Search Results (249)

Traditional grey wolf optimizer (GWO) frequently suffers from insufficient search diversity, unstable stage transition, and premature convergence when addressing complex optimization tasks. To overcome these limitations, this paper proposes an improved grey wolf optimizer with a Q-learning-guided fitness–distance-weighted selector. For the proposed QLFDGWO framework, first, chaotic mapping is introduced to generate a more diverse initial population. A cosine nonlinear convergence factor is employed to improve adjustment capability during the search process. Additionally, a Q-learning-based strategy selection mechanism is constructed to enable adaptive switching between exploration and exploitation. To further improve the leadership structure of GWO, a Q-learning-guided fitness–distance-weighted selection mechanism is designed, in which the beta and delta wolves are selected by jointly considering fitness quality and spatial distance from the alpha wolf. A dynamic threshold-weighted update strategy is designed to enhance the convergence accuracy and stability of the population. Finally, the proposed algorithm is benchmarked against five representative optimization algorithms using the CEC2017 benchmark function set. Experimental results indicate that QLFDGWO achieves satisfactory performance in terms of optimization accuracy, convergence speed, and robustness. In addition, QLFDGWO is applied to three-dimensional (3D) unmanned aerial vehicle (UAV) path planning under a range of complex scenarios. Simulation results demonstrate that the proposed method can generate feasible, safe flight paths that satisfy terrain and obstacle constraints. Full article

High-resolution, long-term gridded meteorological datasets from in situ observations are crucial for ecosystem monitoring, soil diagnostics, hydrological modelling, and Earth system model evaluation. This study presents two enhanced real-time adaptations of Thornton’s Daymet V4 interpolation method. Daymet4-r1 uses a traditional calibration strategy with exhaustive parameter search, while Daymet4-r2 applies a global optimization algorithm (find_min_global from the dlib library) to adjust parameters automatically at each time step. Both methods were tested over Tuscany using high-resolution terrain and a dense observation network. Validation with leave-one-out method was carried out for the period 1995–2011 for both versions, while Daymet4-r2 underwent extended evaluation from 1991 to 2024 to assess seasonal dynamics and long-term variability. Results show that Daymet4-r2 outperforms Daymet4-r1 and the original Daymet V4 for all variables (mean absolute error of 1.24 mm, 1.06 °C, 1.29 °C, 6.26%, 0.78 m/s, and 2.04 hPa for precipitation, maximum and minimum temperature, relative humidity, wind speed, and sea level pressure, respectively). The largest improvement was observed in minimum temperature due to an enhanced approach for detecting and modelling thermal inversions. The high performance, flexibility, and ability of Daymet4-r2 to operate without prior calibration highlight its potential for model verification, real-time environmental monitoring, and integration into climate services. Full article

To address the problems of state estimation bias, dynamic threat response lag, and insufficient safety margin in formation coordination caused by the mismatch between the three-dimensional continuous motion model and the discrete sampling characteristics of sensors in UAV multi-aircraft collaborative inspection missions under complex dynamic environments, this paper studies a trajectory planning method that integrates model predictive control and multi-constraint optimization. By constructing a three-dimensional continuous motion model of the UAV and discretizing it using the Euler integral method, the mapping deviation between the continuous motion characteristics and the discrete working mechanism of the airborne system is solved. Based on the model predictive control method, a patrol trajectory tracking planning model is designed, and state increment and integral augmentation strategies are introduced to transform global reference trajectory tracking into a constrained quadratic programming problem in the rolling time domain, achieving high-precision closed-loop tracking. Furthermore, a dynamic environment model coupling static terrain height field and sudden spherical threat is constructed to systematically characterize the static obstacles and random dynamic threats faced by the UAV in complex scenarios such as mountains and hills. On this basis, multiple constraints such as flight altitude, pitch angle, horizontal turning angle, terrain safety margin, and multi-aircraft collision avoidance are integrated to establish a comprehensive objective function that includes range cost, attitude penalty, and safety cost. Through a collaborative mechanism of global optimization and local online correction, a reference trajectory that meets the requirements of formation safety and flight efficiency is generated and used as the input command for the tracking planning model, forming a closed-loop architecture of global optimization generation, local closed-loop tracking, and dynamic real-time correction for trajectory planning. Experimental results show that the success rate of dynamic obstacle avoidance in complex dynamic environments is always higher than 99.9%, and the mean square error of trajectory tracking is stable in the range of 0.02–0.04 km, which verifies its significant advantages in dynamic adaptability, tracking accuracy and formation safety. Full article

This paper introduces the research progress of path planning, trajectory tracking control, and multi-machine collaborative operation systems for agricultural robots. It summarizes the development laws of 3D terrain modeling and adaptive path planning algorithms for complex agricultural environments such as hills and mountains, and analyzes the dynamic disturbance characteristics of agricultural machinery under slip, sideslip, and dynamic load changes. Through comprehensive analysis, it is found that traditional kinematic control models have limitations in complex and unstructured environments. Combining soil mechanics mechanisms, variable load identification, and robust control strategies is key to improving trajectory tracking stability and operational quality. In terms of multi-machine collaboration, this paper discusses master–slave collaboration, distributed control, and task allocation modes. It further identifies that the stability of collaboration and interoperability standards between devices in weak network environments are currently the main bottlenecks limiting the large-scale application of this technology. Finally, this paper provides prospects for future research directions and suggests strengthening the closed-loop integration of perception, decision-making, and dynamic models, establishing industry unified standards, and enhancing the safety of the entire lifecycle of operations, providing suggestions for the unmanned application of agricultural robots. Full article

