 to 41.4 times slower than the video it consumed. Both gaps must close before a platform of this class can support operational wildfire monitoring.

Drones doi: 10.3390/drones10080634

Authors: Richard Avella Camila A. González Paula N. López

Atmospheric icing is one of the most critical meteorological hazards for unmanned aerial vehicles (UAV), whose operation under adverse conditions—high latitudes, elevated altitudes, long-endurance missions without pilot intervention—particularly exposes them to ice accumulation on aerodynamic surfaces and propellers. Unlike manned aviation, where this phenomenon has been extensively studied and regulated, a significant knowledge gap exists in the UAV domain that limits the development of effective protection systems adapted to energy constraints. This article provides an integrative review—conducted with a systematic search strategy following PRISMA reporting guidelines—of atmospheric ice formation mechanisms, their specific effects on UAV propellers, and the two most promising mitigation approaches: electrothermal modelling for the optimisation of electric heating systems and the development of functional surface materials including superhydrophobic coatings (SHC); composites with conductive nanofillers (graphene, carbon nanotubes); and piezoelectric actuators. The analysis demonstrates that hybrid systems combining passive and active strategies managed by intelligent control represent the most viable solution for extending UAV operational envelopes under known icing conditions, with a projected reduction in anti-icing system energy consumption of at least 40% relative to conventional continuous heating. This estimate is based on the most conservative published evidence: pulsed electrothermal de-icing achieves 40–60% savings versus continuous anti-icingSHC-assisted hybrid heating reduces IPS power by more than 80% on static aerofoils; and rotary-wing pulsed systems reduce mean consumption by 60–75% relative to continuous operation. Key research gaps are identified, and a prioritised future research agenda is proposed to support the development of certifiable anti-icing systems for rotary-wing UAV platforms.

Drones doi: 10.3390/drones10080633

Authors: Wenhui Ma Han Zhang Shuangxi Liu

In view of the task allocation problem under communication constraints, this paper proposes a distributed performance impact (DPI) algorithm for heterogeneous multi-UAV cooperation. First, the task allocation problem is modelled, incorporating task requirements, heterogeneous flight performance and non-ideal networked communication. Then, a multi-phase DPI algorithm is designed in five stages: task inclusion, cluster generation, conflict resolution, coalition formation and task segmentation. Based on task inclusion, a communication-reachable networked cluster generation method is developed. The conflict resolution mechanism is improved based on the removal-performance-impact and task contribution rate. Further considering independently unachievable tasks, a dual-fallback mechanism is constructed by a pruning-redundancy coalition formation scheme and a task- segmentation scheme, thereby enhancing both system operational efficiency and task-completion certainty. Finally, simulations are conducted to demonstrate the effectiveness, superiority and robustness of the proposed DPI algorithm.

Drones doi: 10.3390/drones10080632

Authors: Xiaojia Yan Zhangsong Shi Shiyan Sun Huihui Xu Huimin Zhu Qingping Hu Weiming Zhu Yinglei Li

Cross-view geo-localization (CVGL) is a critical technology used in unmanned aerial vehicles (UAVs) and widely applied in navigation and target localization tasks. However, owing to the extreme perspective disparity between UAV oblique views and satellite vertical views, CVGL still involves significant challenges, including geometric distortion caused by viewpoint differences, drastic appearance inconsistencies, and the difficulty in bridging semantic gaps between heterogeneous data. To address these issues, we propose a novel CVGL method named dynamic-feature collaborative optimization and semantic-alignment network (DFSA), designed to extract robust feature representations and achieve fine-grained alignment. Specifically, the DFSA employs a residual-based vision transformer as the backbone to capture global context while alleviating the training instability and feature collapse often associated with standard transformers. To bridge the semantic gap between global and local features, we design a feature optimization module comprising a local feature enhancer and a global feature aggregator. This module establishes a closed-loop collaborative system that facilitates top-down semantic guidance and bottom-up detail feedback. Furthermore, we introduce a semantic segmentation and alignment module that adaptively partitions images into semantic regions based on feature response distributions, shifting the matching granularity from the global level to the semantic region level to effectively overcome feature mismatches caused by positional offsets and scale variations. Extensive experiments conducted on the University-1652 and SUES-200 datasets demonstrate the superior image retrieval performance of the proposed DFSA. Specifically, DFSA achieves a Recall@1 of 94.87% and an Average Precision (AP) of 95.32% on the University-1652 dataset and maintains highly competitive Recall@1 performances between 96.83% and 99.25% across various altitudes on the SUES-200 dataset. These results validate the model’s effectiveness in handling extreme viewpoint changes for UAV-based cross-view image retrieval tasks.

Drones doi: 10.3390/drones10080631

Authors: Yichi Zhang Jianxiang Xi Wei Li Le Wang Nanchi Liu

For malicious distributed unmanned aerial vehicle (UAV) swarms operating in civilian airspace, this paper proposes a novel dual-channel induced countermeasure strategy (DCICS) with prescribed countermeasure flight trajectories, in which the designable countermeasure signals are injected into the sensor–controller (SC) channel and the controller–actuator (CA) channel of partial or all UAVs, simultaneously. Firstly, we establish a dual-channel induced countermeasure protocol, in which the sensor-induced countermeasure signal is modeled as a continuous function of swarm dynamics, whereas the actuator-induced countermeasure signal is generated through a countermeasure signal generation exosystem (CSGES). Then, we derive a closed-form expression of the countermeasure flight trajectory, explicitly characterizing the effect of the countermeasure signals on the swarm motion. Furthermore, a necessary and sufficient condition for the feasibility of the DCICS is derived. By introducing adjustable performance factors, we construct a robust H∞ framework and provide the design criteria for the dual-channel induced countermeasure signals accordingly. The proposed strategy can drive the formation of the distributed UAV swarm along a prescribed countermeasure flight trajectory while maintaining the original formation structure. Finally, we present numerical simulation examples to validate the effectiveness of the theoretical results.

Drones doi: 10.3390/drones10080630

Authors: Wenbiao Gan Chenxi Guan Junjie Zhuang Jingwei Ma Xiaozhang Liu Shaojiang Dong Zihan Song Jiangtao Zhang Guoqi Zeng

Owing to its inherent flight safety, low takeoff and landing requirements, and favorable economic efficiency, the autogyro UAV, especially its electric and hybrid-electric variants, has become a core platform for low-altitude aviation missions such as transportation, inspection, and surveillance in plateau and offshore regions. However, the low air density and low Reynolds number conditions encountered in plateau regions can induce aerodynamic issues such as premature laminar flow separation, dynamic stall, and increased induced drag, which directly reduce payload capacity and endurance of small electric autogyro UAVs. In offshore environments, strong winds, turbulence, and gust disturbances intensify rotor–wake interactions, cause abrupt variations in aerodynamic loads, and reduce control margins, which severely restricts the mission reliability and flight safety of low-altitude unmanned platforms. These environmental effects collectively degrade rotor performance, including reduced aerodynamic efficiency and insufficient lift generation, and further amplify the energy constraint of electric/hybrid-electric propulsion systems. In response to bottlenecks that restrict the practical application of autogyro UAVs in extreme environments, this paper systematically reviews research progress on lift vector control for autogyro UAV rotors operating under such conditions. First, the typical aerodynamic problems encountered by autogyro UAVs in plateau and offshore environments are summarized, and their underlying physical mechanisms are analyzed from both system-level and local-flow perspectives, with a focus on how environmental factors affect the autorotation stability of unmanned platforms. Subsequently, the development of passive lift vector control technologies is reviewed, with an emphasis on the aerodynamic benefits of passive pitch mechanisms, vortex generators, and blade-tip winglets, as well as their engineering feasibility for small autogyro UAV blades. Active lift vector control technologies are then examined, including air-jet flow control, synthetic jets, and trailing-edge flaps, with discussions of their potential to delay flow separation and stall, enhance rotor aerodynamic efficiency, and an assessment of their adaptability to the energy and structural constraints of unmanned platforms. Finally, a lift vector control strategy suitable for autorotating rotors of autogyro UAVs is proposed, based on careful consideration of energy consumption, structural constraints, and control effectiveness. It provides a reference for aerodynamic optimization and flight control research on electric and hybrid-electric autogyro UAVs operating in extremely low-altitude environments.

Drones doi: 10.3390/drones10080629

Authors: Ravil I. Mukhamediev Valentin Smurygin Liudmila Gorodetskaya Yan Kuchin Adilet Dauletuly Nursultan Kuldeyev Adilkhan Symagulov Irina Fedorovich

This study investigates the automatic classification of crop phenological stages from low-altitude UAV RGB imagery. The dataset included 11,489 images of five crops: sunflower, rapeseed, soybean, wheat, and barley. The images were annotated using the Biologische Bundesanstalt, Bundessortenamt und Chemische Industrie (BBCH) scale, with labels corresponding to either single stages or stage ranges to reflect heterogeneous field conditions and transitional crop states. A pretrained ResNet18 model was adapted to the task using transfer learning. Training was conducted in two stages: first, the classification head was optimized while the backbone remained frozen; second, the entire network was fine-tuned. The model achieved strong internal test accuracy across all crops, with 100% test accuracy for rapeseed and barley, more than 99% for the remaining crops, and a mean accuracy of 99.73% under the studied survey conditions. The results also compare favorably with previously reported studies on UAV-based phenological classification. Overall, the findings support the potential of low-altitude UAV imagery and deep learning for localized phenological assessment of selected field zones in precision agriculture, while broader deployment requires validation across independent fields, seasons, regions, and survey conditions.

Drones doi: 10.3390/drones10080628

Authors: Tianyu Yang Mingyang Zhai Qisheng Wen Yifei Ma Shaoshuai Zhi Shuangfeng Wei

Single-UAV dense visual SLAM is often limited by long trajectory accumulation, incomplete local observations, redundant map growth, and onboard computation constraints. Collaborative mapping can distribute a large mission across several local submaps, but dense 3D Gaussian Splatting (3DGS) maps are expensive to exchange and individual Gaussian primitives are not reliable cross-agent matching units. This paper proposes G3M-SLAM, an anchor-guided Gaussian memory framework for UAV-oriented dense SLAM with representation-level submap fusion. Stable geometric anchors organize local Gaussian primitives and form a compact structural interface for submap exchange, overlap recognition, and correction. A hybrid feature-render tracking strategy combines sparse geometric constraints with Gaussian rendering residuals. A generative completion module predicts candidate Gaussians in weakly observed regions, while multi-view geometric verification and an evidence-aware lifecycle mechanism reject unsupported candidates and control redundant map growth. A dual-graph loop bundle adjustment couples the camera pose graph and the anchor memory graph so that corrections can be propagated to anchor-associated Gaussian structures. Experiments on Replica, ScanNet, TUM RGB-D, and EuRoC MAV evaluate local tracking, dense rendering, runtime, and a split-agent fusion protocol. In the latter protocol, fusion reduces ATE from 0.045 m to 0.031 m, translational RPE from 0.030 m to 0.017 m, and rotational RPE from 1.56° to 0.93°. The serialized anchor packet is 0.66 MB, approximately 101.2× smaller than the complete Gaussian submap. On Jetson AGX Orin, the full system processes EuRoC V101 and V103 at 1.49 FPS and 1.42 FPS, respectively, while ATE increases by only 0.001 m relative to the RTX 3090 Ti workstation results. These gains represent recovery from split-agent degradation and compact representation exchange rather than an improvement over full-sequence single-agent processing. The present study therefore evaluates a representation-level fusion interface and does not reproduce the full conditions of a field-deployed decentralized multi-UAV system.

Drones doi: 10.3390/drones10080627

Authors: Chaofeng Wang Shengming Dai Jie Luo Longfei Zhang

UAV fleet configuration for truck–drone collaborative delivery requires deployment-oriented drone specification choice under wind. In this study, a UAV configuration is defined as a specific hardware parameter bundle comprising cruise speed, battery-block count, and associated mass and energy specifications. Nominal cruise speed or battery-block count alone do not predict synchronized system performance. Nine UAV configurations are compared under a common experimental protocol and the same heuristic on 60 mFSTSP instances (10, 25, and 50 customers) and nine wind scenarios, using paired makespan gaps relative to a baseline of six battery blocks and a 20 m/s cruise speed (Configuration 5). Wind affects drone outcomes through flight time and through expansion or contraction of the wind-feasible sortie set, which reshapes customer allocation and truck–drone synchronization. UAV performance is governed by a time–energy balance rather than any single nominal attribute, and no universally best drone configuration emerges. Balanced high-support UAV configurations are most robust at medium and large customer scales, though rankings remain scenario-dependent. Validation using real-world operational data under three wind conditions confirms the simulation findings, with model predictions achieving acceptable accuracy (3–6% deviation for makespan and energy consumption) and demonstrating wind-aware configuration superiority and practical applicability. The evidence supports predeployment drone fleet screening by expected wind exposure and service scale for effective drone deployment planning.

Drones doi: 10.3390/drones10080626

Authors: Jinting Ye Jie Lang Kefei Liao Ningbo Xie Jiansheng Huang Ranjun Yang Xiansui Wei Ming Li

The growing frequency of unauthorized UAV activities has increased the demand for real-time perception and rapid response on resource-constrained edge devices. This study proposes an edge-deployable UAV detection and net-capture system based on Net-Capture Detection YOLO (NCDet-YOLO). Developed from YOLOv8n, NCDet-YOLO incorporates C2f_Faster, SPD_Conv, EMA, and a lightweight three-scale detection head, with CrossKD used to compensate for accuracy loss caused by structural compression. The dataset contains 6615 images and was divided into 5292 training and 1323 validation images. The self-collected data include DJI Phantom 4 and DJI Inspire 2 UAVs observed at approximately 4–30 m under different daytime backgrounds. NCDet-YOLO achieves an mAP50–95 of 0.6504 with 1.55 M parameters and 4.1 GFLOPs. On a Jetson Orin NX Super under the 15 W power mode, it achieves 31.53 FPS, representing a 31.67% increase over YOLOv8n. The detector is further integrated with target alignment, distance determination, trigger control, and net-capture execution. In 10 real-platform trials, 8 captures were successful, corresponding to an 80.0% success rate, with one false-trigger event and an end-to-end latency from target detection to net-capture firing of approximately 400 ms.

Drones doi: 10.3390/drones10080625

Authors: Aliesha Hvala Anindilyakwa Rangers Hamish A. Campbell

Abandoned, lost, or otherwise discarded fishing gear (ALDFG) is a persistent form of marine pollution requiring survey approaches capable of resolving individual items across large spatial extents. While uncrewed aerial vehicles (UAVs) can capture imagery at resolutions sufficient to resolve individual debris items, their use remains largely constrained to visual line-of-sight (VLOS) operations, limiting large-scale coastal monitoring. This case study develops and field-tests an operational framework for low-altitude beyond visual line-of-sight (BVLOS) UAV surveys, in which DEM-based communication viewshed modelling incorporating first Fresnel zone clearance is used to plan BVLOS missions. A lightweight fixed-wing UAV flown at 60 m AGL completed 20 missions across 210 km of remote northern Australian coastline. Communication viewshed modelling reliably guided mission planning with 90.5% of waypoints placed within predicted high-clearance zones maintaining moderate-to-strong command-and-control (C2) link quality in flight. Manual screening confirmed that the resulting imagery was of sufficient quality, with 291 derelict fishing nets detected. In a simulated VLOS operational scenario, 76.3% of these detections fell beyond VLOS range, and equivalent coverage would require an estimated 8.8-fold increase in mission count. These findings demonstrate that fixed-wing UAVs operating under low-altitude BVLOS conditions can support large-scale image acquisition in remote coastal areas, particularly when enabled by communication-aware mission planning.

Drones doi: 10.3390/drones10080624

Authors: Mingxi Chen Cheng Guo Shaojie Ma Bingting Zha Zhen Zheng

Detecting small unmanned aerial vehicles (UAVs) in ground-to-air imagery is challenging because their weak visual cues must be preserved without imposing excessive computation on resource-constrained platforms. To address the unresolved trade-off between tiny-target representation and deployment efficiency, we propose a Lightweight Feature-Fusion and Small-Target Enhancement Network (LFE-YOLO), a lightweight detector that coordinates partial-channel feature extraction, efficient cross-scale fusion, high-resolution prediction, background-interference suppression, and stable tiny-box regression within a unified architecture. Specifically, C2fFaster and GSConv reduce redundant computation while maintaining multi-scale feature propagation; a P2 high-resolution detection branch and Efficient Multi-scale Attention preserve fine spatial cues and suppress background interference; and Normalized Wasserstein Distance complements Complete Intersection over Union to improve tiny-box localization stability. We also construct Det-UAV by integrating newly collected multi-platform and multi-scene UAV imagery with existing data using scene- and sequence-independent partitioning and duplicate control. Experiments on Det-UAV show that LFE-YOLO improves detection accuracy while reducing parameters and computation relative to YOLOv8s. Zero-shot evaluation on the public DUT Anti-UAV dataset further indicates favorable transferability to an unseen data distribution. TensorRT 8.2.1 deployment experiments on NVIDIA Jetson TX2 show that, under the same evaluation settings, LFE-YOLO achieves higher detection accuracy and inference throughput, lower latency, and a smaller engine size than the comparable-scale YOLOv8s and YOLO11s models. These results support a practical accuracy–efficiency balance for small-UAV detection under constrained resources.

