Drones

31 pages, 3703 KB

Open AccessArticle

CFD-Based Aerodynamic Characterization and Semi-Analytical Modelling of a NACA 0012 Four-Bladed Cyclorotor for Next-Generation UAV Propulsion

by Mădălin Dombrovschi and Daniel-Eugeniu Crunțeanu

Drones 2026, 10(6), 462; https://doi.org/10.3390/drones10060462 (registering DOI) - 13 Jun 2026

Abstract

Next-generation unmanned aerial vehicles require compact propulsion systems capable of providing efficient vertical lift, rapid thrust vectoring, and improved maneuverability. Cyclorotors represent a promising alternative to conventional propellers, but their aerodynamic behavior is governed by highly unsteady blade–wake interactions, making performance prediction challenging. [...] Read more.

Next-generation unmanned aerial vehicles require compact propulsion systems capable of providing efficient vertical lift, rapid thrust vectoring, and improved maneuverability. Cyclorotors represent a promising alternative to conventional propellers, but their aerodynamic behavior is governed by highly unsteady blade–wake interactions, making performance prediction challenging. This study investigates a four-bladed cyclorotor equipped with NACA 0012 airfoils using transient computational fluid dynamics simulations and a calibrated semi-analytical blade-element model. The numerical analysis was performed over a rotational-speed range of 368–2305 rpm and for several pitch-amplitude configurations, including 5°, 7.5°, 10°, 12.5° and 15°. The results showed that the favorable pitch amplitude decreases with increasing rotational speed, shifting from larger amplitudes at low RPM to approximately 5° at higher RPM values. The semi-analytical model reproduced the main CFD trends for lift, drag, moment, and power, providing a reduced-order tool for preliminary cyclorotor performance estimation. The comparison confirmed that pitch-amplitude selection strongly influences aerodynamic loading and efficiency and should therefore be adapted to the operating regime. The proposed CFD-based methodology, supported by semi-analytical modelling, provides a useful framework for the aerodynamic characterization and early-stage optimization of cyclorotor propulsion systems for UAV applications. Full article

(This article belongs to the Special Issue Computational Fluid Dynamics for Next-Generation Unmanned Aerial Vehicles)

32 pages, 2159 KB

Open AccessArticle

Traffic-Predictive Drone Scheduling: Day-Ahead Synchronization of Mobile Depots and Parallel Aerial Sorties in Urban Airspace

by Shihab Hasan, Tarek Sheltami and Ashraf Mahmoud

Drones 2026, 10(6), 461; https://doi.org/10.3390/drones10060461 (registering DOI) - 13 Jun 2026

Abstract

Urban Unmanned Aerial Vehicle (UAV) logistics operations are frequently constrained by the intersection of limited battery endurance and dynamic ground traffic. When mobile depots are delayed by congestion, onboard drone fleets experience extended idling periods, leading to constrained sortie generation and reduced asset [...] Read more.

Urban Unmanned Aerial Vehicle (UAV) logistics operations are frequently constrained by the intersection of limited battery endurance and dynamic ground traffic. When mobile depots are delayed by congestion, onboard drone fleets experience extended idling periods, leading to constrained sortie generation and reduced asset utilization. To address this bottleneck, this paper introduces a traffic-predictive multi-UAV dispatch framework for deterministic day-ahead planning under modeled urban operating conditions. By coupling a count-derived macroscopic speed surrogate learned using XGBoost with a Particle Swarm Optimization (PSO)–Mixed-Integer Linear Programming (MILP) optimization architecture, the framework synchronizes mobile depot trajectories with forecasted low-congestion windows and pre-allocates endurance-feasible parallel aerial sorties. Controlled computational experiments across 30 synthetic routing instances demonstrate the potential value of this approach within the stated modeling assumptions. Compared to baseline clustered deployments, the traffic-aware framework raises mean fleet utilization from 0.43 to 0.63—a 46.2% relative improvement driven by temporal compression of the mission window rather than an absolute increase in flight hours. Furthermore, the proposed framework reduces total mission completion time by 69.87% relative to the conventional truck-only baseline, while achieving a 29.58% incremental gain over static speed drone deployments. These findings suggest that incorporating predictive ground traffic information into day-ahead UAV scheduling can improve modeled fleet efficiency; however, field validation with measured route-level speeds, real delivery demand, and operational constraints remains necessary before deployment-level claims can be made. Full article

(This article belongs to the Section Innovative Urban Mobility)

25 pages, 18006 KB

Open AccessArticle

Multi-UAV Cooperative Localization in Pseudolite-Augmented GNSS-Denied Regions: An Anomaly-Resilient Adaptive Kalman Filter with Group Covariance Compensation

by Chengyan Ji, Xiye Guo, Yuqiu Tang, Xiaohe Han and Yuhang Song

Drones 2026, 10(6), 460; https://doi.org/10.3390/drones10060460 (registering DOI) - 12 Jun 2026

Abstract

In complex low-altitude environments, unmanned aerial vehicles (UAVs) require reliable positioning, yet Global Navigation Satellite System (GNSS) signals are vulnerable to occlusion and interference. Pseudolite-augmented cooperative localization, which combines ground base-station signals with inter-UAV relative observations, can complement GNSS in such environments. However, [...] Read more.