Karst regions are characterized by fragmented topography and significant micro-relief mosaics, leading to prominent spectral aliasing of land features, which can result in insufficient spatial representativeness of remote sensing samples for Land Use and Land Cover (LULC). The accuracy of LULC data directly affects the scientific basis of decision-making for rocky desertification control and ecological conservation. This study selected the Beipanjiang River Basin in Guizhou Province, a typical karst region, as the study area. The study selected the SOS, LOS, OM, and EOS indices from the 2001–2020 MODIS MCD12Q2 phenological dataset, combined with topographic zoning data. This study developed a sample spatial optimization scheme for complex karst terrain by integrating Spearman’s correlation analysis, SKATER spatially constrained clustering, statistical tests, adaptive stratified sampling, and Random Forest classification. The scheme was designed to test a phenology–landform joint stratification strategy for spatial sample allocation. The results indicate that (1) the study area was divided into six phenological pattern subregions, with significant spatial differentiation observed among them; (2) the “phenology–landform joint stratification + dual-weighted sample allocation” method was associated with improved sample representativeness and greater internal homogeneity within sample strata under the current experimental setting; and (3) compared to simple random sampling, the remote sensing phenological pattern-driven spatial optimization scheme improved overall accuracy from 71.33% to 77.55% and increased the Kappa coefficient from 0.43 to 0.62. These results suggest that, under the current study-area, sample-size, and validation settings, the phenology–landform joint stratification and dual-weighted allocation scheme can improve the spatial organization of training samples and classification performance over complex karst terrain, although weakly vegetated or bare classes remain difficult to separate. Full article

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

Managed aquifer recharge (MAR) is increasingly being realized as an important approach to improve water security in arid and semi-arid environments where there is a low amount of surface water and high climatic variability. This paper introduces a unified approach to the process of locating appropriate MAR locations and estimating recharge potential in Central Kazakhstan through a multi-criteria analysis using geographic information systems (GIS) and hydrogeological field exploration, water balance modelling. Remote sensing datasets and evapotranspiration (ET) analyses were conducted for the 2014–2024 period, while field investigations, infiltration tests, and hydrochemical sampling were performed during the 2025 field campaign. The suitability testing was preliminarily performed in the Google Earth Engine (GEE; Google LLC, Mountain View, CA, USA) environment as a weighted overlay test with the combination of terrain, vegetation, hydrological, and land cover parameters. According to the suitability map obtained and patterns of activity in agricultural activities, eleven candidate sites were identified, out of which eight were found to be suitable after hydrochemical analysis. The Nesterov and Boldyrev techniques of field-based infiltration tests produced a range of 0.05 to 1.42 m/day of hydraulic conductivity. Water balance analysis shows that the total amount of water that could potentially be added to groundwater recharge is about 40.2 million m³/year and that the effective amount of water could be recharged is about 11.0 million m³/year, which is limited by the infiltration processes. This means that about 27 percent of the available water is added into ground water recharge, which is a significant boost to the original estimates. The assessment of the storage capacity of the aquifers indicates that at all locations, the pore space is much greater than the recharge volumes that have been calculated and, therefore, storage is not a limiting factor in the implementation of MAR. It is estimated that the potential MAR rates range between 174 and 5282 m³/day depending on local hydrogeological conditions. The suggested method offers a powerful and generalizable site selection and measurement framework of MAR in arid areas with limited data. The findings highlight the significance of combining remote sensing, field measurements, and process-based modeling to aid sustainable groundwater management and climate adaptation strategies. Full article

Robotic visual perception and terrain recognition are critical for autonomous locomotion and adaptive control in complex environments. However, existing models often extract weak features, confuse classes, and deliver unstable recognition. Most prior studies use end-to-end convolutional networks or single-stream feature extraction, which limits the balance between fine-grained visual representation and adaptive discrimination of confusing samples. To solve this problem, this paper proposes a vision model that blends attention mechanisms with a confusion augmentation strategy. Using an improved ResNet50 backbone, we add a local feature sharpening module and a channel–spatial attention module to strengthen edge texture and global context representation. We also design a confusion augmentation strategy based on the similarity of hard samples. It generates mixed samples through cross-perturbation in feature space, thereby improving the discrimination of highly similar terrains. Experiments show that our model achieves an accuracy of over 98.19% on various terrains, including cement, asphalt, sand, and snow. t-SNE visualization and Grad-CAM analysis demonstrate clear class separability and good interpretability, confirming the effectiveness and robustness of the approach for robotic terrain recognition. Full article