Drones doi: 10.3390/drones10080623

Authors: Zihao Wang Ying Ma Ziming Liu Hailong Yan Qiaoyu Zhang Meng Zhang

We propose a high-quality and computationally efficient trajectory replanning method for Unmanned Aerial Vehicle (UAV) swarms, termed RHCR-Opt, which is designed to continuously and efficiently generate multiple collision-free trajectories in dense obstacle environments. RHCR-Opt consists of three layers. The first two layers are the rolling-horizon collision resolution (RHCR) algorithm based on the Conflict-Based Search (CBS), while the third layer focuses on trajectory generation and optimization using Minimum Control (MINCO) trajectories. Within the two-layer RHCR framework, the improved Lifelong Planning A* (LPA*) algorithm incorporating spatiotemporal constraints is proposed and employed as the low-level solver of CBS to satisfy the frequent search requirements for feasible paths under varying spatiotemporal constraints, thereby significantly improving computational efficiency. Furthermore, the rolling-horizon collision resolution concept is adopted in the high-level CBS framework, where only the discovery of collision-free paths within a finite time window is considered. This substantially reduces the computational burden associated with trajectory generation and optimization beyond the time window. At the third layer, a MINCO-based trajectory generation scheme is designed, and a swarm trajectory joint optimization framework with a finite time window is proposed to generate dynamically feasible and collision-free trajectories. In addition, for swarm missions requiring simultaneous arrival, a two-stage temporal coordination optimization method is developed. Extensive simulation experiments demonstrate that, compared with state-of-the-art (SOTA) algorithms on the proposed benchmark, RHCR-Opt achieves significant improvements in both trajectory quality and computational efficiency. In particular, when the swarm size becomes large, the computational efficiency is improved by at least 23.7%.

Drones doi: 10.3390/drones10080622

Authors: Yuhao Gong Jinfu Lin Jiaqiang Zhang Han Wang

Fixed-wing UAVs operating in long-range missions often fly through airspace subject to heterogeneous multi-source constraints that vertically compress the flyable space into a constraint corridor of continuously varying thickness. Conventional path planning methods incur high online computational costs in such scenarios. Deep reinforcement learning can generate reactive decisions from local observations, yet existing approaches predominantly target multirotor obstacle avoidance and rely on observations designed for discrete obstacles, lacking a unified representation for corridor constraints. Moreover, constraint conditions vary across mission scenarios, demanding cross-scenario policy generalization. This paper proposes the Egocentric Constraint Corridor (ECC), which fuses multi-source constraints into upper and lower boundary surfaces defining the corridor, then egocentrically encodes the surrounding corridor relative to the vehicle into a margin field serving as structured policy input. A deep reinforcement learning framework built on ECC is trained end-to-end, with its multi-branch network and composite reward function following from the structure of the corridor encoding. Experiments show that ECC-DRL achieves path efficiency approaching that of globally informed A*, and that it is the only one of the compared methods that computes its decisions online within the decision interval. Ablation studies confirm the margin field is necessary for reliable navigation, and the ECC encoding enables zero-shot transfer to scenarios with unseen terrains and radar deployments without retraining. Hardware-in-the-loop experiments on an embedded platform verify real-time closed-loop feasibility.

Drones doi: 10.3390/drones10080621

Authors: Heungseob Kim Seunghyeon Yu Byoungho Choi

Underwater vehicles must operate efficiently within the limits of their onboard resources, and travel time is a primary operational objective for extending submerged endurance and range. Ocean currents significantly affect vehicle motion and travel time, either aiding or impeding propulsion depending on their direction and magnitude. This study proposes a cell-based, ocean-current-aware cost formulation for offline three-dimensional (3D) underwater path planning. A 3D grid map integrating ocean current vectors and underwater terrain is constructed, and rather than modifying a specific search algorithm, the proposed approach defines a travel-time-based cost at the cell-transition level by incorporating current effects into the vehicle’s effective velocity. Because the cost operates at the cell-transition level, it can be adopted by standard grid-search planners such as Dijkstra’s, A*, and D* without modification. Simulation experiments over the Tsushima–Jeju corridor using two measured current fields from the Korea Hydrographic and Oceanographic Agency show that identical fields aided westbound transits (−2.4% and −4.4% travel time versus a current-unaware baseline) while opposing eastbound transits (+1.3% and +1.8%), with the planner exploiting favorable flows and detouring to mitigate adverse flows; these effects amplified roughly threefold as the vehicle speed decreased from 15 to 6 knots. A* and Dijkstra’s algorithms returned identical optimal costs under the same formulation, confirming planner independence. The results demonstrate that embedding measured current information at the map level yields realistic, condition-dependent, minimum-travel-time routes for offline mission planning.

Drones doi: 10.3390/drones10080620

Authors: Seokwon Yeom

Drones, or unmanned aerial vehicles (UAVs), are widely utilized in a variety of fields [...]

Drones doi: 10.3390/drones10080619

Authors: Xin He Yizhan Ju Yaqing Chen Lingxiao Xue Yumei Zhang

With the increasing scale and density of low-altitude unmanned aerial vehicle (UAV) operations, safety separation between multirotor UAVs has become a critical parameter for low-altitude airspace management. Existing studies mainly consider aircraft geometry, navigation errors, trajectory deviations, and conventional collision risk models, while rotor-downwash-induced aerodynamic interference remains insufficiently addressed. This study proposes a safety separation assessment method for quadrotor UAVs by integrating computational fluid dynamics (CFD) with an improved Event collision model. A small-scale quadrotor UAV is analyzed, and its rotor downwash flow fields under vertical- and horizontal-motion conditions are simulated using the multiple reference frame method. Based on a 5 m/s crosswind-resistance capability threshold, aerodynamic-interference characteristic distances are extracted and used to construct a basic collision box. To better represent the actual aerodynamic hazard region, an I-shaped improved collision box is further developed and incorporated into the Event collision model. Under a target level of safety, the longitudinal, lateral, and vertical minimum safety separations are determined as 1.68 m, 1.72 m, and 1.08 m, respectively. The results show that the proposed CFD–Event coupled method can transform rotor downwash characteristics into collision risk parameters and provide a quantitative basis for safety separation assessment in dense low-altitude multirotor UAV operations.

Drones doi: 10.3390/drones10080618

Authors: Mingxuan Li Liang Xu Shuo Wang Yu Han Huayong Xu Juyong Zhang

Urban unmanned aerial vehicle (UAV) navigation may require path planning that accounts for geometric obstacles and spatially varying communication-related risk. This paper presents a transformer-guided interference-aware 3D path-planning method for urban voxel environments. A 3D convolutional neural network (CNN)–transformer network predicts a dense route probability field from occupancy, electromagnetic risk, start–goal, and auxiliary planning channels. The field is restored to the raw-map resolution and used only as a search prior for A* on the original occupancy and risk maps. Obstacle avoidance, endpoint correctness, 6-connected motion (each move reaches one of six face-adjacent voxels, with no diagonal motion), and final path cost evaluation are enforced by graph search rather than by the neural model. On 320 synthetic urban cases covering four map sizes and four building density settings, Guided A* achieves a 27.7× speedup over A* and an 11.9× speedup over Weighted A*, while reducing expanded nodes by 91.2% relative to A*. The mean path cost and electromagnetic cost increase by 2.7% and 5.7%, respectively. Compared with the rapidly exploring random tree (RRT), the method reduces path cost by 10.1% and electromagnetic exposure by 12.0% at similar runtime. A post-training sensitivity study further identifies an empirical balance between route-prior guidance, electromagnetic risk avoidance, route length, and search effort, while the Manhattan weight exhibits the expected heuristic inflation efficiency–quality trade-off. An extended model trained on a larger mixture of procedural and Sionna RT ray-traced data, including real OpenStreetMap building geometry, is further evaluated without retraining on two real-geometry benchmarks, UrbanRadio3D and an OpenStreetMap–Sionna RT suite, where Guided A* retains a 100% success rate and reduces expanded nodes by 97–99% relative to A* while increasing mean path cost by at most 1.7%.

Drones doi: 10.3390/drones10080617

Authors: Yuhua Cong Xian Zhu Zhisheng Wang Yujia Li

Hierarchical multi-UAV planning can produce scheduled paths that become unsafe during execution because tracking dynamics, actuator limits, sampling, and communication are not fully represented upstream. We introduce a certificate-guided safe-tracking framework that treats each planned path as a versioned execution contract with explicit tube, separation, timing, uncertainty, communication, and input bounds. A rigid-body-derived translational interface supports a command-producing control Lyapunov function–control barrier function quadratic program with hard safety constraints. When a candidate becomes infeasible, a separate minimum-slack program localizes the conflict without passing its command to the plant, and an explicit repair map converts the resulting witness into timing or vertical-spacing updates, with escalation when local repair fails. Randomized comparisons show that certificate feedback removes the observed tube and separation failures of a plain CBF-QP while maintaining reliable completion and competitive tracking relative to a tracking-error-bound comparator. Disturbance and sensor-noise sweeps characterize robustness, and separate indoor flights confirm single-reference trackability. The framework therefore turns execution infeasibility into actionable planning feedback rather than a terminal controller failure.

Drones doi: 10.3390/drones10080616

Authors: Qi Li Tad Gonsalves

The Traveling Salesman Problem (TSP) and the Vehicle Routing Problem (VRP) are two classical combinatorial optimization problems. In recent years, their drone-assisted variants, the Traveling Salesman Problem with Drones (TSP-D) and the Vehicle Routing Problem with Drones (VRP-D) have attracted growing attention. Generally, these problems are solved using exact algorithms or metaheuristic algorithms. However, as the problem complexity increases and the scale of instances grows, these approaches often become less efficient. In this paper, we propose a reinforcement learning method with a shared attention encoder and a hierarchical dual-decoder architecture, where truck–drone coordination is achieved by first decoding the truck’s next node and then conditionally decoding the drone action. To further explore the solution space of large-scale instances, the proposed method adopts a multi-rollout learning strategy. We conducted experiments on large-scale TSP-D and VRP-D instances, and the results show that this model outperforms traditional metaheuristic algorithms in terms of both solution quality and computational efficiency.

Drones doi: 10.3390/drones10080615

Authors: Huiyu Wu Jie Liu Yazhou Wang Yimin Chen Jian Gao

Thruster health monitoring is essential for unmanned underwater vehicles (UUVs), where propulsion degradation can reduce manoeuvrability, tracking accuracy, and mission safety. Practical diagnosis remains difficult because labelled vibration samples are scarce and deployed vehicles may encounter fault states absent from the support library. Closed-set few-shot classifiers are unreliable in this setting because every query must be assigned to a known state. SAGE-PD (Spectral Attention-Guided Episodic Prototype Displacement) is a few-shot open-set method for UUV thruster vibration monitoring. It encodes each vibration window through raw-waveform and STFT branches, constructs episode-specific prototypes for known thruster states, and models their relational geometry. Unknown-state evidence is obtained by replacing the predicted prototype with the query embedding and measuring the displacement of the transformed prototype structure. Spectral attention weights this displacement toward thruster-related time–frequency components. On the analysed UUV thruster dataset, SAGE-PD achieved 0.9078±0.0573 Open OA and 0.9011±0.0407 AUROC in a 5-shot evaluation, and 0.8745±0.0457 Open OA and 0.8953±0.0409 AUROC in a 1-shot evaluation. The results show that SAGE-PD improves both known-state recognition and unknown-state rejection by combining support-structure compatibility with vibration-aware spectral evidence.

Drones doi: 10.3390/drones10080614

Authors: Weijian Jia Fenghua Wang Haiwen Zheng Penglei Hu Xiaobing Wang Yao Zhao Pengdong Zhang Yufei Gao

To address the susceptibility of single-modality approaches to illumination variations and imaging conditions in low-altitude unmanned aerial vehicle (UAV) detection, this study proposes a structure-aware heterogeneous dual-stream detection network based on infrared and visible-light fusion. First, a multimodal UAV detection dataset oriented toward complex low-altitude scenarios is constructed, providing a data foundation for cross-modal detection research. Then, a structure-aware heterogeneous dual-stream feature extraction framework is designed to enable collaborative modeling of visible-light and infrared features through modality-specific encoding. In the visible-light branch, a Structure-Aware Gated Enhancement Block (SAGE Block) is introduced to enhance the representation of fine-grained structural and edge information. In the cross-modal fusion stage, a Bidirectional Wavelet-Guided Fusion Module (BWFM) is proposed to decouple structural semantics and detailed information in the frequency domain. Adaptive fusion is further achieved through low-frequency cross-modal interaction and high-frequency detail-preservation strategies. Finally, the proposed method is experimentally validated on the proposed Multispectral UAV Detection Dataset (MUDD) and the Multi-scenario Multi-Modality Fusion Dataset (M3FD).. The experimental results show that the proposed method achieves an mAP@0.5 of 0.9680 and an mAP@0.5:0.95 of 0.6794 on the proposed MUDD, as well as an mAP@0.5:0.95 of 0.6072 on the M3FD dataset, demonstrating competitive detection accuracy and generalization capability. Ablation experiments further indicate that the SAGE Block, BWFM, and the low-frequency cross-modal fusion and high-frequency detail-preservation strategies within BWFM all contribute positively to performance improvement.

Drones doi: 10.3390/drones10080613

Authors: Xin Fan Hongxing Zheng Zhaoyang Wang

In maritime safety regulation, conventional ballast water inspection faces critical time-window violations due to manual sampling delays across dispersed inspection points. While drone-based inspection significantly enhances multi-point accessibility and timeliness, its operational viability is severely constrained by limited battery endurance during multi-vessel missions. To address this challenge, we propose a novel partial charging strategy enabling adaptive energy replenishment. This strategy is formalized in a model incorporating dual energy consumption (propulsion and hovering) and spatiotemporal constraints for sampling operations. An adaptive metaheuristic algorithm combining dynamic programming with large neighborhood search is developed for efficient resolution. Computational experiments demonstrate that partial charging strategy reduces drone energy consumption by 25% compared to the 70%-threshold charging strategy. This study provides maritime authorities with a quantifiable configuration decision tool to optimize drone fleets and flight parameters, balancing regulatory compliance with operational costs.

Drones doi: 10.3390/drones10080612

Authors: Chenrui Bai Ya Zhang Zehui Yuan Shuwen Zhao Chenghao Yang

Aiming at the problems of traditional Rapidly exploring Random Tree Star (RRT*), its modified algorithms in the three-dimensional complex underwater path planning of autonomous underwater vehicles (AUVs), including sampling redundancy, simplistic expansion mechanism, mismatch between planned paths and AUV motion constraints, and insufficient real-time performance, this paper proposes an improved bidirectional artificial potential field RRT* (Improved BI-APF-RRT*) algorithm. The algorithm adopts a hybrid sampling strategy to concentrate sampling in high-value regions and optimize node distribution. A three-level progressive expansion strategy is designed to balance fast convergence and excellent obstacle avoidance capability. Meanwhile, a dynamic target switching strategy is introduced to enhance the coordination efficiency of bidirectional search trees. Simulation results show that in three-dimensional underwater obstacle environments with different complexity levels, the proposed algorithm outperforms comparative algorithms such as GB-RRT*, BI-RRT*, and APF-RRT* in terms of path length, planning time, number of generated nodes, and iteration times. The proposed algorithm provides an efficient and feasible technical scheme for the deep-sea autonomous navigation of AUVs, and is of great significance for promoting the engineering application of underwater unmanned systems.

Drones doi: 10.3390/drones10080611

Authors: Ziyi Qiu Defu Lin Bo Liu Hui Han Wen Guo Jianjian Liang Zhaojiang Chen Ziheng Yan Haolong Wang Xinghao Yang Zelin Liu Liuhang Zhao

To address the problem of lacking effective evaluation benchmarks for UAV navigation algorithms in complex-structured, GNSS-denied industrial environments (e.g., fully enclosed stockyards), this paper proposes and open-sources a multi-modal UAV navigation dataset. The dataset is collected in a real steel plant enclosed stockyard, integrating LiDAR point clouds, IMU, RGB images, and high-precision total station ground truth trajectories, and specially designs ArUco markers to aid visual localization. Different from existing datasets targeting urban or campus scenes, this dataset realistically reflects the challenges of GNSS-denied signal, weak texture, high dust, and complex spatial grid structures in industrial environments. Through the evaluation of various mainstream LiDAR odometry and fusion navigation algorithms, the difficulties encountered by existing methods in this scenario are highlighted, and the potential of the proposed dataset as a valuable benchmark for developing and quantitatively testing highly robust navigation algorithms is suggested.