In complex low-altitude environments, unmanned aerial vehicles (UAVs) require reliable positioning, yet Global Navigation Satellite System (GNSS) signals are vulnerable to occlusion and interference. Pseudolite-augmented cooperative localization, which combines ground base-station signals with inter-UAV relative observations, can complement GNSS in such environments. However, two practical issues remain in real-world deployment: UAV-to-base-station (U-B) and UAV-to-UAV (U-U) observations have markedly different error statistics that a unified noise adjustment cannot handle, and the conservative covariance estimates produced by Covariance Intersection (CI) fusion bias the innovation-based adaptive noise estimation in distributed architectures. To address these issues, this paper proposes a Distributed Group Covariance Compensation Adaptive Kalman Filter (DGCC-AKF) for collaborative enhancement of UAV regional localization. DGCC-AKF establishes a group adaptive mechanism that independently adjusts the noise covariance matrices of U-B and U-U observations, enabling observation-type-level adaptive weighting that suppresses anomalous U-B or U-U measurements at the group level. In addition, a bounded covariance compensation factor is incorporated to alleviate the CI-induced conservatism in the adaptive noise estimation. The proposed method is evaluated on a 2800 km² semi-physical testbed based on the Ground-based High-precision Local Positioning System (GH-LPS) pseudolite network using measured U-B observations and high-dynamic (>300 km/h) flight trajectories collected from a fixed-wing platform across three independent flight sessions. Results demonstrate that under observation fault periods, the proposed method improves 3D positioning accuracy by up to about 75% over single-UAV extended Kalman filter (EKF). Compared with two advanced algorithms in this field, variational Bayesian adaptive Kalman filter (VBAKF) and maximum correntropy criterion Kalman filter (MCC-EKF), it is the only scheme that remains accurate and stable across all UAVs and fault types. The framework provides a practical step toward field deployment for resilient multi-UAV cooperative navigation in pseudolite-augmented GNSS-denied regions. Full article

(This article belongs to the Topic Navigation and Positioning System: Opportunities and Obstacles)

► Show Figures

Figure 1

23 pages, 93772 KB

Open AccessArticle

TriCross-D2D: A Cross-Scene, Cross-View, and Cross-Weather Dataset for Drone-to-Drone Detection

by Wei Tang, Qilong Li, Yueping Peng, Hexiang Hao, Wenchao Kang, Xuekai Zhang, Liming Hou and Hongyan Lu

Drones 2026, 10(6), 459; https://doi.org/10.3390/drones10060459 (registering DOI) - 12 Jun 2026

Abstract

Drone-to-drone (D2D) detection is a critical yet underexplored task in low-altitude intelligent perception, where UAV targets are often small, weakly textured, motion-affected, and disturbed by complex backgrounds and environmental changes. Existing cross-domain detection datasets mainly focus on ground objects or single-factor shifts, making [...] Read more.

Drone-to-drone (D2D) detection is a critical yet underexplored task in low-altitude intelligent perception, where UAV targets are often small, weakly textured, motion-affected, and disturbed by complex backgrounds and environmental changes. Existing cross-domain detection datasets mainly focus on ground objects or single-factor shifts, making them insufficient for evaluating D2D detection under coupled real-world variations. To address this gap, we present TriCross-D2D, an RGB air-to-air UAV detection dataset and benchmark with three explicit domain shifts: scene, viewpoint, and weather. Built from real flight videos and controlled synthetic fog, TriCross-D2D contains 13 RGB video sequences, 23,403 raw frames, 7045 benchmark images, and 9771 annotated UAV instances. It provides a fixed split of 4045 Source_train images, 2000 Target_train images, and 1000 Target_val images, supporting both unsupervised domain adaptation (UDA) and semi-supervised domain adaptation (SSDA). The dataset is dominated by small objects, with extremely tiny, tiny, and small targets accounting for 73.8% of all instances. Benchmark results show that existing cross-domain detectors still perform limitedly on TriCross-D2D, especially under stricter localization and recall metrics. Single-factor analysis further reveals that the coupled scene–viewpoint–weather protocol is more challenging than isolated shifts, with viewpoint variation producing a particularly strong domain gap. As an exploratory enhanced baseline, SCOPE-DA-RTDETR improves DA-RTDETR from 28.63/13.12/22.39 to 29.94/13.71/23.40 in

{AP}_{50} / {AP}_{50 - 95} / AR

, showing consistent but modest gains. These findings demonstrate that TriCross-D2D provides a challenging and discriminative benchmark for cross-domain D2D small-object detection. Full article

(This article belongs to the Special Issue Detection, Identification and Tracking of UAVs and Drones: 2nd Edition)

► Show Figures

Figure 1

33 pages, 10139 KB

Open AccessArticle

GPOD: Geographic Priors and Object Detection for Candidate-Guided Target Localization in City-Scale UAV Vision-and-Language Navigation

by Yuze Liu, Changming Xu, Kewen Xiao, Yuhua Wu and Ziyu Li

Drones 2026, 10(6), 458; https://doi.org/10.3390/drones10060458 - 11 Jun 2026

Abstract

City-scale unmanned aerial vehicle vision-and-language navigation (UAV-VLN) requires accurate upstream target localization from an overhead map, onboard observation, and language description. Existing VLM-based methods often treat road names, landmarks, and spatial relations as raw text, leaving the model to search a large map [...] Read more.

City-scale unmanned aerial vehicle vision-and-language navigation (UAV-VLN) requires accurate upstream target localization from an overhead map, onboard observation, and language description. Existing VLM-based methods often treat road names, landmarks, and spatial relations as raw text, leaving the model to search a large map and implicitly infer geometric constraints. This paper proposes GPOD, an inference-time candidate-prior interface for the upstream target-localization stage in city-scale UAV-VLN. GPOD converts language anchors, spatial relations, target-category cues, static map objects, and vehicle detections into ranked candidate priors through branch-specific candidate generation, thereby reformulating unconstrained full-map coordinate regression as candidate-prior-conditioned coordinate prediction. The static branch aligns language constraints with map-object geometries, while the dynamic branch uses YOLOv8l-VisDrone with Slicing Aided Hyper Inference (SAHI) to construct detection-conditioned vehicle candidates. In the GPOD-VLM setting, ranked candidates are injected as structured spatial prompts and the base VLM predicts the final continuous coordinates; GPOD-Direct is a candidate-direct diagnostic variant that directly uses candidate centers without VLM coordinate regression. On the CityNav localization protocol, GPOD improves FlightGPT Overall SR@20m from 15.23% to 25.61% and consistently reduces Mean Navigation Error (Mean NE) across splits and backbones. On Val-Unseen, GPOD-Direct (Top-1) reaches 32.59% SR@20m, showing that ranked candidate priors provide strong discrete localization signals. These results show that inference-time candidate priors can reduce city-scale search ambiguity without updating the base VLM parameters, while also revealing a candidate-utilization gap in the current prompt-based continuous coordinate-regression interface. Full article

(This article belongs to the Topic Advanced Methods in Unmanned Aerial Vehicle Control, Navigation, and Safety)

► Show Figures

Figure 1

34 pages, 4235 KB

Open AccessArticle

A Multimodal Data Fusion Algorithm for Urban Low-Altitude UAV Perception

by Bowen Xu, Peinan He, Xu Wang, Yixiao Zhang and Yuanjie Zhao

Drones 2026, 10(6), 457; https://doi.org/10.3390/drones10060457 - 11 Jun 2026

Abstract

Accurate Unmanned Aerial Vehicle (UAV) position estimation is the cornerstone of urban low-altitude safety management systems. Time Difference of Arrival (TDOA) and Remote Identification (Remote ID) are widely used surveillance technologies with complementary characteristics. TDOA provides high-rate updates but suffers from geometry-induced horizontal–vertical [...] Read more.

Accurate Unmanned Aerial Vehicle (UAV) position estimation is the cornerstone of urban low-altitude safety management systems. Time Difference of Arrival (TDOA) and Remote Identification (Remote ID) are widely used surveillance technologies with complementary characteristics. TDOA provides high-rate updates but suffers from geometry-induced horizontal–vertical anisotropy and multipath effects, while Remote ID supplies absolute state information yet struggles with intermittent sampling and packet loss. Existing fusion schemes typically address these issues in isolation: sequential filtering manages asynchrony but assumes Gaussian noise, robust estimators suppress outliers at the cost of discarding valid data, and coupled-filter architectures allow vertical anomalies to contaminate horizontal estimates through the Kalman gain cross-coupling. No prior framework jointly handles structural TDOA altitude jumps, stochastic Remote ID timing jitter, and the geometric anisotropy between estimation subspaces within a single coherent pipeline. To bridge this gap, we propose a Hybrid Conditional Kalman Filter (HCKF) framework comprising three integrated modules. First, a kinematics-based temporal alignment module maps asynchronous measurements onto a uniform timeline and predicts missing samples, resolving cross-modal time mismatches. Second, a measurement quality evaluation mechanism detects TDOA altitude steps via robust two-layer stratification and scores Remote ID timing irregularity through a confidence mapping, converting these anomalies into dynamic covariance adjustments and weight caps without discarding observations. Third, a Subspace-Decoupled Fusion strategy exploits the physical insight that TDOA horizontal precision derives from hyperbolic intersection geometry, whereas its vertical estimates suffer from weak observability due to near-coplanar ground-station deployment . By applying entropy-guided weighting in the horizontal plane and a conditional Remote ID-dominant rule in the vertical axis, this design prevents cross-dimensional error propagation. The framework was validated using three real-world flight missions at distinct altitudes (255 m, 345 m, and 440 m) totaling 13.51 km of flight distance, with RTK serving as ground truth. HCKF reduces the Root Mean Square Error by over 40% relative to single-source baselines (95% bootstrap confidence interval: [35.2%, 48.7%]), and paired Wilcoxon signed-rank tests confirm statistically significant improvement (

p < 0.01

) over standard EKF, Covariance Intersection, and Iterative CI across all three tracks. Full article

(This article belongs to the Special Issue Detection, Identification and Tracking of UAVs and Drones: 2nd Edition)

29 pages, 31575 KB

Open AccessArticle

DCA-DeepLab: Dual-Coordinate Attention DeepLab with Adaptive Focal Loss for Cotton Growth Semantic Segmentation from UAV Remote Sensing Images

by Liruizhi Jia, Jiazhan Gao, Zuolong Li, Heng Shi and Jihong Zhu

Drones 2026, 10(6), 456; https://doi.org/10.3390/drones10060456 - 11 Jun 2026

Abstract

UAV remote sensing provides centimetre-level imagery for fine-grained cotton growth monitoring, yet existing segmentation models face three challenges: cotton fields exhibit a pronounced row and column structure that standard convolutions struggle to capture; conventional decoders fuse features statically, suppressing fine boundary cues; and [...] Read more.