Drones doi: 10.3390/drones10080610

Authors: Jingjing Qian Yang Cheng Junhong Wu Bing Liu Wei Dai

Autonomous UAV landing in urban scenes requires a perception module that can identify candidate landing surfaces, reject structural and dynamic hazards, and express uncertainty before a downstream controller commits to a landing maneuver. This study reformulates UAV landing perception as a unified three-class landing-safety segmentation problem by relabeling UAVid, UDD6, and VDD into candidate landing area, structural obstacle, and critical hazard classes. The Landing3 dataset is constructed with an implementation-consistent relabeling protocol in which critical hazards override other labels and only sufficiently large connected components of source-specific candidate classes are retained as landing candidates. Two real-time segmentation models are evaluated under this unified task: PIDNet, a CNN-based multi-branch model with explicit boundary modeling, and SCTNet, a Transformer-guided model with semantic alignment for long-range context modeling. A post hoc conformal prediction (CP) module then converts softmax outputs into pixel-wise prediction sets, and the final candidate landing region is extracted only from pixels whose prediction set is the singleton candidate-landing class. Experiments compare the two models in terms of best validation mIoU, class-wise IoU, resolution-dependent accuracy–speed trade-offs, power-constrained FPS and complexity, qualitative candidate-area visualization, and CP-derived safe-area quality. A fixed-checkpoint sensitivity analysis further shows that the comparative model ranking remains stable across the tested connected-component threshold settings. SCTNet provides a lighter model and higher throughput under most tested resolutions and power limits, whereas PIDNet preserves higher safe-area recall, safe IoU, and spatial coherence after CP filtering. These results show that reliable UAV landing perception requires joint consideration of cross-dataset task definition, real-time model efficiency, and uncertainty-aware candidate-area extraction, rather than semantic segmentation accuracy alone.

Drones doi: 10.3390/drones10080609

Authors: Cuauhtemoc Acosta Lúa Bernardino Castillo-Toledo Stefano Di Gennaro Ulises Larios

This paper addresses the problem of autonomous landing of a quadrotor unmanned aerial vehicle (UAV) on a moving ground vehicle subjected to unknown vertical oscillations generated by road irregularities. The proposed approach considers simultaneous longitudinal, lateral, heading, and altitude regulation in the presence of nonlinear coupled dynamics, aerodynamic effects, and platform motion disturbances. A nonlinear control architecture is developed by combining backstepping techniques, super-twisting sliding-mode control, and an adaptive internal model regulator. The longitudinal, lateral, and heading subsystems are stabilized through a block backstepping–sliding-mode framework, whereas the altitude subsystem is regulated using an adaptive internal model controller capable of compensating unknown multi-frequency oscillatory disturbances without prior knowledge of their amplitudes or frequencies. The complete UAV dynamics are derived from the Newton–Euler formulation, including aerodynamic forces, gyroscopic effects, and coupled translational–rotational dynamics. To improve robustness and avoid algebraic differentiation, exact first-order differentiators based on the super-twisting algorithm are incorporated into the control implementation. The proposed adaptive regulator is compared against a robust super-twisting sliding-mode altitude controller under low- and high-frequency oscillatory platform motions. Simulation results demonstrate that the adaptive internal model regulator achieves accurate trajectory tracking and consistently lower accumulated tracking errors than the robust super-twisting sliding-mode controller under both low- and high-frequency platform oscillations. These results highlight the suitability of adaptive output regulation techniques for autonomous UAV landing operations under oscillatory platform conditions with measurement noise.

Drones doi: 10.3390/drones10080608

Authors: Zishuang Pan Dazhao Yu Wei Han Xichao Su Jie Wang Shansong Song Bing Wan

Carrier-based fixed-wing UAV landing is challenged by deck motion, carrier airwake, gust disturbances, strong trajectory–attitude coupling, and actuator constraints. To address these issues, this paper proposes a Predefined-Time Direct Lift/Side-Force-Integrated Approach Landing (PTDIAL) method. An integrated direct-force architecture is constructed using the trailing-edge flap for direct lift and the spoiler for direct side force, thereby reducing the dependence of trajectory correction on angle-of-attack- and bank-angle/sideslip-mediated regulation. A preview-based reference glide slope is generated from the predicted Ideal Touch Point (ITP) sequence to improve the response to deck motion. Predefined-time control laws are developed for the cascaded position, trajectory, attitude, angular rate, and velocity loops, with prescribed-performance constraints imposed on the attitude response. A predefined-time disturbance observer is introduced to estimate the lumped aerodynamic disturbances, while an auxiliary anti-saturation mechanism compensates for the effect of trailing-edge flap saturation. Lyapunov analysis establishes the practical predefined-time stability of the closed-loop system under bounded disturbances and actuator constraints. Various simulations demonstrate that the proposed architecture improves lateral and vertical tracking while preserving the UAV attitude, and Monte Carlo simulations further confirm the robustness of PTDIAL.

Drones doi: 10.3390/drones10080607

Authors: Ahmed Alamouri Cosima Berger Mohammad Shafi Bajauri Konstantin Wenzlaff

Risk assessment of Unmanned Aerial Vehicle (UAV) path planning is a crucial step towards ensuring a safe UAV operation. However, achieving a reasonable risk assessment of UAV flight paths remains challenging because it involves multiple responsibilities and processes that extend beyond a single agency or organization. Additionally, it must balance various complex factors and data from social, technical, political, and economic sources. Most existing works on flight path planning evaluate flight risks at a global level and generalized geometric representations of the UAV operating environment with respect to the current applicable UAV regulations. However, geometric information alone does not provide sufficient insight for comprehensive and safe path planning. Therefore, there are other ideas and concepts for using semantic data to characterize objects, obstacles and actions in the UAV environment. Incorporating semantic information into path planning enables more meaningful scene descriptions and a better representation of relevant constraints, obstacles within UAV environment that may influence the safety level of UAV operation, and the relevant risk assessment process. In this paper, we propose the development of methods and frameworks integrated into a flight planning prototype designed to generate safe two-dimensional UAV routes within a local, fine-grained planning context. The prototype incorporates safety considerations to enable a comprehensive assessment of UAV operational risks. It leverages both geometric and semantic datasets to characterize objects and obstacles within the UAV environment. These datasets are processed and stored in a relational database to support structured access and long-term usability. All concepts and experiments were implemented using datasets from a study area in the city of Brunswick, Germany.

Drones doi: 10.3390/drones10080606

Authors: Xujun Guan Junwu Luo Hai Zhang

In multi-UAV airborne radar sensing, target-level fusion cannot start until detections reported in different UAV coordinate frames have been associated. This paper focuses on a single-epoch case: two UAV-mounted radars observe a common target group, but the reported point sets are sparse, unordered, only partially overlapping, and corrupted by angular noise, missed detections, and additional observations. In this setting, appearance cues, long tracks, and dense point-cloud neighborhoods are absent, so the association problem is mainly a sparse geometric registration problem. We propose a Triangle Attribute-Guided Robust Association (TARA) framework. TARA describes each target triple by its absolute scale and two normalized side-length ratios. These attributes are not used as final identity labels. Instead, they retrieve, verify, and rank plausible cross-view triangle pairs, from which a fixed number of rigid-transformation hypotheses are estimated. The best coarse transformation is then refined by the trimmed ICP, and the final correspondences are obtained by gated one-to-one matching in the aligned frame. The ablation results show why the attribute-guided step matters. With only 100 evaluated hypotheses, TARA reaches 93.7% F1 in the medium setting, while blind RANSAC needs about 10,000 random hypotheses to reach comparable accuracy. In composite degradation tests, TARA obtains mean F1-scores of 98.7%, 94.2%, 69.3%, and 69.1% from low to extreme conditions. The advantage is clearest when clutter, missing detections, and angular uncertainty appear together because the triangle attributes make the limited hypothesis budget much less likely to be spent on geometrically implausible triples.

Drones doi: 10.3390/drones10080605

Authors: Xi You Wenjun Yi

Urban low-altitude unmanned aerial vehicle (UAV) navigation combines partial observability, building occlusion, wind disturbance, and continuous control. This study develops and evaluates HW-GA-SAC, a multimodal history-window Soft Actor-Critic (SAC) policy for procedurally generated three-dimensional MuJoCo cities. A Gated Transformer-XL (GTrXL)-inspired gated-attention encoder processes a fixed eight-step navigation history, while a current-frame safety branch supplies vertical clearance, sparse Light Detection and Ranging (LiDAR)-like range sectors, and handcrafted safety cues directly to the actor and critic. The policy uses obstacle-related observations and reward shaping to support collision avoidance; it does not include constrained policy optimization or a separate runtime safety filter. In a seven-method comparison using five training seeds and five evaluation layouts, HW-GA-SAC achieved a 96% ± 3% success rate, 207 ± 16 average return, and 3% ± 4% timeout rate. Feedforward SAC achieved 92% ± 11% success and a 7% ± 10% timeout rate, but its successful paths were more direct. Five-seed learning curves, city-split evaluation, wind sensitivity, sensing perturbations, inference profiling, and ablation studies further characterize the method. Within this simulation protocol, HW-GA-SAC provides the strongest completion-oriented performance, with a measurable trade-off between task completion and path directness.

Drones doi: 10.3390/drones10080604

Authors: Shang-En Tsai Chia-Han Hsieh Wei-Cheng Sun Sin-Dao Shen

This paper presents a distributed counter-UAV early-warning and response-decision support framework for low-altitude UAV defense. Acoustic–visual edge nodes generate local state packets and exchange compact information-filter parameters through Multi-Target Information Consensus (MTIC), avoiding centralized fusion and reducing payload bandwidth. The MTIC naïvety-handling mechanism is extended to heterogeneous acoustic–visual sensing, improving robustness to partial observations, packet loss, and node disconnection. A Fuzzy–Bayesian threat-assessment layer converts fused distance, velocity, and heading cues into interpretable response recommendations with calibrated confidence. Implemented on ROS 2/Fast DDS with tiered QoS, software-assisted IEEE 1588 synchronization, and Preempt-RT scheduling, the framework achieves within about 5% of centralized accuracy while reducing payload bandwidth by up to about 97% relative to the main centralized baseline. Simulation and hardware-in-the-loop tests on three- and five-node mesh topologies show software-assisted sub-millisecond synchronization (200–500 μs offset), bounded latency, gradual AUC degradation under association mismatch, and end-to-end feasibility under controlled packet loss. Overall, the system provides a resilient, deployment-oriented architecture for distributed C-UAV early warning.

Drones doi: 10.3390/drones10080603

Authors: Hongbin Liu Liang Zeng Mengyuan Lu Ke Tang Bo Li Xinping Zhu

Urban low-altitude UAV missions require efficient, risk-aware path planning and rapid response to dynamic airspace changes. This study proposes a hierarchical planning and dynamic replanning framework based on the Geographic coordinate Subdivision grid with One-dimensional integer coding on a 2n-Tree (GeoSOT) and height-layer encoding (GeoSOT-H). The framework constructs a multi-granularity 3D semantic-risk voxel model and uses semantic-triggered refinement to limit fine-resolution modeling to flight-relevant high-risk regions. The Hierarchical Semantic-risk-aware Path Planning with Corridor-constrained A* (HSPC-A*) algorithm generates a macro-corridor and conducts fine-level search to balance path length, semantic-risk exposure, and vertical maneuvering cost while satisfying no-fly constraints. Its output is a connected L24-H voxel-center path for subsequent navigation or post-processing. Experiments in a 2.89 km2 urban area show that explicit storage is reduced to 16.4% of full-domain L24-H voxels. Compared with conventional 3D A*, HSPC-A* slightly increases path length from 2183.06 m to 2202.38 m, while reducing average semantic risk from 8.5927 to 5.2595, eliminating high-risk samples, and reducing search time from 164.09 s to 3.74 s. Code-based updating achieved a 104.6-fold speedup, and two-branch replanning handled both corridor-retained and corridor-disconnecting no-fly events, jointly demonstrating the trade-offs among path length, semantic-risk exposure, computational efficiency, and compliance with modeled flight-safety constraints.

Drones doi: 10.3390/drones10080602

Authors: Emmanouel T. Michailidis Irene S. Karanasiou

Unmanned aerial vehicles have become indispensable components of modern defense systems by conducting Intelligence, Surveillance, and Reconnaissance (ISR) missions to collect critical operational information through onboard sensing technologies, supporting secure communication and information sharing among distributed military assets, and enhancing battlefield situational awareness through real-time sensing and data fusion. In addition, UAVs are increasingly capable of executing a wide range of defense missions, including target search and tracking, electronic warfare, search-and-rescue, and combat support. The integration of AI into UAV-based systems has the potential to enhance these operational capabilities by enabling intelligent perception, autonomous decision-making, adaptive mission planning, autonomous navigation, resilient communications, and cooperative multi-UAV coordination, thereby enabling the autonomous and collaborative execution of complex defense missions. This paper presents an up-to-date review of AI-enabled UAV-based defense systems, focusing on major operational domains including autonomous air combat and cooperative UAV operations, path planning and autonomous navigation, target tracking/detection/classification, cybersecurity, electronic warfare protection, and resilient UAV operation. In addition to surveying the recent literature, this paper provides an integrated system architecture, a functional classification framework, and an analysis of the AI paradigms enabling next-generation UAV-based defense systems. Furthermore, this review synthesizes the key technological trends, lessons learned, and cross-domain research challenges identified across the reviewed studies, providing a unified perspective on the current state of the field. Finally, this paper highlights promising future research directions for resilient, scalable, secure, and intelligent next-generation AI-enabled UAV-based defense systems.

Drones doi: 10.3390/drones10080601

Authors: Jianfeng Wang Jun Li Yuze Liu Cheng Wang Chen Bu Shuai Feng Mingying Huo

Low-aspect-ratio flying-wing unmanned aerial vehicles (UAVs) are attractive drone platforms for civilian remote sensing, environmental monitoring, infrastructure inspection, disaster assessment, and persistent public-service monitoring because their integrated tailless layout offers high aerodynamic efficiency and payload volume. A trajectory-command-based three-loop nonlinear dynamic inversion (NDI) control architecture enhanced by a nonlinear disturbance observer (NDO) is designed to address the critical challenges of rapid time variation, strong nonlinearity, strong coupling, and restricted yaw authority in low-aspect-ratio flying-wing UAVs. The core innovation lies in the development of a trajectory-command-to-attitude kinematic mapping mechanism, integrated with the NDO for active torque compensation of lumped uncertainties and time-varying external disturbances. Leveraging a mathematical model of a low-aspect-ratio flying-wing UAV standard model, a three-loop NDI controller comprising angular rate, attitude, and trajectory command loops was designed based on the time-scale separation principle. The NDO was further designed to estimate lumped disturbances and provide feedforward compensation, thereby establishing an NDI-DO system that mitigates the high sensitivity of conventional NDI to modeling inaccuracies. Simulation and robustness tests involving typical high-angle-of-attack maneuvers (e.g., Cobra and Split-S maneuvers) demonstrated that the NDI-DO system achieved a reduction in angular-rate tracking error by over 77.2% compared to the baseline NDI. Furthermore, the permissible range of aerodynamic parameter perturbations was improved by 23%, significantly enhancing tracking fidelity and disturbance rejection. In a 3-DOF wind tunnel free-flight test, the NDI-DO system achieved a substantial expansion of the controllable angle-of-attack (attitude-stability) envelope from 72.9° to 99.19°, substantiating the high reliability and engineering utility of the control framework in post-stall nonlinear regimes. These results indicate that the proposed NDI-DO framework can support safer envelope expansion, autonomous upset recovery, and robust flight control for civilian flying-wing drones operating under uncertain aerodynamic and environmental conditions.

Drones doi: 10.3390/drones10080600

Authors: Dong Zhao Huaizhi Dong Wenjing Ren

With the increasing deployment of multi-unmanned aerial vehicle (multi-UAV) systems in dynamic environments, the problem of efficient cooperative path planning has emerged as a critical challenge requiring urgent solutions. To address this issue, this paper proposes a novel joint optimization framework, named spatio-temporal attention-based multi-agent deep deterministic policy gradient (STA-MADDPG). Rather than proposing a new reinforcement learning algorithm in the strict sense, this work integrates advanced spatial-temporal feature extraction with heuristic gradient guidance. First, a cascaded architecture combining multi-head attention and Long Short-Term Memory (LSTM) networks is utilized to extract key local and temporal features, thereby mitigating the dimensionality curse in dense multi-agent observations. Second, an improved dynamic artificial potential field (DAPF) is integrated into the reinforcement learning framework as a state augmentation mechanism, providing heuristic guidance vectors that accelerate convergence and improve obstacle avoidance. Furthermore, to balance computational complexity and adaptive behavior, a rule-based hierarchical formation strategy is designed. The framework maps predefined formations (elliptical, chain, or wedge) to specific environment categories, while the underlying MARL policy governs the dynamic trajectory planning and topology maintenance. Finally, rigorous comparative and ablation experiments are conducted to evaluate path length, search time, and relative position errors. Statistical analysis demonstrates the effectiveness of the proposed framework, achieving up to a 67.3% reduction in search time and a 91.56% search success rate compared with standard MARL baselines in complex environments.

Drones doi: 10.3390/drones10080599

Authors: Bing Han Jianbin Chen Jianxin Peng Haojie Man

In complex operational environments, signal ambiguity and false alarms make task allocation for multiple unmanned aerial vehicles (UAVs) challenging. To address this problem, we propose the Bayesian Rolling Information-Theoretic Reconnaissance Action Assignment (RITRA) framework. RITRA integrates Bayesian belief tracking, expected information gain (EIG), and a cooperative criticality model based on Shapley values in a unified receding-horizon strategy. By assessing reconnaissance and intervention utilities for each UAV–hazard pair, RITRA casts mission planning as a dynamic optimization problem. We use an enhanced Hungarian algorithm with idle states, allowing UAVs to remain on standby when no deployment has positive expected value. Ablation results indicate that combining Bayesian tracking, active information collection, and risk-aware execution contributes to robust performance under the evaluated uncertainty conditions. In the 200-run paired Monte Carlo stress test, RITRA achieved a mean mission utility of 9.4362, a normalized mission score of 0.7890, and 0.245 false deployments per run. Against the best-performing baseline, a CBBA-based allocation method, RITRA achieved a mean paired mission-utility gain of 1.4925. On publicly released dynamic MRTA instances augmented with false-alarm uncertainty, RITRA achieved the highest aggregate mission utility, normalized mission score, and mission accuracy among the compared controllers, providing additional evidence of its effectiveness. These results indicate that RITRA supports risk-aware multi-UAV allocation under the evaluated uncertain conditions.