UAV remote sensing provides centimetre-level imagery for fine-grained cotton growth monitoring, yet existing segmentation models face three challenges: cotton fields exhibit a pronounced row and column structure that standard convolutions struggle to capture; conventional decoders fuse features statically, suppressing fine boundary cues; and the pixel-level class distribution is severely imbalanced. We present DCA-DeepLab, built on DeepLabv3+ with three task-specific components: a Dual-Coordinate Attention Gating (DCAG) module that decouples horizontal and vertical dependencies to encode row and column structures; a Multi-Scale Attention-Guided Modulated Feature Merging (MSAM-MFM) module that reweights semantic and detail features at each location; and an adaptive pixel-level modulated focal loss (APMFL), which focuses training on hard, minority-class pixels. We construct a cotton growth dataset of 11,745 UAV patches with four semantic classes. On this dataset and the public LoveDA benchmark, DCA-DeepLab attained the highest mIoU among the compared methods (51.74% and 51.71%), exceeding the strongest cotton baseline by 1.10 percentage points. Relative to DeepLabv3+, the Vigorous and Sparse minority-class IoUs improved by 3.51 and 1.91 percentage points, respectively, and Vigorous recall rose from 51.85% to 60.04%, with only 3.9% more parameters. These results show that encoding directional structure and adaptively balancing class contributions benefits fine-grained UAV crop segmentation. Full article

(This article belongs to the Section Drones in Agriculture and Forestry)

► Show Figures

Figure 1

40 pages, 27368 KB

Open AccessArticle

Intelligent Maintenance and Routing Decision Making for UAV Clusters in Harsh Environments

by Sujunjie Sun, Cong Cao, Yan Lyu, Xueyong Xu, Renzhi Huang, Yuhang Xu, Shenquan Tang, Chenchen Fu and Weiwei Wu

Drones 2026, 10(6), 455; https://doi.org/10.3390/drones10060455 - 10 Jun 2026

Abstract

UAV clusters operating in harsh environments must maintain connectivity under sudden node failures and dynamic traffic demands. Existing studies often address maintenance and routing separately, which leads to unnecessary reorganization overhead and forwarding bottlenecks around gateway nodes. To overcome these limitations, we propose [...] Read more.

UAV clusters operating in harsh environments must maintain connectivity under sudden node failures and dynamic traffic demands. Existing studies often address maintenance and routing separately, which leads to unnecessary reorganization overhead and forwarding bottlenecks around gateway nodes. To overcome these limitations, we propose a unified UAV Cluster Maintenance and Routing System (UCMR) that combines a firefly-inspired clustering method, a DQN–GCN-based medium-level interference recovery strategy (DGMR), and load-aware routing over forwarding neighbors. The maintenance component improves backbone repair by integrating global topology features with local observations, while the routing component exploits edge and node features to distribute forwarding pressure beyond gateway nodes. Extensive experiments demonstrate that UCMR outperforms representative benchmark methods in terms of average end-to-end delay and packet delivery ratio. Full article

► Show Figures

Figure 1

14 pages, 2758 KB

Open AccessArticle

Liquid Time-Constant Network-Enhanced INS/SAR Integrated Localization Method for UAVs in Degraded Scenarios

by Jing He, Rui Li, Chunlei Pang, Peiran Li and Chenhao Zhao

Drones 2026, 10(6), 454; https://doi.org/10.3390/drones10060454 - 10 Jun 2026

Abstract

Synthetic aperture radar (SAR) can acquire navigation data to correct inertial navigation system (INS) errors even under global navigation satellite system (GNSS)-denied conditions. However, when unmanned aerial vehicles (UAVs) may deactivate the SAR system to maintain radio silence, or the SAR sensor may [...] Read more.

Synthetic aperture radar (SAR) can acquire navigation data to correct inertial navigation system (INS) errors even under global navigation satellite system (GNSS)-denied conditions. However, when unmanned aerial vehicles (UAVs) may deactivate the SAR system to maintain radio silence, or the SAR sensor may be subjected to transient interference, the INS/SAR integrated navigation system transitions to degraded scenarios without SAR navigation data. Furthermore, the irregular sampling characteristics of SAR navigation data pose significant challenges to the localization performance of the INS/SAR integrated navigation system. In order to address the above challenges faced by UAVs, we propose a liquid time-constant (LTC) network-enhanced INS/SAR integrated localization method. The method adopts a loosely coupled integration strategy with training and prediction modes. During training, an LTC-assisted localization prediction network (LTC-ALPN) is designed to model input–output relationships using prior flight data while explicitly accounting for the non-uniform temporal sampling characteristics of SAR measurements. In prediction mode, the trained LTC-ALPN forecasts missing SAR navigation information, which is subsequently fused with INS outputs via a Kalman filter to maintain high-precision positioning during SAR outages. Experimental results demonstrate that, compared to pure INS localization in degraded scenarios, the proposed method reduces northward error MAE and RMSE by approximately 92.8% and 93.9% and eastward error MAE and RMSE by 54.1% and 67.1%. Against suboptimal network baselines, further improvements of 50.8%/38.1% (north) and 17.1%/16.7% (east) in MAE/RMSE were achieved. Full article

► Show Figures

Figure 1

24 pages, 4438 KB

Open AccessArticle

Dynamic Self-Organization and Safe Navigation for Hierarchical Embodied Swarms

by Lanbo Wu and Chen Wei

Drones 2026, 10(6), 453; https://doi.org/10.3390/drones10060453 - 10 Jun 2026

Abstract

This paper is concerned with cooperative multi-UAV navigation in a planar obstacle environment. A hierarchical embodied swarm framework with leader, subleader, and follower roles is proposed. At the high level, a passable-corridor-driven decision layer is developed to perform split–merge reconfiguration and navigate/encircle mode [...] Read more.