Drones doi: 10.3390/drones10080598

Authors: Ishaq Ahmed Jun Lu Talha Younas Ghulam Farid Muhammad Bilal Sohaib Tahir Chauhdary

Reliable fault detection and isolation (FDI) for actuators and sensors in underactuated autonomous underwater vehicles (AUVs) is challenging because nonlinear hydrodynamic coupling redistributes fault signatures, and persistent ocean currents can trigger false alarms. This paper presents a reinforcement learning (RL)-augmented dual-layer FDI framework for actuator and sensor faults in underactuated AUVs. The actuator layer uses a robust adaptive propagated interval observer (RAPIO) that evaluates thruster and control-surface residuals against a calibrated dynamics-consistency tube. The sensor layer forms estimator-consistency residuals for Doppler velocity log (DVL), depth, and inertial measurement unit (IMU) measurements against a reference-separated finite-time extended state observer (FTESO). An offline-trained soft actor–critic (SAC) policy schedules bounded actuator uncertainty margins and sensor alarm thresholds according to operating confidence. The scheduled actuator error dynamics remain Metzler and Hurwitz, preserving positive interval propagation and center-error input-to-state stability (ISS) independent of policy convergence. A Schmitt-trigger alarm and signal-space disambiguation rule classify healthy, actuator-only, sensor-only, and simultaneous-fault conditions under explicit residual-separation and persistence conditions. Across 72 simultaneous-fault episodes over a 4×6 uncertainty–current grid, the proposed method achieved 100% detection coverage for both actuator and sensor faults with only five false-alarm events, retaining full coverage in the severe-current, high-uncertainty subset where the selected actuator and sensor baselines achieved only 88.9% and 70.4% detection, respectively, with more false alarms. These results indicate that the proposed bounded RL scheduler can deliver reliable, certifiable actuator and sensor fault diagnosis under significant operational uncertainty.

Drones doi: 10.3390/drones10080597

Authors: Xiao Yan Yue Ma Jinwen Zhou

Motion control of tilt-servo fully vectoring UAVs (TSFV-UAVs) is challenging due to their highly nonlinear dynamics. This paper proposes a reinforcement-learning-based training framework for the hierarchical control architecture of TSFV-UAVs. The framework introduces an adaptive guidance mechanism and a varying-gradient reward function to improve training convergence. In addition, a wrench residual penalty term is incorporated into the reward function to help the agent identify the boundary of the reachable wrench set, thereby reducing the frequency of actuator saturation and improving the task success rate of the UAV. The results show that the proposed method not only effectively improves training convergence, but also enables the trained agent controller to achieve significant improvements in tracking performance and task success rate (93.0%), while exhibiting good robustness. Finally, hardware-in-the-loop experiments verify the practical deployability of the trained controller on embedded systems.

Drones doi: 10.3390/drones10080596

Authors: Joel Samu Chuyang Yang Kush R. Poddar

Unmanned Aerial Vehicle (UAV) swarms are increasingly deployed in search and rescue (SAR) missions to rapidly locate survivors. However, deploying autonomous swarms in low-altitude airspace shared with crewed rescue aircraft poses significant algorithmic and safety challenges. Following PRISMA 2020 guidelines, this systematic review synthesizes 44 peer-reviewed studies (2022–2026) to evaluate the literature across algorithmic optimization, reality-gap limitations, tactical deconfliction, and validation maturity. The synthesis reveals a consistent trend toward decentralized swarms, driven by Deep Reinforcement Learning in dynamic environments and by bio-inspired metaheuristics for static coverage. Despite these algorithmic advancements, the literature exhibits a severe reality gap: approximately 86% of evaluated models rely exclusively on idealized software simulations, abstracting away critical constraints like communication denial and sensor noise. Furthermore, most models assume uncontested airspace and lack the Manned–Unmanned Teaming (MUM-T) and tactical deconfliction protocols necessary for safe coexistence with rescue helicopters. To achieve true operational readiness within the critical “Golden 72 Hours” of disaster response, the discipline must transition toward hardware-in-the-loop and physical field trials, natively integrating airspace deconfliction into core swarm optimization loops.

Drones doi: 10.3390/drones10080595

Authors: Zhengyang Cao Liang Wang

Air-ground collaborative missions impose coupled constraints on unmanned aerial vehicles (UAVs), including restricted platform envelopes, rapid payload reconfiguration, transient interface impacts, and center of gravity (CG) shifts. Conventional serial design workflows fixing overall layout before local lightweighting are insufficient to satisfy structural mass reduction, interface safety, and attitude recovery requirements simultaneously. This study proposes a task-constraint-embedded multi-level lightweight structural design method for UAVs operating with ground mobile platforms. The layout-evaluation framework incorporates ground-platform envelope limits, UAV payload distribution, CG migration, and attitude-stability requirements through a hierarchical screening and score-based selection process. Mass reduction rate, CG shift, and layout compactness are used as the main layout-evaluation indicators, and a feasible layout is selected for subsequent structural zoning and verification. Subsequently, the airframe is partitioned into load-bearing, non-load-bearing, and interface zones, for which main-skeleton topology optimization, honeycomb sandwich lightweighting, and local interface reinforcement are applied, respectively. The optimized prototype is validated through finite-element simulation, interface-impact testing, and full-scale flight trials. Relative to the defined traditional baseline, the selected design reduces the whole-airframe mass by 19.2%, limits the payload-switching-induced CG shift to 2.80 cm, and shortens the attitude-settling time by 39.7%. The main verification indicators show simulation-test deviations below 4.2% for the selected layout and representative verification cases. These results indicate that embedding task-specific constraints into the overall layout-evaluation stage can improve the balance among lightweighting, interface-load resistance, and dynamic-stability requirements.

Drones doi: 10.3390/drones10080594

Authors: Xingda Li Jianqiang Zhang Yiping Liu Pengfei Zhang Ling Tan

Unmanned surface vehicle (USV) swarms operating in communication-denied maritime environments face degraded formation control when inter-agent state exchange is disrupted. This paper presents a three-layer control architecture integrating (1) LLM-assisted strategic mission planning with formal safety verification, (2) predictive tactical coordination combining physics-based motion extrapolation with online-learned neighbor behavior models, and (3) DMPC + ADMM execution for constrained formation control. The predictive coordination module in simulation-based evaluation across five representative scenarios reduces formation error by 76.0% (under simulation conditions) during 60 s communication outages compared to zero-hold prediction (Cohen d = 0.88, p < 0.001, DMPC + ADMM validated). The adaptive topology manager dynamically selects among star, mesh, and tree configurations via a utility function balancing communication quality, threat exposure, and overhead, reducing communication overhead by 71.7% (for the particular cases investigated) under the evaluated conditions while maintaining formation accuracy. Stability analysis using multiple Lyapunov functions and average dwell time theory guarantees global uniform exponential stability under topology switching, with explicit error bounds under communication denial derived via Gronwall-type arguments. The framework is validated through 1250 simulation trials across five scenarios with rigorous statistical analysis. The three-layer temporal decoupling architecture provides a practical template for safely integrating LLM-assisted planning with real-time multi-agent control in contested environments.

Drones doi: 10.3390/drones10080593

Authors: Fabrizio Ricardo Cahuas-Talledo Juan Eduardo Velázquez-Velázquez Alberto Luviano-Juárez

This article examines the challenge of trajectory inference for an unknown system affected by external disturbances, with the objective of reconstructing the trajectory of a quadrotor using a reference model. The proposed methodology extends the Closed-Loop Output Error (CLOE) scheme through two complementary contributions: an identified gain, incorporated into the reference model to guarantee the Hurwitz condition of the closed-loop error dynamics, and a set of sliding-mode correction terms that further accelerate error convergence and enhance robustness against bounded disturbances. The stability of both contributions is formally established via Lyapunov-based analysis. The proposed approach is validated through realistic simulations carried out in the CoppeliaSim robotics environment, considering both constant and time-varying trajectory scenarios. Results show that the hybrid approach improves trajectory inference accuracy and convergence speed, maintaining resilience under adverse conditions, making it a promising alternative for autonomous quadrotor monitoring.

Drones doi: 10.3390/drones10080592

Authors: You Wang Jiayi Xu Mengting Lin Lu Dong Gui Fu Keye Yan

To address the performance degradation of UAV object detection under low-light conditions, we develop an end-to-end object detection network. This proposed method integrates contrastive learning into the detection framework and establishes feature consistency constraints between low-light and normal-light images through a hierarchical contrastive selection encoder. Since the encoder is required only during training and removed during inference, the proposed framework improves feature robustness without introducing additional inference cost. To further improve object detection accuracy, Frequency Guided Dynamic Attention (FGDA) is introduced into the object detection network, focusing on resolving the issue of redundant interference during feature transmission and enhancing feature representation capability. To improve multi-level spatial feature fusion, an Adaptive Gated Dual-Spatial Fusion (AGDSF) module is further developed, which adaptively strengthens target-relevant responses while weakening background noise. According to the experiments on the VisDrone (dark) dataset and a self-collected nighttime UAV-dark dataset illustrate that the proposed method ensures heightened detection accuracy with low computational overhead, complying with the real-time and robustness requirements of UAV perception in low-light contexts.

Drones doi: 10.3390/drones10080591

Authors: Shaochun Qu Yuhuan Cai Shiyan Wan Yanfang Fu

To address the challenges faced by UAV swarms in local narrow-space penetration missions, including constrained passages, dense obstacles, and the difficulty of balancing formation stability and traversability, this paper proposes a local penetration method that integrates virtual–center consensus-based formation control with a V-shaped formation self-reconfiguration strategy. First, a virtual geometric center is introduced as the consensus reference to replace the traditional physical leader node, thereby reducing the risk of single-point failure and improving swarm coordination consistency. Second, geometric constraints, including the effective channel width, lateral formation width, and safety margin, are incorporated to construct a channel-constraint-driven formation self-reconfiguration mechanism, enabling the swarm to contract its formation, avoid obstacles during traversal, and recover the formation after passing through the constrained region. Finally, a digital twin-based virtual–real interactive validation platform is constructed to verify the formation maintenance, formation reconfiguration, and virtual–real trajectory consistency of the proposed method. Experimental results show that the proposed method can maintain favorable formation consistency and safe inter-UAV distances in constrained channel environments while demonstrating good stability and scenario adaptability during formation adjustment and recovery.

Drones doi: 10.3390/drones10080590

Authors: Minghao Bai Xiaojin Lv Haomeng Ma Chenghai Gu Peiyan Liu Yang Zhang Fulai Liang

Bio-radar is widely used for casualty search and rescue. However, its short-range and handheld operation mode limits the efficiency of wide-area detection. This paper applies UAV-borne fully polarimetric synthetic aperture radar (PolSAR) to the detection of stationary human targets at long ranges and experimentally validates the feasibility of this framework. To address the difficulty of detecting weakly scattering stationary human targets in strong clutter backgrounds, we propose a novel framework that integrates clutter suppression, target enhancement, and false-alarm suppression. The method introduces, for the first time, polarimetric scattering mechanism analysis into human target enhancement. This improves the SCNR of stationary human targets by approximately 13 dB on average. Then, using neighborhood density features, we reduced the number of non-zero pixels in non-human-target areas (NPINA) by 93.8% and 96.9% in the two passes, respectively. Finally, through dual-pass local binary-map correlation, we obtained the locations of the human targets. This study serves as a proof-of-concept and presents preliminary experimental results on the feasibility of UAV-borne PolSAR for long-range stationary human target detection. The experiment was conducted in a parking lot adjacent to a highway, with three subjects lying supine or prone on a gravel surface, and the scene included buildings such as garages and vegetation such as grass. The experimental results indicate that, under the tested conditions, UAV-borne fully polarimetric SAR is feasible for long-range stationary human target detection. This study expands the application scope of SAR and provides a new technical approach for rapid UAV-based human target detection.

Drones doi: 10.3390/drones10080589

Authors: Mariángeles del-Valle-García Gabriel Navarro Alejandro Román Juan Antonio López-Ramírez Luis Barbero Alberto Chisvert Guillem Peris-Pastor Gonzalo Martínez Marco Talone Antonio Tovar-Sánchez

Sunscreen-derived ultraviolet (UV) filters are emerging pollutants of increasing concern in coastal ecosystems. This study presents the development, optimization, and field validation of Mar_Spectra, a compact spectroradiometric system mounted on an unmanned aerial vehicle (UAV) designed to detect sunscreen UV filters in surface waters. The system integrates an Ocean Optics microspectrometer with a Raspberry Pi microprocessor, a GPS receiver, an altimeter, and an RGB camera for synchronized image acquisition. Spectral measurements were processed to derive remote sensing reflectance (Rrs) using a Zenith Lite reference panel. Controlled calibration experiments confirmed the system’s sensitivity in the 300–380 nm range and allowed the definition of a convenient index iS exhibiting a strong linear relationship with total UV-filter concentration (R2 = 0.964). Environmental tests indicated that wind, turbidity, and algal presence reduced detection capacity, while stable solar conditions improved spectral consistency. Despite its sensitivity to environmental variability, the Mar_Spectra device demonstrated consistent and reproducible performance under real field conditions, providing a cost-effective and scalable approach for near-real-time monitoring of UV filters and enhancing spatially resolved coastal pollution surveillance

Drones doi: 10.3390/drones10080588

Authors: Dongxu Yang Wanbing Ren Yanren Li Shengmao Zhang Zuli Wu Tianfei Cheng Fei Wang

Accurate spatial localization of small, transient targets in low-texture aquatic environments remains a fundamental challenge in UAV-based remote sensing, where open-water surfaces often lack stable tie points, degrading exterior orientation estimation and conventional photogrammetric georeferencing. An integrated UAV framework combining DG/AAT-BA georeferencing with deep-learning-based oriented bounding box (OBB) detection was implemented for high-precision localization, validated on the Critically Endangered Yangtze finless porpoise (YFP, Neophocaena asiaeorientalis) in the Yangtze–Poyang Lake system. The georeferencing component selects direct georeferencing (DG) in open-water scenes and automated aerial triangulation with bundle adjustment (AAT-BA) in feature-rich nearshore scenes. Validation using two static verification points showed that, relative to DG, AAT-BA reduced geometric georeferencing RMSE from 2.59 to 0.62 m under straight-flight conditions and from 3.61 to 0.67 m under turning-flight conditions. For target detection, a lightweight Laplacian edge-enhancement convolution module (LapConv) was incorporated into YOLO-OBB backbones, amplifying weak-edge and low-contrast features of partially submerged targets. Across four representative YOLO-OBB models and three group-constrained partitions, LapConv consistently improved the mean mAP@0.5, with gains of 0.026, 0.024, 0.019, and 0.026 for YOLOv8, YOLO11, YOLO12, and YOLO26, respectively. Applying this framework to six UAV missions across three ecologically and hydrologically distinct subregions enabled georeferenced mapping of porpoise distributions and visualized spatial distribution characteristics during the survey period. The approach is reproducible, minimally invasive, and potentially transferable to UAV-based monitoring of other small aquatic wildlife, providing a methodological basis for fine-scale spatial surveys and subsequent habitat analysis.

Drones doi: 10.3390/drones10080587

Authors: Baogang Li Bing Sun Daqi Zhu Wenyang Gan

Autonomous underwater vehicles (AUVs) are increasingly used in seabed inspection, underwater search, ocean observation, and infrastructure maintenance. However, feasible and safe trajectory planning in dense three-dimensional underwater environments remains challenging because AUVs must avoid multiple irregular static obstacles, react to moving obstacles, and maintain robust navigation performance under ocean-current disturbances. Traditional path planning methods can suffer from high computational cost or poor trajectory smoothness in dense 3-D environments. The interfered fluid dynamical system (IFDS) provides a promising flow-field-based planning mechanism by treating obstacles as disturbance sources in a virtual fluid field. Nevertheless, the performance of IFDS strongly depends on the repulsive and tangential parameters, which are usually selected empirically and may not provide an optimal trade-off among path length, smoothness, safety margin, and energy consumption. To address these issues, this paper proposes a grey wolf optimization-enhanced IFDS method, termed GWO-IFDS. First, static and dynamic underwater obstacles are modeled using unified super ellipsoid implicit functions, allowing spheres, cylinders, ellipsoids, reefs, and seabed mounds to be described in a common mathematical form. Second, a 3-D IFDS planner is developed to generate collision-free streamlines by combining the attractive flow toward the target and obstacle-induced modulation matrices. Third, a grey wolf optimization algorithm is introduced to optimize the IFDS repulsive and tangential parameters by minimizing a scalarized multi-criteria fitness function that considers path length, terminal error, trajectory smoothness, energy proxy, and minimum obstacle clearance. Finally, simulation studies are conducted under four representative scenarios: multiple static obstacles, mixed static and dynamic obstacles, mixed obstacles with ocean-current disturbances, and parameter-optimization comparison among GWO, PSO, DE, BO, GA, and fixed-parameter IFDS. The results demonstrate that the proposed GWO-IFDS method can generate smoother and safer trajectories with lower steering-effort-related cost than fixed-parameter IFDS. Additional tests under sonar-like perception uncertainty, bounded steering constraints, and different weight settings further verify the robustness and feasibility of the proposed framework.

Drones doi: 10.3390/drones10080586

Authors: Dong Xie Xinyi Cheng Yuqing Feng

Submerged aquatic vegetation (SAV) is a sensitive indicator of shallow-water condition, but operational mapping has relied on multispectral, red-edge, or near-infrared sensing, leaving low-cost consumer RGB drones largely unused for the task. We present an RGB-only detector for SAV. It treats SAV as local darkening rather than a greenness signal, and sets the detectable patch scale with two label-free mechanisms, a ground-sample-distance (GSD)-anchored minimum mapping unit that rescales with altitude and a CFAR-inspired scale criterion N·(s/σ)2 ≥ τ driven by per-scene clutter σ. Across four densely annotated scenes most SAV is small, with 73–98% of labelled patches falling below the 0.43 m2 that a conventional fixed lower limit on pixel number retains at 4 cm GSD. Anchoring the minimum unit correctly raises pixel recall from 0.52 to 0.71 with no false positives inside the annotated open-water rectangles, and reaches a cross-scene SAV-versus-open-water AUC of 0.969, whereas an excess-green index performs near chance (AUC 0.438). Consumer RGB drones can therefore map binary SAV presence and extent within a darkness × scale detectability envelope, recovering the small, fragmented stands that coarser sensors miss, while cover fraction, biomass, species identity and the faintest SAV remain out of reach. The reference standard throughout is expert photointerpretation of 4 cm imagery, so what is mapped is image-interpreted SAV-like dark patches rather than field-confirmed SAV, and all results come from one lake system, one camera and one flight altitude.

Drones doi: 10.3390/drones10080585

Authors: Jianfeng Xie Siyuan Wang Jiandong Zhang Qiming Yang Shuling Dai

To address the limitations of static-obstacle-based route planning in forest fire missions, this study develops a three-dimensional route-planning method for fixed-wing unmanned aerial vehicles (UAVs) that accounts for time-varying fire and smoke threats, complex terrain, and flight-dynamics constraints. A cellular automaton models fire spread with wind, slope, fuel, and moisture effects, while a Gaussian plume model estimates smoke concentration. The resulting burning and high-concentration smoke cells are encoded as dynamic three-dimensional threat envelopes and local grid masks. A hierarchical proximal policy optimization (H-PPO) architecture then combines a high-level stateful long short-term memory (LSTM) policy for route-subgoal generation with a pretrained low-level flight controller that produces continuous throttle and control-surface commands in JSBSim. In 100 independent simulation tests, the complete H-PPO model achieved a 100% task success rate, a mean terrain clearance of 771.92 m, and an average online decision time of 1.290 ms. Compared with A* and RRT*, H-PPO provided higher task reliability, greater mean terrain clearance, and lower online computational cost. The results show that hierarchical temporal decision making improves safety-prioritized planning in evolving fire and smoke environments, although conservative avoidance increases route length and mission duration. Further real-world and flight-test validation is required.

Drones doi: 10.3390/drones10080584

Authors: Kun Zhang Huayu Gao Tianzhu Ren Yimeng Liu

Resource assignment in heterogeneous multi-UAV systems changes both current task outcomes and the task chains that remain feasible across a mission region. Completion, delay, utilization, and role-specific reachability metrics do not retain this spatially resolved option space. This paper proposes the grid-based functional feasible domain (G-FFD), which represents capability- and time-constrained sensing–execution UAV chains subject to an intermediate decision/communication delay. A generalized index aggregates profile-labeled feasible chains and recovers feasible-chain count as the single-profile, unit-weight case. Shared resources and chain overlap describe functional coupling between grids. Controlled simulations show that local task bursts remove alternatives from external grids, with realized loss depending on the candidate-domain size, coupling, accepted demand, and occupied resources. Across 5 × 5, 10 × 10, and 20 × 20 grids, the high-G-FFD region has the largest mean peak external loss, although magnitudes and other regional rankings remain resolution-dependent. In a 20-seed paired stress test, G-FFD-guided selection reduces the cumulative external deficit relative to earliest-finish selection by a mean of 2.519 retention-minutes (bootstrap 95% confidence interval: 1.073–4.089). The differences in overall completion and post-burst completion of external medium/high-profile tasks remain inconclusive. G-FFD thus complements task outcomes with information about remaining regional alternatives and assignment consequences.

Drones doi: 10.3390/drones10080583

Authors: Hao Xiong Minghu Tan Xiaoyu Liu Haoyu Li

A cooperative guidance law based on the experience-guided multi-agent proximal policy optimization (E-MAPPO) algorithm is proposed for multiple unmanned aerial vehicles (UAVs) to track dynamic points of interest in civilian applications, such as collaborative search and rescue and environmental monitoring. In multi-UAV cooperative tracking, accurate arrival-time coordination is important for improving collaborative task execution, but it remains challenging because of continuous action spaces, target maneuvering, uncertain time-to-go estimation, and inefficient exploration in multi-agent reinforcement learning. Specifically, a multi-UAV cooperative guidance environment is formulated, and the problem is modeled as a Markov decision process. To address the challenges of large action spaces and poor convergence in multi-agent reinforcement learning, an experience-guided MAPPO framework is introduced to enhance training efficiency and policy stability. Different from standard MAPPO, the proposed E-MAPPO introduces proportional-navigation-guided experience only during the early training stage to guide exploration, while the final policy is still optimized through the MAPPO objective. Subsequently, a composite reward function is designed by integrating distance-based heuristic terms with auxiliary guidance signals, thereby improving exploration efficiency and facilitating coordinated rendezvous and tracking of dynamic references. Comparative simulations with cooperative proportional navigation guidance (CPNG), sliding mode control (SMC), and standard MAPPO are conducted under different target motion scenarios. The results show that E-MAPPO reduces the average convergence step by 17.07% compared with MAPPO. In the straight-moving target scenario, E-MAPPO reduces the cooperative time error by 55.10% compared with CPNG and by 8.33% compared with MAPPO. In the S-type maneuvering target scenario, E-MAPPO reduces the cooperative time error by 55.81% compared with CPNG and by 9.52% compared with MAPPO. Monte Carlo experiments further verify its effectiveness and robustness. Additional robustness tests under Gaussian measurement noise, observation bias, and communication delay show that the proposed method maintains acceptable tracking accuracy and cooperative timing performance under different uncertainty conditions. In addition, the results indicate that the proposed method generalizes well to different types of maneuvering targets.

Drones doi: 10.3390/drones10080582

Authors: Seyedreza Fattahi Massoum Hugh H.-T. Liu

This paper studies energy-aware trajectory generation for the gate-passage maneuver of a multirotor-slung load system with a varying-length cable. The mission requires load and drone gate clearance while enforcing positive cable tension and recovered thrust, body torque, and winch torque limits, resulting in a highly constrained nonconvex problem. Using conic programming, a conservative convex optimization formulation is developed whose surrogate objective provides an upper bound on the original objective over a restricted feasible set. The conic construction enables fast solution and global optimality with primal–dual certificates for the solved restriction. In the two numerical scenarios, it reduces solution time by approximately 74% and 90% relative to the locally optimal nonconvex solutions, while increasing the actuator effort proxy and relative electrical energy estimate by only approximately 4%–5%. Post-optimization checks verify positive actuator margins, while closed-loop simulations demonstrate trackability and retain positive gate and cable tension margins. The proposed conic formulation provides a significant computational improvement at the cost of only a modest increase in relative energy demand.

Drones doi: 10.3390/drones10080581

Authors: Anran Du Huiqi Xu Wenqiang Yao

Detecting ship critical components in UAV-based maritime monitoring remains challenging because of small-object detail loss, multi-scale semantic misalignment, and limited computational resources on embedded platforms. To address these issues, this study proposes D-MobileNetV3, a lightweight anchor-free detection algorithm based on an enhanced CenterNet framework. In the proposed framework, MobileNetV3 is used as the lightweight backbone, and an RFB module is introduced to improve multi-scale contextual representation. A High-Frequency and Spatial Perception Feature Pyramid Network (HS-FPN) with a serially coupled dual-path architecture is then constructed by integrating the High-Frequency Perception (HFP) and the spatial dependency perception module (SDP), enabling edge-detail enhancement and spatial-semantic alignment during multi-scale fusion. In addition, a lightweight decoupled detection head with SIoU loss is designed to improve the localization of slender components such as waterlines and masts. Experiments on the in-house VitalShips dataset and the public SeaShips dataset show that D-MobileNetV3 improves detection accuracy and generalization relative to the baseline models while maintaining low computational cost. UAV flight tests on the Jetson AGX Xavier platform achieve an average processing time of 16.3 ms and an mAP50 of 83.5%, indicating that the proposed method provides a high-precision, low-latency solution for maritime surveillance.

Drones doi: 10.3390/drones10080580

Authors: Samuel Cartwright Maxime Collignon Adolfo Perrusquía Antonios Tsourdos

The increasing frequency and intensity of wildfires has created an urgent demand for scalable and autonomous wildfire response systems. While recent advances in multi-agent reinforcement learning (MARL) have demonstrated promise for collaborative uncrewed aerial vehicle (UAV)-based wildfire suppression, most existing approaches rely on simplified fire propagation dynamics and highly centralised learning architectures that are difficult to deploy in realistic operational settings. This paper presents a decentralised MARL framework for wildfire suppression that combines stochastic wildfire propagation, wind-driven spread dynamics, and communication-aware multi-agent coordination. The proposed framework extends an existing probabilistic wildfire environment through the incorporation of wind speed and directional effects, producing highly asymmetric and stochastic wildfire behaviour that more closely resembles real wildfire propagation. A decentralised Deep Q-Network (DQN) architecture is then introduced in which UAV agents learn independently through individual replay buffers. To mitigate the sparse-learning challenges introduced by decentralisation, selective experience sharing based on the SUPER algorithm is incorporated, enabling agents to exchange only high-value experiences under realistic communication constraints. Experimental results demonstrate that selective communication significantly improves containment performance and learning efficiency while preserving decentralised execution. The work highlights both the feasibility and challenges of realistic UAV swarm coordination for wildfire suppression, particularly the trade-offs between communication bandwidth, environmental stochasticity, and collaborative performance.

Drones doi: 10.3390/drones10080579

Authors: Weizhong Chen Zeyan Li Shuo Dong

This paper presents an observer-based active fault-tolerant control scheme for leader–follower formations of differential-drive unmanned ground vehicles under the Separation-Bearing Control framework. The relative pose is described by inter-vehicle distance, bearing angle, and heading difference, which are kinematically coupled, so a single actuator fault propagates through all three Separation-Bearing Control channels at once. The nonlinear kinematics are linearized about the rigid-formation operating point; for a constant-speed reference, this operating point is stationary and the resulting fault-augmented error model is linear time-invariant, explicitly accounting for additive actuator faults and bounded external disturbances. A joint estimation observer reconstructs the unmeasured fault signal and filters the measured state with a guaranteed disturbance-attenuation level obtained from a single convex linear matrix inequality, and a state-feedback law with online fault compensation is derived. The linear closed loop is certified asymptotically stable with a prescribed H∞ attenuation bound, which establishes local asymptotic stability of the corresponding equilibrium of the nonlinear formation. Numerical simulations on a three-vehicle circular convoy, together with Monte Carlo, fault-variation, measurement-noise, model-mismatch, and trajectory-change studies, and comparisons against non-fault-tolerant, passive, and integral-action baselines, delineate the specific benefit of online fault reconstruction and active compensation.

Drones doi: 10.3390/drones10080578

Authors: Ahmad Irham Jambak Ismail Bayezit Mahmut Reyhanoglu

Autonomous sailing sea drones cannot sail directly into the wind. Headings inside the no-go zone produce too little aerodynamic force for propulsion. An escape heading that is geometrically safe can be unsailable for this reason, a complication that propeller-driven vessels never face. WF-RS, the Wind-Feasible Recovery Supervisor developed in this paper, computes tangent recovery headings around an inflated obstacle boundary, replaces wind-infeasible candidates, and holds one avoidance side through commitment and hysteresis until nominal guidance can resume. A reduced-order planar sailing model is used to compare WF-RS with nominal LOS/tacking guidance, three structural ablations, and a reactive potential-field baseline with no-go projection. Across a deterministic scenario set and 500 paired randomized obstacle encounters under goal-oriented target-completion scoring (each trial continued to goal reach), WF-RS reaches the highest clean target-completion rate, 96.0% (95% Wilson interval 93.9–97.4%), with a 4.0% obstacle-contact rate against 11.6–53.6% for the comparators, and every trial reaching the goal. Prolonged residence inside the no-go sector is logged as a diagnostic. The performance statistics come from simulation with a non-identified model. A processor-in-the-loop (PIL) test additionally verifies execution of the controller on a TI LAUNCHXL-F28379D C2000 target over a 115,200 baud serial link. These results are limited to simulation and PIL testing. Sea trials are future work.

Drones doi: 10.3390/drones10080577

Authors: Ruixin Fan Sai Huang

The rapid evolution toward sixth-generation (6G) wireless networks is reshaping the role of unmanned aerial vehicles (UAVs) in future communication systems [...]

Drones doi: 10.3390/drones10080575

Authors: Kunpeng Qi Chunhong Liu Zhihong Liu

Robust quadrotor trajectory tracking under wind disturbances is challenging because real outdoor wind is multimodal with time-varying and heavy-tailed characteristics, whereas existing solutions suffer from insufficient disturbance observability and poor out-of-distribution robustness. To this end, this paper presents a robust quadrotor trajectory-tracking method based on residual-aware deep reinforcement learning for multimodal wind disturbances. The proposed method augments a standard tracking policy with a compact online acceleration-residual feature, which provides disturbance-related information without requiring an explicit wind sensor, a full disturbance observer, or a long-history recurrent estimator. To reduce excessive dependence on wind-specific temporal patterns, a residual-input regularization term is introduced during policy optimization. In addition, a tail-risk-aware reward is designed to balance nominal tracking accuracy, control smoothness, and suppression of large tracking deviations. The proposed method is evaluated under in-distribution wind, held-out out-of-distribution wind, and measured real-wind disturbances. The results show that the proposed method achieves the most balanced robustness under multimodal wind conditions compared with baselines.

Drones doi: 10.3390/drones10080576

Authors: Ayman Massaoudi Walid Aydi

The integration of Unmanned Aerial Vehicle (UAV) swarms with Cell-Free massive Multiple Input Multiple Output (CF-mMIMO) networks offers promising prospects for large-scale crop monitoring in precision agriculture. CF-mMIMO provides macro-diversity and uniform channel quality across large agricultural fields. However, practical deployment demands jointly optimizing 3D trajectories and transmit power to maximize energy efficiency and field coverage simultaneously. This is challenging due to the limited battery capacity, mandatory return-to-depot constraints, and collision avoidance requirements. In this paper, we introduce a joint sensing–communication utility function that captures the trade-off between energy efficiency and field coverage completeness. To provide a scalable and distributed solution for rotary-wing UAV swarms, we develop a multi-agent deep reinforcement learning (MADRL) methodology based on the Multi-Agent Proximal Policy Optimization (MAPPO) approach. We adopt the Centralized Training with Decentralized Execution (CTDE) strategy, in which a CF-mMIMO central processing unit (CPU) serves as a global critic during training. At execution time, each UAV independently runs a lightweight local policy that adapts its trajectory and transmit power in real time based on battery state and air-to-ground channel variations. Simulation results reveal that the proposed MAPPO-CTDE approach outperforms existing benchmarks. Unlike prior methods that require instantaneous global CSI or neglect the sensing–communication coupling, the proposed approach simultaneously achieves high field coverage completeness, robust communication energy efficiency, and a high depot-return rate under hard battery constraints without any inter-UAV communication overhead at execution time.

Drones doi: 10.3390/drones10080574

Authors: Dipraj Debnath Fernando Vanegas Sebastien Boiteau Julian Galvez-Serna Juan Sandino Felipe Gonzalez

Multi-UAV missions for remote sensing and environmental monitoring under extreme conditions require task allocation and path optimisation to efficiently distribute goals across vehicles. These methods must also be executed reliably inside an autonomous robotics framework. Several methods for the multiple travelling salesman problem (mTSP) show robust offline routeing efficiency. However, system-level validation under realistic operational conditions including waypoint management and inter-UAV separation remains limited. This research transforms the previously proposed Distance Efficient Clustering Kmeans Genetic Algorithm (DECK_GA) from an offline model into a deployment-focused multi-UAV remote sensing framework implemented in ROS2, Aerostack2, and Gazebo. A uniform waypoint management interface integrates planning, Rviz visualisation, and autonomous execution. The system combines Dynamic Centroid Kmeans (DCKmeans) for spatially coherent waypoint allocation with a Distance Efficient Genetic Algorithm (DEGA) for individual UAV route optimisation. The evaluation is conducted in a high-fidelity Antarctic environment where waypoints represent survey desired objectives in moss regions, and altitude is managed using terrain-referenced control involving two to five UAVs and 30 to 120 waypoints. The framework was evaluated against two baselines under identical mission configurations, with 10 trial runs for each: a Traditional GA Divide & Conquer planner and a Classical Kmeans DEGA planner, which utilises the same route optimisation method and differentiates the outcomes of the allocation stage. DECK_GA showed reduced mean planned and executed distances compared to the Traditional GA Divide & Conquer baseline across all configurations, achieving planned distance reductions ranging from 15.99% to 75.36%. Additionally, it produced shorter path than Classical Kmeans DEGA in 14 out of 16 configurations. The average minimum inter-UAV separation was greater than the Traditional GA Divide & Conquer baseline in 15 of the 16 configurations and higher than Classical Kmeans DEGA in 14 of the 16, which demonstrates that the DCKmeans allocation improves spatial separation. This research focuses on the framework for planning to execution instead of the introduction of a new optimisation method, as DECK_GA was proposed in previous research and is now incorporated and tested within an autonomy framework. This evaluation is simulation only. Real world flying, hardware in the loop testing, wind, communication latency, and location error prediction tend to be future developments.