This paper is concerned with cooperative multi-UAV navigation in a planar obstacle environment. A hierarchical embodied swarm framework with leader, subleader, and follower roles is proposed. At the high level, a passable-corridor-driven decision layer is developed to perform split–merge reconfiguration and navigate/encircle mode switching. At the low level, a multi-term force synthesis controller is constructed for formation maintenance, inter-agent collision avoidance, obstacle avoidance, and sub-swarm cohesion. To accommodate both rule-based and local large language model (LLM) decisions, a feasibility projection operator is introduced so that only kinematically admissible structural actions are executed. In addition, a LiDAR-based obstacle-repulsion term and an occlusion-attenuated attraction mechanism are incorporated to improve navigation safety in cluttered environments. A Lyapunov analysis of the smooth controller core further certifies that, for a known (possibly time-varying) cruise velocity compensated by feedforward, the formation tracking error is uniformly bounded by the initial energy. Finally, multi-seed numerical simulations verify the proposed framework in standard, ablated, and complex scenarios. In the hardest alternating-gate scenario, the LLM-assisted variant raises mission success from

0.000

to

0.100

, increases the goal-reaching ratio from

0.025

to

0.125

, and reduces the mean terminal error from

44.738 m

to

39.851 m

, showing the value of semantic high-level reconfiguration under tight passage constraints. Full article

(This article belongs to the Special Issue UAV Swarm Intelligent Control and Decision-Making)

► Show Figures

Figure 1

49 pages, 37729 KB

Open AccessFeature PaperArticle

Comparative Evaluation of Classical, Hybrid, and RL-Based 3D Trajectory Planning for Multi-UAV Systems

by Ilya Mashkov, Angelika Kochetkova, Valerii Serpiva, Grigoriy Yashin and Pavel Golikov

Drones 2026, 10(6), 452; https://doi.org/10.3390/drones10060452 (registering DOI) - 9 Jun 2026

Abstract

This study investigates offline trajectory planning strategies for multi-UAV missions in complex 3D environments, with the aim of systematically comparing classical, hybrid, and reinforcement learning-based approaches under unified evaluation conditions. Two simulation scenarios were considered: an uneven terrain environment with elevation-induced constraints and [...] Read more.

This study investigates offline trajectory planning strategies for multi-UAV missions in complex 3D environments, with the aim of systematically comparing classical, hybrid, and reinforcement learning-based approaches under unified evaluation conditions. Two simulation scenarios were considered: an uneven terrain environment with elevation-induced constraints and a planar obstacle-rich environment. The evaluated planners include graph-based (A*), sampling-based (RRT, RRT*), gradient-based (APF), a hybrid APF B-RRT* method, and a DQN-based reinforcement learning planner with spatial attention and reward shaping. Performance was assessed using geometric, safety, energetic, and computational metrics. The results show that A* consistently produces the shortest and most stable trajectories with low energy consumption but at increased computational cost in high-resolution environments. Sampling-based planners exhibit higher variability and planning time, while APF achieves computational efficiency but may violate safety margins. The hybrid planner provides improved robustness across scenarios. The reinforcement learning planner demonstrates consistent safety compliance and strong inter-UAV separation in both environments, also with longer trajectories and higher energy usage. Overall, the study highlights trade-offs between determinism, scalability, safety, and adaptability across planning paradigms. Full article

(This article belongs to the Special Issue Advances in Cartography, Mission Planning, Path Search, and Path Following for Drones: 2nd Edition)

► Show Figures

Figure 1

35 pages, 1068 KB

Open AccessReview

UAV-Based Remote Sensing and Artificial Intelligence for Climate-Smart Agriculture: A Systematic Review of Technologies, Analytics, and Applications in Smallholder Systems

by Andrew Manu, Jeff Dacosta Osei and Thomas Lawler

Drones 2026, 10(6), 451; https://doi.org/10.3390/drones10060451 - 9 Jun 2026

Abstract

Unmanned aerial vehicle (UAV)-based remote sensing combined with artificial intelligence (AI) has emerged as a key enabler of climate-smart agriculture (CSA). However, the extent to which these technologies operationalize CSA’s three pillars, productivity, adaptation, and mitigation, remains unevenly assessed. This study presents a [...] Read more.

Unmanned aerial vehicle (UAV)-based remote sensing combined with artificial intelligence (AI) has emerged as a key enabler of climate-smart agriculture (CSA). However, the extent to which these technologies operationalize CSA’s three pillars, productivity, adaptation, and mitigation, remains unevenly assessed. This study presents a PRISMA-guided systematic review of 59 peer-reviewed studies examining UAV–AI applications in agricultural systems. The synthesis categorizes platform configurations, sensor modalities, analytical architectures, geographic distribution, and data integration strategies, and evaluates their alignment with CSA objectives. Results indicate that productivity-oriented applications, including yield estimation, biomass mapping, and nutrient assessment, are the most mature, while adaptation-focused stress detection is also well established. In contrast, mitigation-oriented applications, such as carbon quantification and greenhouse gas monitoring, remain comparatively underrepresented. The analysis further reveals a growing convergence toward multimodal sensing and cross-scale data integration linking UAV observations with satellite and environmental datasets. However, substantial variability in validation approaches and dataset representativeness limits generalizability and scalability. Advancing UAV–AI contributions to CSA therefore requires methodological standardization, interoperable data governance, and strengthened institutional capacity. Collectively, the findings position UAV–AI systems as emerging components of climate-smart agricultural intelligence infrastructure rather than isolated monitoring tools. Full article

(This article belongs to the Special Issue Advances in UAV-Based Remote Sensing for Climate-Smart Agriculture)

► Show Figures

Figure 1

30 pages, 3689 KB

Open AccessArticle

Resource-Aware Surprise Reinforcement Learning for Collision Avoidance in Maritime UAV Encounters

by Zuocheng Liu, Qi Feng, Zidong Wang and Xiaoguang Gao

Drones 2026, 10(6), 450; https://doi.org/10.3390/drones10060450 - 9 Jun 2026

Abstract

Collision avoidance in maritime unmanned aerial vehicle (UAV) operations must satisfy two competing objectives: ensuring reliable safety separation and minimizing unnecessary maneuver commands that increase operator burden and communication overhead. While deep reinforcement learning (DRL) has shown promise in handling high-dimensional encounter states, [...] Read more.

Collision avoidance in maritime unmanned aerial vehicle (UAV) operations must satisfy two competing objectives: ensuring reliable safety separation and minimizing unnecessary maneuver commands that increase operator burden and communication overhead. While deep reinforcement learning (DRL) has shown promise in handling high-dimensional encounter states, standard DRL approaches often prioritize safety at the cost of operational suitability, leading to frequent, oscillatory, or unnecessary avoidance commands that erode remote operator trust and consume limited communication bandwidth. To address this challenge, this paper proposes Resource-Aware Intrinsic Surprise Exploration (RAISE), a unified framework that balances collision avoidance performance with command economy. We conceptualize the issuance of avoidance maneuvers as a consumable “virtual resource”, compelling the agent to optimize its intervention budget. RAISE integrates this mechanism into the Soft Actor–Critic (SAC) architecture, augmented by a surprise-based intrinsic reward derived from the ensemble forward dynamics prediction error. This allows the agent to efficiently explore complex encounter scenarios driven by curiosity, while a resource-aware coefficient adaptively suppresses redundant actions when the communication or operational budget is constrained. Furthermore, an adaptive exponential moving average (EMA) scaling mechanism is introduced to stabilize the interplay between intrinsic and extrinsic rewards. Extensive simulations under diverse resource constraints and encounter geometries demonstrate that RAISE outperforms state-of-the-art baselines. It significantly reduces maneuver reversal rates and strengthens command stability without compromising safety margins. Specifically, under resource-constrained settings, RAISE suppresses excessive and unstable advisory behavior by reducing strengthening and reversal commands while maintaining effective collision avoidance; under resource-rich settings, it flexibly enhances safety buffers, demonstrating superior adaptability and operational realism for autonomous maritime UAV systems. Robustness evaluation confirms that RAISE maintains stable performance under sensor noise and wind disturbances. Full article

(This article belongs to the Special Issue Artificial Intelligence-Driven Drones Systems for Marine Engineering Applications)

► Show Figures

Figure 1

24 pages, 62342 KB

Open AccessArticle

DCAFuse: A Differential Cross-Attention Transformer Network for Infrared and Visible Image Fusion in UAV-Based Wilderness Search and Rescue

by Yu Jing, Yili Yan, Zhao Li, Fugui Qi, Tao Lei, Jianqi Wang and Guohua Lu

Drones 2026, 10(6), 449; https://doi.org/10.3390/drones10060449 - 9 Jun 2026

Abstract

Infrared and visible image fusion is critical for unmanned aerial vehicle (UAV) wilderness search and rescue. By integrating thermal radiation of the targets and texture details of the scenario, it enables accurate search for the wounded and comprehensive perception of disaster areas, thereby [...] Read more.

Infrared and visible image fusion is critical for unmanned aerial vehicle (UAV) wilderness search and rescue. By integrating thermal radiation of the targets and texture details of the scenario, it enables accurate search for the wounded and comprehensive perception of disaster areas, thereby significantly improving emergency rescue efficiency. To alleviate data scarcity, we construct UAV-MSR, an infrared-visible dataset for casualty search, comprising 3889 paired images captured under diverse weather, illumination, and scenarios. Existing Transformer-based fusion methods mainly focus on high-intensity pixels while inadequately modeling low-intensity complementary features, resulting in blurred details and degraded target contrast in fused images. To this end, we propose a novel differential cross-attention Transformer network to address the issue of complementary information loss. Specifically, the encoder integrates convolution operations for local detail extraction and self-attention mechanisms for global context modeling. Then, we design a differential cross-attention guided feature fusion module to enhance the representation and preservation of detailed complementary features. Furthermore, a pixel loss function with a segmentation strategy is employed to improve the saliency of the target, enabling the fused image to facilitate subsequent target detection tasks. Experimental results and ablation studies demonstrate that the proposed method achieves notable performance and generalization ability. In summary, this work delivers a multimodal dataset and an efficient infrared-visible image fusion network to enable comprehensive perception for UAVs in wilderness search and rescue scenarios. Full article

► Show Figures

Figure 1

33 pages, 5566 KB

Open AccessReview

A Review of Reinforcement Learning for Multirotor UAVs from a Hierarchical Control Perspective: Biomimetic Architecture and Sim-to-Real

by Wei Wei, Xubo Zhao, Yongjie Shu, Qingkai Meng, Mingkai Ding, Yunyi Wang and Qingdong Yan

Drones 2026, 10(6), 448; https://doi.org/10.3390/drones10060448 - 8 Jun 2026

Abstract

As unmanned aerial vehicle (UAV) systems evolve from automated execution toward autonomous decision-making, multirotor UAVs increasingly face complex dynamics, uncertain sensing conditions, and task-level autonomy demands. Reinforcement learning (RL) has emerged as a promising learning-based paradigm for addressing these challenges. Existing surveys on [...] Read more.