Drones doi: 10.3390/drones10080573

Authors: Chetan Yumnam Satya Prakash Kumar Bikram Jyoti Ramesh Kumar Sahni Manoj Kumar Karan Singh Kamal Nayan Agrawal Shishupal Pal Avinash Sahu Manish Kumar

Drone-based chemical spraying in agriculture faces challenges related to operator health hazards, spray deposition, application efficiency, and environmental safety. Electrostatic charging system is a novel technology that minimizes off-target spraying losses and increases droplet deposition on the plant canopy. In electrostatic spraying, chemical consumption and application rates are reduced due to the uniform distribution and enhanced deposition of charged droplets on plant surfaces, thereby improving spraying efficacy. In this study, a drone-based air-assisted electrostatic sprayer was developed to investigate the effect of operational parameters that include forward speed, discharge rate, applied voltage (charged condition) and propeller speed on the droplet density (drops/cm2) in a cotton crop under laboratory conditions. The charge-to-mass ratio (CMR) of the developed air-assisted electrostatic nozzle was found in the range between 1.8–2.5 mC/kg. An Artificial Neural Network–Ant Colony Optimization (ANN–ACO) method was used to optimize the operational parameters for obtaining the highest charged droplet density on the plant canopy surfaces. Results showed that the charged droplet density was significantly affected by discharge rate (DR) and applied voltage (AV) followed by forward speed (FS) and was slightly influenced by the propeller speed (PS). Optimal performance was achieved at FS = 2.58 km/h, PS = 1204 rpm, DR = 558.45 mL/min and AV = 6.18 kV under the charged conditions. At these optimized parameters, an average charged droplet density of 185.83 ± 4.25 drops/cm2 (mean ± SE) was achieved. For the uncharged conditions, the optimal performance was achieved at FS = 2.80 km/h, PS = 1065 rpm and DR = 556.75 mL/min, corresponding to an average droplet density of 108.49 ± 2.15 drops/cm2. The integration of the ANN–ACO optimization algorithm with the drone-based air-assisted electrostatic spraying system can enhance precision chemical application on the cotton, improving efficiency and sustainability.

Drones doi: 10.3390/drones10080571

Authors: Yihao Zhong Changsheng Yin Ruopeng Yang Yuantao Yang Yiwei Lu Yongqi Wen Yongqi Shi Bo Huang Yu Tao Jinyin Bai

The growing affordability, autonomy, and swarming of small unmanned aerial vehicles (UAVs) turn low-altitude defense from single-shot interception into a multi-node cooperative decision problem, in which the loss of sensing, coordination, or engagement nodes breaks the closed loops linking them. This study formulates their recovery as the dynamic reconfiguration of cooperative counter-UAV task chains. Given a pre-disturbance plan and a set of failed defending nodes, reconfiguration is modeled as a constrained bi-objective optimization balancing recovered engagement effectiveness against the change to the baseline plan and is solved by Multi-Agent Heuristic Evolution (MAHE), an automated heuristic design framework whose evolution, coordinator, repair, and reflection agents—driven by a large language model—evolve scoring heuristics for a fixed reconfiguration solver. Across instances of varying scale and under light-to-heavy node loss conditions, MAHE outperforms both a single-agent heuristic design counterpart and a range of hand-crafted solvers: on ten held-out test instances spanning 8–320 targets it attains the highest overall normalized hypervolume (0.947, versus 0.935 for the single-agent counterpart and 0.30–0.45 for the hand-crafted solvers) and the best mean rank (1.43 of six methods, p<10−5); the hand-crafted solvers lose most of their solution quality as the problem grows, whereas MAHE preserves it and sustains high recovery at a nearly constant reconfiguration cost. An ablation confirms that its agents contribute complementary gains. These simulation results indicate that automatically generated, reconfiguration-specific heuristics offer a scalable algorithmic foundation for dynamic, heterogeneous, and constraint-intensive counter-UAV task-chain reconfiguration.

Drones doi: 10.3390/drones10080572

Authors: Hongwei Han Guanghong Gong Ni Li

Coordinating multiple unmanned aerial vehicles (UAVs) for cooperative missions requires agents that perceive their environment, reason about objectives, and generate joint actions. Vision–language–action (VLA) models unify these capabilities but lack a principled source of multi-agent training data and suffer from a training–inference discrepancy in closed-loop control. We propose MiniUAV-VLA, a compact centralized VLA controller for simulated multi-UAV search-and-elimination based on multi-agent reinforcement learning (MARL) expert distillation. A QMIX expert policy achieving 100% mission success generates multimodal demonstrations pairing rendered tactical map images with structured textual state prompts. A 158 M-parameter VLA model with approximately 65 M trainable parameters in the MiniMind-3V backbone and vision projection is fine-tuned with a multi-agent discrete action head that jointly predicts actions for all UAVs in a single forward pass. We identify a training–inference feature mismatch in behavior cloning and address it via prompt-end action pooling, which extracts action-relevant hidden states at the user–prompt boundary rather than after the generated response. In closed-loop evaluation with four drones and six mobile targets averaged over five evaluation seeds, MiniUAV-VLA reaches 74.4 ± 4.6% mission success against 9.4 ± 2.1% for a random policy and 16.2 ± 3.2% for an observation-limited greedy baseline. Across five independent training runs, prompt-end action pooling improves mean closed-loop success from 40.6% to 76.2% over the last-token alternative. These results support MARL expert distillation as a data-efficient route to compact multi-agent VLA control in this simulated setting.

Drones doi: 10.3390/drones10080570

Authors: Hailu Xin Weidong Bao Hui Yan Ji Wang Xiaoqing Li Yanjie Song Lining Xing

Reliable data collection is essential for disaster-oriented Internet of Things (IoT) systems, where damaged terrestrial communication infrastructure often leaves sensed data buffered at disconnected end devices. In Unmanned Aerial Vehicle (UAV)-Internet of Things device (IoTD) collaborative data collection, random UAV faults and limited energy and buffer resources further complicate mission execution, making fault-tolerant scheduling crucial for robust data recovery. To address these issues, a unified framework is developed by integrating dynamic UAV reliability modeling, Maximum Distance Separable (MDS)-coded fault-tolerant backup, and collaborative scheduling optimization. Within this framework, a data fault-tolerance mechanism, termed MFTB, and a bilevel collaborative scheduling algorithm, termed LP-DCFS, are proposed. Simulation results indicate that, in the evaluated scenarios, the proposed methods achieve better overall performance than the considered baselines. In a representative high-load, high-failure scenario, MFTB reduces data loss by 4.8% and 37.5% compared with Buffer-Limited Retransmission (BLR) and Replication, respectively, while LP-DCFS increases the amount of recovered data by 33.9%, 32.3%, and 53.1% compared with ACEPSO, ADE-DMRM, and DQN, respectively. Under the modeled independent random crash and non-return faults and the evaluated simulation settings, these results suggest that coordinating failure-risk characterization, data-protection mechanisms, and task-scheduling strategies can improve the robustness and data-recovery capability of disaster-oriented UAV-assisted data collection.

Drones doi: 10.3390/drones10080569

Authors: Yuchao Liu Shuguo Xie Xiao Sun Qinglong Wu

The rapid proliferation of drones has raised growing concerns regarding low-altitude airspace safety. In practical UAV communication scenarios, relying solely on a single time-domain IQ signal is often insufficient to fully characterize UAV RF fingerprints. Moreover, increasingly subtle hardware differences among UAV transmitters make RF fingerprint identification more challenging. Other representations, such as STFT-based features, are commonly converted into image-like inputs for neural networks, increasing deployment complexity on edge devices. To address these challenges, this paper proposes a UAV recognition framework based on multi-domain signal representations. The proposed framework employs a multi-domain input strategy and structural reparameterization to reduce the number of parameters, computational cost, and deployment latency. Experiments under AWGN conditions demonstrate that the proposed model achieves superior recognition performance in both UAV classification and individual identification tasks. The proposed model is further deployed on the Ascend 910B platform to verify its deployment feasibility.

Drones doi: 10.3390/drones10080568

Authors: Sonia Mami Karem Chokmani Ridha Guebsi

Surface water monitoring often requires frequent and reliable sampling at locations that are difficult or unsafe to access using conventional methods. To address these constraints, unmanned aerial vehicles (UAVs) have increasingly been used as platforms for automated water collection. This paper presents a UAV-assisted water sampling system designed to provide a high degree of operational flexibility through control of sampling depth and collected volume. The proposed system can collect up to 3 L of water distributed across six individual containers, enabling a variety of sampling strategies, including discrete, composite, and multi-depth sampling. Sampling depths of up to 3.25 m can be achieved, supporting depth-resolved investigations such as stratified water-column analyses. Compared with existing UAV-based samplers, the proposed system combines configurable sampling depth, configurable sampling volume, and a multi-bottle architecture within a single platform. To support sample integrity, an automated cleaning sequence is executed before each collection step to reduce the risk of cross-contamination between samples. Experimental validation demonstrated repeatable depth deployment under controlled conditions and consistent volume-control performance, with relative errors generally within ±3% and a maximum observed deviation of 5%. In addition, a task-based analysis provided an initial characterization of the energy demand associated with the sampling process. Overall, the results indicate that the proposed architecture can support flexible and protocol-oriented water sampling operations while extending the sampling capabilities of existing UAV-based systems.

Drones doi: 10.3390/drones10080567

Authors: Nan Sun Xin Zhao Bing He Zixin Jiang

Cooperative jamming task allocation for UAV swarms must jointly consider radar priority, the effectiveness of different jamming modes, and heterogeneous resource limits. This paper establishes a radar threat assessment model from radar operating parameters and evaluates each candidate jamming assignment in the time, frequency, and power domains. Based on these evaluations, a binary integer programming model is formulated to jointly determine radar selection, UAV assignment, and jamming mode under coverage, capacity, and mode availability constraints. To solve the model, an improved multi-population genetic algorithm (IMPGA) is developed using crossover on radar task blocks, mutation at the task level, stochastic feasibility repair, and cooperative evolution among multiple subpopulations. Experimental comparisons are conducted under identical function evaluation budgets, and each representative scenario is evaluated through 100 independent Monte Carlo runs. The results show that the IMPGA reliably obtains exact or near-optimal solutions and provides improved solution quality and consistency, particularly as the problem scale and resource coupling increase. The scalability experiments further demonstrate that the algorithm maintains small optimality gaps in larger instances. Ablation results confirm that the radar task operators and stochastic repair make important contributions to the final solution quality and convergence process.

Drones doi: 10.3390/drones10080566

Authors: Yadong Wang Zhe Chen Wen Wang Ye Yang Yunfei Mu Peng Gu

Drones are increasingly being used for tasks such as power grid maintenance, routine inspections, and forest monitoring. These tasks often involve large areas that need to be surveyed, requiring drones to be recharged frequently. However, these areas, such as mountainous regions and forests, have rugged terrain and humid environments, making it unsuitable to place charging systems on the ground. Furthermore, misalignment can easily occur when drones are docked. Therefore, we propose an underground wireless charging system for drones with a frustum magnetic coupling structure to adapt to rugged terrain and enhance the wireless charging capability of drones. On the transmitting side, a frustum magnetic coupling structure based on a tap coil is proposed. To enhance the drone’s alignment capability, the receiving coil is designed as a circular planar coil that matches the frustum structure of the transmitting coil, achieving self-alignment through its physical structural characteristics. The parameters of the magnetic coupling structure are designed and optimized through finite element simulation analysis, and the output characteristics of the equivalent T/S type compensation network are analyzed. A physical prototype is built, achieving a maximum efficiency of 90.4%. The feasibility of the proposed system has been verified.

Drones doi: 10.3390/drones10080565

Authors: Xin Zhang Bing Li Changwei Mi Yangyang Zhao Junqi Dong Zhenyang Cui Zeyuan Zhao

Cooperative UAV missions often require multiple vehicles to reach a target area at nearly the same time. For small-turbofan UAVs flying along recorded routes, this timing task is affected by path-length differences, spatially varying wind, and phase-circle segments kept for waiting or coordination. Under these conditions, using only a fixed cruise setting often leaves limited room for arrival-time adjustment. This paper develops a wind-aware synchronized-arrival optimization method for small-turbofan UAVs based on segmented RPM scheduling and phase-circle retention. The method combines local wind projection along the path, an empirical RPM-to-airspeed model, and discrete flight-time integration, and it evaluates three timing-adjustment modes: fixed-alpha segmented RPM optimization, fixed-RPM phase-circle retention optimization, and their joint optimization. The method is tested on the recorded trajectories of UAV 13 and UAV 14, using a reconstructed wind field obtained from multi-UAV flight data. In the wind-aware constant-RPM baseline, both UAVs keep the full phase-circle segments and fly at 6500 rpm, which gives a predicted synchronization error of 1.192 s. Segmented RPM optimization reduces the error to 0.013 s. Phase-circle retention optimization reduces it to 0.007 s and shortens the total path length from 37,752.3 m to 26,103.4 m. The joint optimization gives the best result, with a synchronization error of 6.58×10−4 s and a total path length of 26,094.4 m. A no-wind ablation study also shows that ignoring wind leads to a total flight-time bias of about 29–31 s. These results indicate that phase-circle retention is effective for coarse flight-time adjustment through path shortening, whereas segmented RPM scheduling is better suited to fine timing correction. Using the two together improves synchronized-arrival prediction and adjustment for recorded UAV paths in wind.

Drones doi: 10.3390/drones10080562

Authors: Maxim Saveliev Vladyslav Shtefan Thomas B. Scott Viktor Grechaninov Oleksandr Mykhailov Anatolii Doroshenko Maksym Pantin

On 14 February 2025, the outer cladding of the Chornobyl New Safe Confinement (NSC) was damaged by an explosion caused by a one-way attack unmanned aerial vehicle (UAV), creating a hole of about 15 m in diameter and requiring about 300 penetrations to be made in the cladding during firefighting. This created an urgent need to assess radiation dose rates above damaged areas to support repair planning and worker radiation protection. This study presents a UAV-based survey of the equivalent gamma dose rate distribution above the damaged northern side of the NSC outer cladding. The survey used a bespoke system, integrating a multirotor UAV, an AccuRad Personal Radiation Detector (PRD), onboard data acquisition and transmission modules, and ground-based and server-side analytical components. Measurements were performed under real post-incident field conditions, including restricted flight zones, wind-induced turbulence, proximity to large metallic structures, and electronic warfare interference. The dataset was filtered for Global Positioning System (GPS) reliability, transformed into a metric coordinate system, and processed for spatial interpolation and mapping. The resulting distribution showed a spatially non-uniform radiation field: the main damage zone had relatively low equivalent gamma dose rates, whereas the highest values, up to 1092 μSv/h, were recorded over areas of the NSC closest to the Shelter Object. The study demonstrates UAV-based radiation mapping of a damaged large-scale confinement structure and provides data supporting Chornobyl Nuclear Power Plant repair planning.

Drones doi: 10.3390/drones10080564

Authors: Xirui Xu Dongxu Yang Wei Fan Weimin Quan Ruiliang Fan Fei Wang

Monitoring the spatial distribution of oyster reefs in an accurate, efficient, and flexible manner is crucial for assessing changes in nearshore fishery habitat environments. However, traditional optical remote sensing methods are often affected by illumination variation, tidal fluctuation, and spectral confusion in complex intertidal environments. An automated extraction framework based on the SAM (Segment Anything Model) using UAV (Unmanned Aerial Vehicle) DOM (Digital Orthophoto Map) imagery is proposed to achieve high-precision oyster reef identification and area estimation. Multi-resolution UAV imagery was processed, and key SAM parameters were systematically optimized under different illumination conditions. The results show that spatial resolution significantly influences segmentation accuracy, and appropriate resolution selection improves both stability and reliability. Based on UAV data acquired in 2025, the total oyster reef area in the study region was estimated to be 2.24 ha. After parameter optimization, segmentation accuracy improved from 93.22% to 97.61% in illuminated areas and from 93.30% to 96.67% in shaded areas. The proposed method demonstrates strong robustness under varying environmental conditions and effectively enhances boundary detection accuracy. A scalable and reliable approach is provided for coastal habitat monitoring and offers new insights into automated object extraction in complex remote sensing imagery.