As unmanned aerial vehicle (UAV) systems evolve from automated execution toward autonomous decision-making, multirotor UAVs increasingly face complex dynamics, uncertain sensing conditions, and task-level autonomy demands. Reinforcement learning (RL) has emerged as a promising learning-based paradigm for addressing these challenges. Existing surveys on RL-based UAV control predominantly classify methods from an algorithmic or learning-paradigm perspective, while relatively little attention has been paid to the functional roles of RL policies within the control loop. This often leads to an unclear correspondence between algorithmic characteristics and the requirements of different control layers. To address this gap, this review proposes a biomimetic “spinal cord–cerebellum–cerebrum” framework, organizing existing RL studies into low-level dynamic stabilization, mid-level perception–action coordination, and high-level task planning and decision-making. The proposed hierarchy emphasizes the functional role and intervention depth of RL policies within the control architecture, further supporting a layer-wise analysis of sim-to-real challenges. This review aims to provide a structured understanding of the roles of reinforcement learning in hierarchical UAV control and to highlight future research directions toward robust real-world deployment. Full article

(This article belongs to the Special Issue Advances in Deep Learning for Drones and Its Applications: 2nd Edition)

► Show Figures

Figure 1

22 pages, 30453 KB

Open AccessArticle

CPD-UAV: A Benchmark Dataset for Detecting Personnel Visually Blended with the Environment Under UAV Perspective

by Xuekai Zhang, Wenchao Kang, Yueping Peng, Wei Tang, Qilong Li, Hexiang Hao, Liming Hou and Xin Ying

Drones 2026, 10(6), 447; https://doi.org/10.3390/drones10060447 - 8 Jun 2026

Abstract

Camouflaged object detection (COD) is important for intelligent UAV monitoring and search-and-rescue operations. However, existing benchmarks focus primarily on natural camouflage, creating a noticeable domain shift for specific applications such as the search and rescue of individuals visually similar to their surroundings due [...] Read more.

Camouflaged object detection (COD) is important for intelligent UAV monitoring and search-and-rescue operations. However, existing benchmarks focus primarily on natural camouflage, creating a noticeable domain shift for specific applications such as the search and rescue of individuals visually similar to their surroundings due to their clothing. To investigate this shift, we introduce CPD-UAV, a benchmark comprising 1061 high-resolution images with detailed pixel-level annotations across diverse terrains and flight altitudes. Benchmarking of seven state-of-the-art models on this dataset reveals specific challenges. Specifically, the scale variations and “vanishing boundaries” inherent in aerial perspectives can lead to boundary localization inaccuracies. Furthermore, this evaluation observes the deceptive nature of traditional metrics, such as Mean Absolute Error (MAE), when targets occupy small image proportions. To address the degradation of weak target signals during feature integration, we propose a lightweight, plug-and-play component: the Residual Gated Alignment Module (RGAM). RGAM handles scale variations by establishing semantic anchors in deep network layers, mitigating signal dilution and highlighting micro-targets against complex backgrounds. By integrating RGAM into three representative baselines, we demonstrate that the enhanced architectures achieve a competitive performance level. Quantitative results show consistent improvements in structural integrity (structure-measure,

S_{m}

) and boundary localization. Ultimately, this work provides a practical data platform and an effective algorithmic solution for advancing aerial monitoring systems. Full article

(This article belongs to the Special Issue Advances in Deep Learning for Drones and Its Applications: 2nd Edition)

► Show Figures

Figure 1

21 pages, 3695 KB

Open AccessArticle

Joint Position–Orientation Deployment Design of UAV-Borne Linear-Array Angle-of-Arrival Sensors for Target UAV Localization

by Jiawei Tang, Tian Chang, Haiqi Liu, Zhe Yu, Dekang Liu and Xuhui Ding

Drones 2026, 10(6), 446; https://doi.org/10.3390/drones10060446 - 7 Jun 2026

Abstract

This paper investigates joint deployment of unmanned aerial vehicle (UAV)-borne linear-array angle-of-arrival (AOA) sensors for localizing a target UAV in three-dimensional space. Since each sensing UAV carries a lightweight one-dimensional (1-D) AOA array, each measurement provides only one angular constraint, and its information [...] Read more.