Drones doi: 10.3390/drones10080563

Authors: Luthfi Fikri Baskoro Nurdianti Rizki Hapsari Puguh Triwinanto Erinna Dyah Atsari Adi Maimun

This study presents an experimentally evaluated reduced-order maneuvering framework for a trimaran Autonomous Underwater Vehicle (AUV) by integrating CFD-derived hydrodynamic coefficients with experimentally characterized thruster forces for low-speed pivot turn prediction. Unlike previous studies that primarily focused on hydrodynamic coefficient identification or computational fluid dynamics (CFD)-based hydrodynamic analyses, the proposed approach evaluates the capability of independently derived hydrodynamic parameters to reproduce experimentally observed maneuvering behavior within a computationally efficient three-degrees-of-freedom (3DOF) dynamic model. The maneuvering formulation incorporates nonlinear hydrodynamic derivatives adopted from a previously published virtual Planar Motion Mechanism (PMM) investigation together with experimentally measured thrust characteristics. Numerical simulations were performed in MATLAB using a planar 3DOF maneuvering model, while an experimental evaluation was conducted through a representative pivot turn maneuver in a controlled pool environment using video-based trajectory tracking. The simulated and experimental trajectories exhibited consistent turning behavior, yielding Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) values of 0.0992 m and 0.0905 m, respectively. Although discrepancies were observed in the heading response during the later stages of the maneuver owing to simplified hydrodynamic assumptions and unmodeled nonlinear effects, the proposed framework successfully reproduced the dominant trajectory and yaw characteristics of the investigated maneuver. These results indicate that the proposed reduced-order framework was able to reproduce the investigated low-speed pivot turn maneuver under the validated operating conditions, while broader experimental validation remains necessary to establish its general applicability.

Drones doi: 10.3390/drones10080561

Authors: Mosab Alrashed Humoud Aldaihani Mohammad Alqattan

Long-endurance unmanned aerial vehicle (UAV) missions beyond 72 consecutive flight hours face a reliability boundary at which thermal gyroscope drift drives the inertial navigation system (INS) position error into rapid nonlinear divergence at a predictable threshold tc. This paper presents BAZ II, a simulation-validated multi-agent navigation system that extends the analytical BAZ (bifurcation-aware zonal navigation) framework. Its central idea is to treat communication quality as a planning resource and combine it with multi-agent collaboration, making the navigation cliff a manageable degradation event rather than a hard operating limit. Four contributions support this idea: a thermalhysteresis MEMS gyroscope drift model reproduces the analytical cliff in simulation and supplies its physical mechanism; a distributed collaborative simultaneous localization and mapping (SLAM) filter coupled to a stochastic continuous-time Markov chain (CTMC) interagent channel sustains GPS-denied localization within the operational accuracy budget; a 3D Gaussian process RF-aware model predictive controller (MPC) with cognitive radio frequency-hopping restores link availability under jamming, while an analytic hierarchy process (AHP)-weighted multi-objective communication cost improves latency and jitter at negligible signal-to-noise ratio cost; finally, the integrated controller executes within the onboard real-time budget of an NVIDIA Jetson Xavier NX. All results are obtained in simulation, with hardware-in-the-loop and field testing remaining as priority future work.

Drones doi: 10.3390/drones10080560

Authors: Lijia Lin Xiaopei Chen Wenhao Wu Zhijian Lin

Bolstered by emerging sixth-generation (6G) communication technology, space–air–ground– sea integrated networks (SAGSINs) are reshaping edge computing through the synergistic use of space, aerial, terrestrial, and maritime platforms. However, in such highly dynamic and heterogeneous network environments, long-term energy-efficient computation offloading in drone-enabled SAGSINs has not yet been thoroughly explored, particularly when dynamic task demands from user equipment (UE) are served under the constraints of energy-limited drones. To fill this gap, the problem of service node association and computing-frequency allocation under dynamic computation offloading demands at the edge of the networks is studied in this paper. However, the related problem turns out to be a stochastic optimization problem. To this end, through in-depth mathematical analysis based on the Lyapunov optimization method, it is found that the multi-time-slot long-term optimization problem can be transformed into several single-time-slot optimization problems, which enables an efficient solution to the original problem. The single time-slot problem is solved using graph theory, convex optimization, and optimization theory, where Lyapunov optimization is utilized to achieve queue stability and energy efficiency. Simulation results demonstrate that the proposed strategy effectively reduces energy consumption, ensures low delay, and maintains long-term queue stability in drone-enabled SAGSINs under dynamic task demands from UE.

Drones doi: 10.3390/drones10080559

Authors: Ioana-Raluca Adochiei Daniel Andrei Avram Felix-Constantin Adochiei

Autonomous takeoff and landing (ATOL) remains one of the most challenging tasks for fixed-wing unmanned aerial vehicles (UAVs), particularly in environments where Global Navigation Satellite System (GNSS) signals are degraded or unavailable. This paper presents an LiDAR-assisted terrain-relative navigation framework for the autonomous takeoff and landing of a 5 kg fixed-wing UAV operating under degraded navigation conditions. The proposed architecture integrates a downward-facing LiDAR rangefinder with barometric altitude sensing, INS/GNSS navigation, optical-flow measurements, and airspeed information within a multi-sensor fusion and flight-control framework. The system combines a Pixhawk-based autopilot with a companion-computer architecture responsible for real-time sensor processing, altitude estimation, mission supervision, and MAVLink-based communication. A dedicated filtering strategy and sensor fusion approach enable reliable terrain-relative altitude estimation during critical low-altitude flight phases, while fault-tolerant command-management mechanisms improve operational robustness in the presence of temporary communication losses and sensor disturbances. The proposed framework was validated through Software-in-the-Loop (SITL), Hardware-in-the-Loop (HITL), and real-flight experiments. Experimental results demonstrated stable and repeatable autonomous landing performance. Comparative analyses showed that the LiDAR sensor provided the most accurate and responsive terrain-relative altitude measurements during takeoff, flare, and landing operations, particularly over irregular and vegetation-covered surfaces. In contrast, barometric sensing provided greater long-term stability during cruise flight, highlighting the importance of multi-sensor fusion for reliable altitude estimation throughout the mission profile. The results confirm that LiDAR-based terrain-relative sensing significantly improves autonomous takeoff and landing performance for fixed-wing UAVs operating in GNSS-degraded environments. The proposed architecture offers a practical and low-cost solution for the autonomous takeoff and landing of fixed-wing UAVs operating in GNSS-degraded environments while demonstrating the benefits of integrating LiDAR, inertial, barometric, and GNSS measurements within a unified multi-sensor autonomous flight framework.

Drones doi: 10.3390/drones10080557

Authors: Harrison Dean Srikanth Bashetty Hana Forrester Juan Bernal Palacios Tristen Portis

Unmanned Aerial Vehicles (UAVs) are seeing increased use in agricultural settings due to their potential to be integrated with systems for applying pesticides. They can target specific areas while offering the potential to reduce chemical waste and improve application efficiency. However, this means that spray deposition efficiency is strongly influenced by rotor-induced downwash, which affects droplet transport, drift, and uniformity. This study presents a combined computational and experimental investigation of downwash–spray interactions in a hexacopter platform. CFD is used to predict the performance of various sprayer configurations that differ in the number, spacing, and positioning of nozzles. Rotor-induced airflow is modeled using an actuator disk approach in ANSYS Fluent 2025, and spray behavior is predicted using the Discrete Phase Model. Pure water was used as the working fluid for both the CFD simulations and experimental validation to ensure consistency between numerical and physical testing conditions. Numerical results indicate that a two-nozzle under-rotor setup maximizes performance characteristics such as deposition area, density, and uniformity for the designed agricultural UAV, providing a theoretically effective deposition area of 9.375 m2, an effective application rate of 0.03387 mL/m2, and a coefficient of variation of 45.3%. Compared to the best-performing boom configuration, this represents an approximately 13.5% improvement in spray uniformity. These results are validated through experimental testing using a modular UAV sprayer system and deposition measurements obtained from water-sensitive paper in controlled indoor conditions, achieving a droplet size of 502 µm, swath width of 1.8 m, 0.8% area coverage, and a coefficient of variation of 36.5%. While differences were observed between predicted and measured droplet size distributions, the CFD and experimental results demonstrated similar trends in spray coverage and deposition uniformity. Future work will refine simulations to better match experimental conditions and investigate canopy interaction, crosswind effects, and field-scale performance.

Drones doi: 10.3390/drones10080558

Authors: Fangqi Zhang Yi Hu Qiang Wang Yuanjing Ma

UAV path planning in complex urban environments faces significant challenges due to dense obstacles, narrow corridors, and stringent safety requirements. To address these issues, this paper proposes LLSOA, a modified Seagull Optimization Algorithm that integrates Latin Hypercube Sampling (LHS) for population initialization and Lévy Flight for global search. The key innovation lies in the problem-driven design: LHS ensures uniform coverage in dense urban maps, while Lévy Flight helps escape local optima. Compared with four state-of-the-art swarm intelligence algorithms (DBO, GWO, PIO, and PSO) across four urban scenarios, LLSOA achieves the best comprehensive fitness. Considering multiple constraints including path length, curvature, collision avoidance, and obstacle-avoidance logic, the trajectories generated by LLSOA show competitive overall performance, with no unsafe points recorded in the test scenarios and the best fitness values among the compared algorithms, albeit with a slight trade-off in path length. High-fidelity AirSim simulations with GPS/IMU noise further demonstrate that the planned trajectories remain within engineering acceptable limits. Compared with the noise-free baseline, the maximum trajectory deviation increases by 2.3% and the average deviation increases by 2.8% under high GPS/IMU noise. The main contributions are: (1) a problem-driven LLSOA that combines LHS and Lévy Flight, specifically tailored to dense urban environments; (2) theoretical analysis and simulation verification demonstrating its feasibility for multi-constraint path planning under the tested conditions; (3) high-fidelity (UE+AirSim) validation showing that the generated trajectories retain stability even under realistic sensor noise.

Drones doi: 10.3390/drones10070556

Authors: Kim I. Monteforte Paul A. Butcher Brendan P. Kelaher

Unprovoked shark bites are one of the most recognised human–wildlife conflicts and present a significant concern for beach safety. Lethal methods of shark mitigation have previously been implemented to reduce the risk of such incidents; however, due to their destructive impacts on vulnerable marine wildlife, non-lethal approaches are increasingly preferred. In recent years, drones have emerged as an effective, minimally invasive tool for real-time shark surveillance in surf zones. Drones are also used to collect valuable data on shark ecology and behaviour in nearshore environments, which can inform evidence-based policies. This review examines the utility of drones for shark surveillance programs by identifying key operational parameters and associated challenges of drone-based methods. We investigate emerging technologies, including long-range drones, remotely operated or autonomous flight missions, and the use of artificial intelligence for shark detection and species identification. We also outline current drone licensing, laws, and regulations, noting that these vary across administrative regions (i.e., countries and states). Overall, this review provides insight into the expansion of drone-based shark surveillance in nearshore areas and its potential to enhance beach safety, support management decisions, and advance scientific knowledge without negatively impacting shark populations.

Drones doi: 10.3390/drones10070555

Authors: Jan M. Kelner Sławomir Klimaszewski Sławomir Brzózka Wojciech Stecz

The paper presents a certification-oriented methodology for requirements management and model verification in the design of unmanned aerial vehicle (UAV) systems. The proposed approach addresses challenges in certifying safety-critical UAV platforms whose architectures include custom-developed hardware and software components. The methodology integrates requirements engineering, functional hazard analysis (FHA), and formal model verification into a unified design and validation workflow. Methods for categorizing, prioritizing, and decomposing system requirements are presented to support the development of reliable UAV software and operational procedures. A systematic approach for mapping certification requirements and FHA safety functions to UAV operational scenarios and system use cases is introduced, enabling traceability between safety requirements, system behavior, and verification artifacts. The proposed framework employs Unified Modeling Language (UML) state-machine models, validated using linear temporal logic (LTL) and computation tree logic (CTL). Model verification is performed using the NuSMV symbolic model checker to assess the correctness, completeness, consistency, and safety properties of operational scenarios. A practical case study concerning UAV handover between ground control stations (GCSs) demonstrates the applicability of the proposed method. The analysis identifies inconsistencies in the operational model and shows how formal verification supports the early detection of unsafe or incomplete system behaviors. The presented approach supports the development of certification-ready UAV systems by improving requirements traceability, reducing verification ambiguities, and facilitating the validation of safety-critical flight-control procedures.

Drones doi: 10.3390/drones10070554

Authors: Ziping Wang Henan Zhu Kofi Nyarko Xiaozheng He

While drone delivery has gained significant scholarly and industrial interest, rural residents’ acceptance of these systems remains underexplored, despite the region’s acute logistics challenges and service inequities. Using survey data from rural U.S. residents, this study first estimates an ordered logistic regression (OLR) model to identify factors associated with five-level drone delivery acceptance. The study then compares OLR, multinomial logistic regression (MNL), Random Forest (RF), XGBoost, and LightGBM under matched feature sets to evaluate whether nonlinear machine-learning models improve prediction beyond the interpretable statistical baseline. Open-ended responses are coded into LLM-derived sentiment labels and added as supplementary predictors to test whether unstructured feedback improves acceptance prediction. Results show that willingness to pay is the strongest predictor of acceptance, while equitable same-day delivery demand and post-pandemic attitude adjustment are also positively associated with higher acceptance. Household disability status and urban accessibility are not significant after adjustment. In the five-level analysis, OLR provides a strong ordinal baseline, while XGBoost and other tree-based models improve selected class-level prediction metrics. In the binary high-acceptance analysis, machine-learning models show stronger predictive performance, especially when structured predictors are combined with sentiment features. This study contributes to rural drone-delivery literature by linking service equity, perceived value, and LLM-derived sentiment within a comparable statistical and machine-learning framework for rural service design.

Drones doi: 10.3390/drones10070553

Authors: Wenqin Han Shuangxi Liu Xianyu Wu Wei Zhao

High-speed Unmanned Aerial Vehicles (UAVs), characterized by high velocity and maneuverability, represent critical strategic threats within the aerospace security domain. Accurate trajectory prediction is a fundamental prerequisite for effective early warning and decision-making in defense. Despite increasing research in the domain, model-, data-, and hybrid-driven methods have not yet been comprehensively examined under a unified framework, limiting the understanding of their relative strengths and applicability. To address this gap, this paper systematically reviews the evolution and current technical status of these paradigms, categorized by the core logic underlying prediction methods. The principles and applicability limits of physical process modeling and state estimation algorithms are analyzed, along with emerging applications of machine learning and deep learning for trajectory feature extraction and pattern recognition. State-of-the-art architectures involving the integration of physical constraints and data-driven learning are discussed. Standard evaluation metrics are introduced to facilitate performance benchmarking of existing methods. Comparative analysis reveals that no single technical route can fully address the coupled challenges of uncertainty, accuracy, and real-time performance, underscoring that hybrid frameworks are essential for balancing these competing requirements. Lastly, key challenges are summarized, and future research directions are outlined to advance trajectory prediction methodologies. The provided insights can inform method selection and promote the development of high-accuracy prediction systems.

Drones doi: 10.3390/drones10070552

Authors: Zeyu Li Hui Li Chenhong Xiangli Yuanyuan Shen Yi Yu Fei Li

UAV bridge inspection requires centimeter-level global positioning in the engineering coordinate frame to associate detected defects with BIM component IDs and mileage stakes. Intermittent GNSS outages beneath beams, inside box girders, and in pier-dense regions cause conventional navigation methods to accumulate drift or produce discontinuous pose estimates. This paper presents BCEF-Nav, a BIM-constrained elastic fusion navigation framework for UAV bridge inspection. The framework adaptively adjusts GNSS constraints according to signal availability and replaces degraded GNSS references with semantic–geometric constraints from a high-precision as-built BIM model within a sliding-window optimizer. Semantic-guided feature matching suppresses false correspondences in repetitive bridge structures. BCEF-Nav uses only low-cost commercial sensors and is compatible with mainstream inspection UAVs. Digital twin simulations and field experiments achieved an ATE RMSE of 5.7 cm after 120 s of complete GNSS occlusion and a 100% positioning success rate with positioning errors below 10 cm, outperforming seven recent state-of-the-art baselines.

Drones doi: 10.3390/drones10070551

Authors: Yigit Mahmutoglu

Accurate underwater acoustic localization is a key enabling capability for autonomous underwater vehicles and underwater drones, which cannot rely on satellite positioning while submerged and therefore depend on acoustic methods to determine their position. Localization based on time-of-arrival (TOA) measurements requires precise time synchronization between the source and the receivers, which is difficult to maintain in practical deployments. The time-difference-of-arrival (TDOA) representation removes this requirement but discards part of the absolute timing information, reducing localization accuracy. This study investigates a physics-based feedforward multilayer perceptron (FF-MLP) framework for two-dimensional range–depth underwater localization that learns directly from the arrival-time structure induced by sound-speed variability and multipath, with the receiver-array geometry treated as a central design variable for improving synchronization-free TDOA localization. Using multi-receiver arrival times generated with the BELLHOP beam-tracing model under a representative Mediterranean underwater environment, synchronous TOA, biased TOA, and TDOA measurement representations are compared on a common footing, and the effects of the receiver depth distribution, the number of receivers, and the reference-receiver position are systematically examined through Monte Carlo evaluation. The results show that the receiver-array geometry, rather than the measurement representation alone, is decisive for TDOA-based localization: with an appropriately designed geometry, synchronization-free TDOA localization achieves a median two-dimensional RMSE of 11.28 m, approaching the accuracy attainable with synchronous TOA, which requires precise time synchronization. These findings indicate that careful receiver-geometry design can make synchronization-free TDOA a practical alternative to synchronous TOA for the acoustic localization of autonomous underwater vehicles.