This paper investigates joint deployment of unmanned aerial vehicle (UAV)-borne linear-array angle-of-arrival (AOA) sensors for localizing a target UAV in three-dimensional space. Since each sensing UAV carries a lightweight one-dimensional (1-D) AOA array, each measurement provides only one angular constraint, and its information contribution depends jointly on the UAV waypoint and array pointing direction. This leads to a coupled coordinate–orientation design problem that differs from conventional full-AOA deployment. We formulate a Cramér–Rao lower bound (CRLB)-based framework under A- and D-optimality criteria, covering both free-flight and constrained hovering regions. By exploiting the structure of the 1-D AOA Fisher information matrix, we show that, for fixed UAV coordinates, the orientation block can be exactly eliminated through a low-dimensional eigenproblem. The resulting reduced coordinate problem is then solved by a geometry-structured sequential quadratic programming (SQP) method, whose curvature model captures the radial and tangential sensitivities induced by line-of-sight geometry. Numerical simulations further validate the effectiveness of the proposed approach. Full article

(This article belongs to the Section Drone Communications)

► Show Figures

Figure 1

28 pages, 22349 KB

Open AccessArticle

Real-Time Elevation and Orientation-Aware Visual Localization for GNSS-Denied Drone Navigation

by Hadi Fares, Ammar Mohanna and Bilal Kaddouh

Drones 2026, 10(6), 445; https://doi.org/10.3390/drones10060445 - 6 Jun 2026

Abstract

Global Navigation Satellite Systems (GNSS)-denied environments pose significant challenges for autonomous drone navigation, requiring robust visual localization systems capable of real-time performance. Existing approaches either sacrifice accuracy for speed or fail to adapt to varying flight altitudes and orientations, limiting their practical deployment. [...] Read more.

Global Navigation Satellite Systems (GNSS)-denied environments pose significant challenges for autonomous drone navigation, requiring robust visual localization systems capable of real-time performance. Existing approaches either sacrifice accuracy for speed or fail to adapt to varying flight altitudes and orientations, limiting their practical deployment. We present Real-Time Elevation and Orientation-Aware Localization Architecture (REOLA), a visual localization system that combines similarity-driven autonomous window sizing, element-wise correlation-based orientation detection, and reinforcement learning with human feedback (RLHF) enhancement for publicly available satellite imagery. On desktop hardware (i7-10700K + RTX 3070), the REOLA achieved approximately 59 FPS performance with sub-5-m accuracy across diverse flight conditions through intelligent similarity-based matching, combined with efficient MobileNet-V3 embeddings and FAISS similarity search. For embedded deployment on NVIDIA Jetson Orin Nano, the system achieved 22.5 FPS, meeting real-time requirements for autonomous drone localization. The system autonomously selects optimal window sizes corresponding to the current elevation and determines drone orientation through element-wise correlation scoring across discrete rotation angles. Enhanced through RLHF, the REOLA achieved a

97.1 %

success rate (sub-5-m localization) while processing frames in 17 milliseconds on desktop hardware (44.4 ms on embedded hardware), providing a substantial margin over real-time requirements. The approach demonstrates particular superiority over traditional keypoint-based methods in challenging environments with repetitive patterns such as agricultural fields, rocky mountains, dense forests, and grasslands, where conventional keypoint detection struggles. We explicitly identify featureless sand dune deserts and open-sea or coastal water flights as out of scope, since the reference satellite imagery in those regimes does not contain stable landmarks. Full article

(This article belongs to the Topic Advanced Methods in Unmanned Aerial Vehicle Control, Navigation, and Safety)

► Show Figures

Figure 1

30 pages, 11527 KB

Open AccessArticle

Intent-Aware CNN–Informer for Long-Horizon Trajectory Prediction of Cross-Domain Unmanned Aerial Vehicles in Constrained Environments

by Yichen Liu, Chijun Zhou, Lei Shao, Yangchao He, Xueqian Wang and Jikun Ye

Drones 2026, 10(6), 444; https://doi.org/10.3390/drones10060444 - 6 Jun 2026

Abstract

Long-horizon trajectory prediction for unmanned aerial vehicles (UAVs) operating in constrained environments remains challenging because of strongly nonlinear dynamics, hidden control effects, and evolving destination-oriented behavior. This challenge is particularly pronounced for highly maneuverable cross-domain unmanned aerial vehicles (CDUAVs), whose glide trajectories are [...] Read more.

Long-horizon trajectory prediction for unmanned aerial vehicles (UAVs) operating in constrained environments remains challenging because of strongly nonlinear dynamics, hidden control effects, and evolving destination-oriented behavior. This challenge is particularly pronounced for highly maneuverable cross-domain unmanned aerial vehicles (CDUAVs), whose glide trajectories are strongly coupled with control and environmental constraints. To address this problem, this paper proposes an intent-aware CNN–Informer framework for accurate long-horizon trajectory prediction. First, a control-affine reformulation of the vehicle dynamics is used to construct physically interpretable DBL control parameters, which reduce the learning difficulty associated with hidden control effects. Second, three continuous intent features—tangential no-fly zone avoidance distance, heading error angle, and relative closing velocity—are introduced to encode destination tendency and avoidance requirements. These features are fused with historical trajectory states and fed into a hybrid CNN–Informer network, where the CNN extracts local maneuver patterns and the Informer captures long-range temporal dependencies. Experiments on a constrained trajectory dataset demonstrate that the proposed method achieves the best performance among all compared models, including SSD-LSTM, Transformer, iTransformer, DLinear, and Informer. Compared with Informer, the proposed approach reduces the average prediction error by 17.2% and significantly improves terminal and maximum prediction errors. These results indicate that the proposed framework provides an effective and physically interpretable solution for long-horizon UAV trajectory prediction in constrained flight scenarios, with potential extensions to behavior-aware forecasting and guidance support in autonomous aerial systems. Full article

(This article belongs to the Section Artificial Intelligence in Drones (AID))

► Show Figures

Figure 1

Journal Description

Drones

Latest Articles

Journal Menu

Journal Browser

Highly Accessed Articles

Latest Books

E-Mail Alert

News

Topics

Conferences

Special Issues

Topical Collections

Further Information

Guidelines

MDPI Initiatives

Follow MDPI