Drones doi: 10.3390/drones10070550

Authors: Mei You Huihui Xu Zhangsong Shi Xiaopeng Bao Chengfei Wang Hao Wu

To address the fundamental trade-off between real-time responsiveness to high-priority missions and long-term overall economic efficiency of the system in multi-UAV dynamic task assignment, we propose a hybrid intelligent scheduling algorithm abbreviated as DPSO-Greedy. The algorithm performs periodic global batch optimization for regular orders using an improved discrete particle swarm optimization (DPSO) method, and realizes instantaneous allocation of emergency orders via an adaptive multi-factor Greedy strategy, thus enabling efficient collaborative processing of differentiated tasks. Targeting the trade-off between real-time response and long-term system efficiency, this paper proposes a hybrid DPSO-Greedy algorithm with decoupled task scheduling mechanisms. Comparative simulation results demonstrate that compared with mainstream metaheuristic algorithms (Greedy, SSA, GWO and RHS), the proposed method reduces the average response time of emergency orders by 33.2–68.2%, achieves an emergency order completion rate exceeding 90%, and improves system load balancing performance by 24–35% in dynamic scenarios characterized by burst and tidal demands. This study provides a promising solution for dynamic UAV assignment problems and offers valuable insights for a broader range of real-time resource collaborative decision-making applications.

Drones doi: 10.3390/drones10070549

Authors: Mehmet Konar Seda Arık Hatipoğlu İsmail Erol Ömer Çam Sümeyra Tuna Mustafa Fenerci

Battery state estimation is the main safety constraint for electric rotary-wing unmanned aerial vehicles (UAVs): mission decisions depend on both the instantaneous State of Charge (SOC) and the Remaining Useful Life (RUL). The present study compares seven machine learning and deep learning models (LR, SVM, k-NN, GBT, EL, LSTM, and a simplified RWKV) on real flight data from a rotary-wing helicopter testbed with a Pixhawk autopilot and an NVIDIA Jetson Nano mission computer. The dataset has 1310 samples (∼262 s) of nine on-board sensor signals. Mission-based RUL is defined as the projected time until SOC reaches a 20% safe-landing threshold. All models use an 80/20 random split, five regression metrics (RMSE, MAE, R2, MSE, PRMSE), and five random seeds. GBT wins on SOC with R2=0.9943±0.0014, MAE =0.25%, and 3.8μs per-sample inference on a workstation CPU; this latency leaves headroom for on-board mission planning. Battery temperature and voltage together carry over 90% of the predictive signal. GBT wins again on RUL (R2=0.596±0.042, MAE =583 s). The same ordering (tree ensemble ≻ recurrent ≻ linear) holds for both tasks; the remaining RUL gap reflects the single-flight dataset. The SOC labels originate from the on-board autopilot’s Coulomb-counting-based fuel-gauge estimator, so the SOC numbers should be read as a reproduction of that on-board trace at sub-microsecond inference latency rather than as independent accuracy; the calibration-free Coulomb-counting baseline reaches a marginally higher R2 (0.9950, MAE =0.35%) on the same task.

Drones doi: 10.3390/drones10070548

Authors: Rinkuben Patel Areej Salaymeh

Coordinating multiple unmanned aerial vehicles (UAVs) at scale remains challenging through centralized control or fixed rule sets, particularly when vehicles must operate under intermittent communication links, incomplete observability, and constrained onboard computational resources. Graph Neural Networks (GNNs) have emerged as a promising framework for addressing these challenges; however, existing surveys do not systematically relate GNN architectural decisions to the operational constraints imposed by UAV platforms during deployment. This survey reviews 196 scholarly studies published between 1987 and 2026 to develop such a framework. A GNN architecture and deployment taxonomy is organized into six major categories—Convolutional, Attentional, Sampling-Based, Spatio-Temporal, Distributed, and Resource-Efficient—each examined through dedicated architectural subsections and evaluated in the context of UAV system constraints. Four primary application domains are examined: multi-UAV trajectory planning, cooperative target tracking, communication-aware network optimization in Flying Ad Hoc Network (FANET) environments, and spatio-temporal airspace traffic prediction. Within these domains, the analysis highlights how architectural choices influence scalability, adaptability to dynamic conditions, and computational efficiency. Several deployment challenges consistently emerge, including maintaining tractable inference as swarm size increases, adapting graph representations under high mobility, and operating within the limitations of onboard computational resources. Based on these findings, a set of architecture-selection guidelines is derived to support deployment under varying operational conditions. Emerging research directions are also discussed, particularly the integration of GNNs with reinforcement learning, federated edge computing, and next-generation wireless communication systems. Overall, this survey bridges the gap between methodological development and practical deployment, providing a structured foundation for evaluating GNN suitability in real-world multi-UAV environments.

Drones doi: 10.3390/drones10070547

Authors: Ali Güneş

UAV-based weed detection in precision agriculture is constrained by asymmetric error costs: a missed weed patch causes herbicide under-treatment and yield loss, whereas a false alarm only prompts an unnecessary spot treatment. Cost-sensitive active learning (cAL) addresses this through an asymmetric misclassification penalty r+, but the optimal value is scene-dependent and cannot be determined without domain expertise or advance knowledge of scene difficulty—a fundamental barrier to autonomous UAV monitoring workflows. We propose AdaRisk-Agent, the first LLM-orchestrated framework for adaptive r+ calibration in cAL-based UAV weed detection. We validate the framework on four UAV multispectral scenes from two public datasets—WeedsGalore (Germany, five-band maize) and WeedyRice (Vietnam, four-band paddy)—spanning two crop types, two sensor configurations, and weed prevalence from 3.1% to 30.5%. Adaptive calibration reduces the false-negative rate (FNR) by up to 80% relative to symmetric-cost baselines across all scenes. The deterministic surrogate (AdaRisk-Rule) surpasses the fixed-policy oracle (cAL r+=7) on two of four scenes without advance scene knowledge, achieving a 50% FNR reduction on the most spectrally challenging scene. A context-feature ablation confirms that budget urgency is the primary calibration signal and that test-set-independent deployment is feasible. Each calibration decision is accompanied by a natural-language justification, enabling auditable deployment in operational precision agriculture workflows. Future work will extend AdaRisk-Agent to multi-class weed species detection and multi-scene meta-learning for compact offline surrogate policies.

Drones doi: 10.3390/drones10070546

Authors: Zhi Liu Yong Xian Shaopeng Li Ming Wang Liying Qian

Rapid preliminary reconnaissance of Critical Transportation Hubs (CTHs) by UAV swarms is vital during post-disaster rescue operations. However, limited camera fields of view, large heading variances, and GNSS multipath errors near massive steel-concrete structures complicate multi-view cooperative perception. This paper introduces a discrete, vision-based cooperative perception framework utilizing a decentralized anchor-wingman architecture. The pipeline integrates a Prob-IoU-optimized YOLO26m-OBB detector to extract oriented infrastructure footprints. To handle severe rotational discrepancies without IMU priors, a global scene registration cascade—combining SuperPoint and an Optimal Transport-driven LightGlue—is employed to establish robust geometric correspondences. Furthermore, a Projected Polygon Intersection over Union (Proj-IoU) mechanism, coupled with an RMSE-weighted spatial fusion strategy, dynamically associates and deduplicates overlapping targets across distributed views. Experimental results indicate that the framework achieves a low pixel-level RMSE of 2.12 pixels on the source domain and maintains a highly stable 2.36 pixels during zero-shot cross-domain testing (SUES-200 dataset), successfully resolving extreme heading variances up to 270°. The Proj-IoU mechanism resolves multi-source redundancies—collapsing overlapping projections by over 50%—bounding the localization error to approximately 1.06 m. Operating at 6.7 FPS on edge hardware via low-bandwidth tensor transmission, this system provides a rigorous geometric foundation for autonomous swarms, enabling downstream collision-free trajectory planning and Multi-Target Task Allocation (MTTA) in GNSS-denied environments.

Drones doi: 10.3390/drones10070545

Authors: Mohamed Fawzy El-Khatib M. Abdelfattah Mohamed M. El-Sotouhy S. Shaaban A. Abdellatif

Accurate trajectory tracking of quadrotor unmanned aerial vehicles (UAVs) remains a very challenging problem because of their inherent nonlinear, strongly coupled and underactuated dynamics. In order to overcome these limitations, a real-time Model Reference Adaptive Control (MRAC) strategy is proposed in this paper for better tracking performance in the presence of parametric uncertainties and external disturbances. The controller is cascaded, and adaptive laws based on Lyapunov stability theory are used to control the translational and rotational motions separately and guarantee closed-loop stability. The proposed approach is benchmarked against a tuned Particle Swarm Optimisation (PSO) PID controller under the same operating conditions to evaluate its efficacy. The validation is performed via extensive numerical simulations and real-time Hardware-in-the-Loop (HIL) experiments on an OPAL-RT platform, confirming enhanced disturbance rejection and transient response in the studied deterministic HIL conditions. The results show that the MRAC controller converges faster and has higher tracking accuracy than the PSO-based PID controller. Settling times are reduced from 9–12 s to 5–7 s with negligible steady-state error in setpoint tracking tests. The tracking errors for the multi-axis trajectory-tracking experiments, including the square and three-dimensional trajectories, are kept within 0.1–0.3 m; larger tracking deviations are observed with the benchmark controller. The quantitative performance evaluation demonstrates approximately 60–70% reduction in RMSE together with lower MAE, IAE, and ITAE values compared with the optimised PSO-based PID controller. Also, disturbance experiments under 1 N external force demonstrate the improved disturbance-rejection performance of the adaptive controller with performance degradation of about 25–30% compared to 38–45% for the PSO-based PID controller. The overall results obtained under deterministic real-time HIL conditions indicate that the proposed MRAC strategy provides improved trajectory-tracking performance compared to the benchmark PSO-based PID controller. Further statistical validation and physical flight experiments will be considered in future works.

Drones doi: 10.3390/drones10070544

Authors: Chaofeng Wang Longfei Zhang Jie Luo Shengming Dai

When natural disasters strike, the destruction of terrestrial communication infrastructure creates urgent demands for emergency networks. Efficient UAV deployment in capsule airport–UAV hierarchical networks has emerged as a critical challenge due to limited aerial resources and stringent quality-of-service requirements. This paper develops a QoS-aware joint optimization model for UAV deployment, integrating air-to-ground (A2G) channel modeling with resource allocation, where upper-level position optimization is coordinated with lower-level frequency allocation and power control through a hierarchical decomposition strategy. The proposed QoS-TLK-VNS-K algorithm combines graph coloring for interference mitigation with iterative power control for SINR guarantee. Empirical evaluation using multi-scenario simulations demonstrates that the proposed approach significantly outperforms the traditional distance-based coverage method. Statistical validation over 30 independent runs demonstrates significant improvements in QoS satisfaction (+23.8%, p<0.001), average SINR (+104.0%, p<0.001), minimum user rate (+194.9%, p<0.001), and Jain’s fairness index (+16.2%, p<0.001) compared to the distance-based baseline. These results demonstrate that the framework effectively addresses the trade-off between interference suppression and network connectivity in multi-UAV emergency communication systems.

Drones doi: 10.3390/drones10070543

Authors: Emilia Georgiana Prisăcariu Raluca Andreea Roșu Oana Dumitrescu Romeo Robert Ciobanu

The rapid expansion of Unmanned Aerial Vehicle (UAV) applications in both civilian and military sectors has intensified the demand for propulsion systems capable of delivering higher speed, increased endurance, and improved payload capacity. While battery-electric propulsion remains dominant for small UAV platforms, its limited energy density restricts operational range and mission flexibility. As a result, micro gas engines have emerged as a viable alternative for applications requiring high power-to-weight ratios and sustained high-speed operation. This review examines the fundamentals, scaling effects, and classification of micro gas turbine propulsion systems used in UAV applications, with emphasis on micro turbojets and related hybrid configurations. The paper discusses the thermodynamic principles governing micro gas engines and analyzes the aerodynamic, thermal, and combustion challenges associated with miniaturization, including low Reynolds number effects, tip leakage losses, thermal management limitations, and combustion instability. Furthermore, the study reviews the operational characteristics and mission suitability of different propulsion architectures for reconnaissance UAVs, high-speed UAVs, including reconnaissance and loitering platforms, target drones, and hybrid-electric aerial platforms. Recent developments involving additive manufacturing, advanced control systems, recuperated cycles, and hybrid-electric integration are also evaluated as enabling technologies for next-generation UAV propulsion. The findings demonstrate that although micro gas turbines continue to face important efficiency and manufacturing challenges at reduced scales, they remain essential for mission profiles that exceed the capabilities of purely electric propulsion systems.

Drones doi: 10.3390/drones10070542

Authors: Ziqi Geng Haixin Tian Ye Jin Jianli Song

The application of unmanned aerial spraying systems (UASS) in hilly orchards is challenged by terrain-induced airflow variability, which affects droplet transport and deposition. This study investigated the effects of terrain slope and flight patterns on rotor downwash airflow and droplet deposition using airflow measurements, computational fluid dynamics (CFD) simulations, and field experiments. Adjustable slope platforms (0°, 10°, 20°, and 30°) were used to characterize airflow behavior, while droplet deposition was evaluated under flat, uphill, downhill, and contour-parallel flight conditions, both outside and within citrus canopies. Results showed that increasing slope transformed the downwash airflow from an axisymmetric structure to a downslope-biased asymmetric pattern. Flight patterns significantly influenced deposition distribution. Uphill flight enhanced deposition in upslope and upper-canopy regions, whereas downhill flight increased deposition in rear and lower-canopy regions due to stronger recirculation. During contour-parallel flight, airflow shifted downslope, resulting in higher deposition on the downslope side of the canopy. Under the experimental conditions investigated in this study, the effective spray swath width during single-flight-line operations perpendicular to the contour lines (uphill and downhill flights) was approximately equivalent to the width of one individual tree canopy, whereas contour-parallel flight resulted in a narrower effective spray swath width due to terrain-induced airflow redistribution. An upslope route offset of 0.3–0.5 m improved droplet deposition uniformity between the upslope and downslope canopy regions. These findings provide guidance for optimizing UASS spraying strategies by adjusting flight trajectories, route offsets, and operational parameters according to terrain slope and canopy position.

Drones doi: 10.3390/drones10070541

Authors: Yang Zhao Mingying Huo Naiming Qi Liguang Wang Zongquan Xia Bo Qi Ge Yang

Challenges in trajectory planning are encountered by fixed-wing unmanned aerial vehicle (UAV) swarms operating in environments with unknown obstacles. In this study, a distributed real-time trajectory-planning method that integrates a distributed model predictive control (DMPC) framework with an adaptive Gaussian collocation strategy (DA-GCMPC) was developed. This method leverages a distributed iterative computational framework based on DMPC to reformulate trajectory planning as an optimal control problem. To address the fixed-resolution limitation of conventional distributed MPC formulations, a complexity-aware adaptive collocation mechanism is introduced. The novelty of the method lies in adapting the collocation transcription resolution of each local MPC problem according to the instantaneous planning complexity. This mechanism selects the collocation type online according to maneuvering demand, obstacle density risk, and neighboring-UAV interaction risk, enabling the planner to balance real-time computation and constraint-handling capability under limited perception. We decomposed the UAV energy consumption and formulated the total energy consumption of the swarm as the objective function. An optimal control sequence was derived using the Gaussian collocation method by integrating obstacle avoidance constraints for fixed-wing UAVs and environmental limitations. Comparative simulations against the implemented fixed-discretization interior-point and SQP baselines showed that the proposed DA-GCMPC method achieved lower computation time and better trajectory quality metrics under the tested simulation settings, with average per-step computation times below 80 ms. In addition, an eight-UAV semi-physical hardware-in-the-loop validation was conducted to verify the real-time executability of the proposed method in a closed-loop flight control system.

Drones doi: 10.3390/drones10070540

Authors: Tong Jiang Xuechen Gu Jianchuan Ye Zengzhen Mi Tao Jiang

Active area protection against unauthorized UAV swarms requires coordinated target assignment strategies that account for both low-level physical capabilities and the stochastic, consumptive nature of physical interceptions. This paper presents a decentralized target assignment framework based on probabilistic flow optimization. By utilizing an aerodynamics-aware flight model, we derive a probabilistic prior to capture the geometric dependency of terminal interception success under high-velocity maneuvers. Modeling the defense process as a probabilistic consumption flow couples initial tactical assignments with conditional transition flows, allowing surviving defensive assets to be proactively redistributed to secondary unauthorized intrusions. To resolve this problem under practical communication and sensing constraints, we develop the Distributed Flow-regularized Market-based Consensus (DFMC) algorithm. The proposed algorithm decomposes the global optimization into localized subproblems and employs a water-filling projection to plan secondary paths. Simulation results demonstrate that the proposed framework yields improved interception rates and better spatial resource dispersion compared to conventional auction-based baselines, while maintaining stable scalability in dense interception scenarios.