Agricultural cold chain logistics is characterized by inherent challenges—product perishability, high carbon emissions, and stringent time windows—which are further exacerbated by dynamic disruptions. Existing methods suffer from slow adaptability, unstable multi-objective convergence, and severe cold-start issues. This work falls within the broad scope of biomimetics—the science of emulating nature’s time-tested strategies to solve complex engineering problems—and bio-inspired data-driven methods and their applications in engineering control, optimization, and artificial intelligence. The proposed H-MODRL framework embodies core biomimetic principles: the Genetic Algorithm (GA) mimics Darwinian natural selection and genetic inheritance, the Sparrow Search Algorithm (SSA) abstracts the cooperative foraging and anti-predation behaviors of sparrow populations in nature, and the Arrhenius-based freshness-decay model captures the biochemical kinetics governing perishable biological products. By synergistically integrating these biological evolution principles, swarm intelligence, and deep learning, the framework tackles real-world logistics complexity in a manner directly inspired by living systems. This study presents a well-organized hybrid optimization framework (H-MODRL) that couples a three-stage hybrid evolutionary mechanism, synergistically integrating heuristic warm-start, evolutionary policy guidance, and deep reinforcement learning decision-making. First, an improved genetic algorithm combined with the earliest deadline first strategy constructs a feasible initial population satisfying hard time-window constraints. Second, a large neighborhood search-enhanced chaotic sparrow search algorithm builds a high-quality elite guidance set for policy learning. Third, a physics-based multi-objective proximal policy optimization model embedded with Arrhenius equation-derived freshness-decay kinetics performs online decision-making. Experiments demonstrate that pre-computed all-pairs shortest paths and an O(1) hash-based dynamic-disruption indexing mechanism support fast online replanning. On heterogeneous simulated terrains based on real Chinese geospatial data, H-MODRL outperforms state-of-the-art algorithms across four objectives—logistics cost, carbon emissions, terminal freshness, and delivery time—while exhibiting compact, low-variance performance distributions, thereby validating its engineering robustness and practical value in complex agricultural cold chain environments. Full article

Cultural heritage in major river basins serves as an important spatial carrier of historical civilization evolution, and the spatial differentiation characteristics and influencing factors of its activation level are closely related to heritage conservation, utilization, and sustainable development. This study focuses on the Henan section of the Yellow River Basin and selects 344 cultural heritage sites as the research objects. A comprehensive evaluation system for cultural heritage activation was constructed from three dimensions—culture, society, and economy. By integrating GIS-based spatial analysis with the GWR model, the study reveals the spatial differentiation characteristics of cultural heritage activation levels and their influencing factors. The results indicate that the activation level of cultural heritage exhibits a dual-core-dominated and multi-level spatial agglomeration pattern. Zhengzhou and Luoyang function as dual high-density core clusters with elevated heritage activation levels, while a continuous cultural heritage corridor has gradually formed along Sanmenxia, Luoyang, Zhengzhou, Jiaozuo, Hebi, and Puyang. Furthermore, heritage agglomeration, heritage spatial radiosity, per capita GDP, transportation accessibility, terrain relief, and NDVI on the activation level of cultural heritage demonstrate significant spatial heterogeneity. Based on the identification of spatial heterogeneity, this study proposes a core–corridor–node spatial pattern and a factor-adaptive targeted strategy for cultural heritage activation. These findings provide a scientific basis for differentiated conservation and precise activation of cultural heritage under the national strategy of ecological protection and high-quality development in the Yellow River Basin, while also offering valuable insights for the collaborative governance of cultural heritage in major river basins worldwide. Full article

To address the challenges of time-dependent error divergence in Strapdown Inertial Navigation Systems (SINS) and the insufficient accuracy of traditional terrain matching algorithms in feature-sparse flat terrain environments, this paper proposes an intelligent terrain-aided navigation method integrating an Improved Osprey Optimization Algorithm (IOOA), Distribution Estimation, and Q-learning. Utilizing terrain information entropy as a robust matching metric, the algorithm establishes a two-phase evolutionary framework comprising Lévy flight-based random search (exploration phase) and elite-guided Gaussian Estimation of Distribution (exploitation phase). By introducing a Q-learning mechanism to adaptively regulate exploration parameters, an intelligent balance between population diversity and convergence speed is achieved. Under a unified computational benchmark, systematic multi-scenario simulations were conducted using datasets from simulated moderately undulating foothill terrain, the Libyan Sahara, and the real Digital Elevation Model (DEM) of the Junggar Basin in Xinjiang, China. Experimental results demonstrate that, compared to traditional TERCOM and mainstream swarm intelligence algorithms, the proposed algorithm drastically reduces positioning errors in the aforementioned complex terrains and significantly enhances matching accuracy. Robustness and real-time performance tests indicate that the algorithm achieves an average single-match processing time of only 0.08 s and maintains error variability as low as ±0.83 m under random perturbations. Furthermore, an ablation study confirms the necessity of the multi-strategy fusion mechanism in suppressing local optima entrapment and non-convergent oscillations. This study validates the engineering feasibility of the algorithm under conditions of low computational dependency, providing an effective technical approach for high-precision autonomous navigation in GPS-denied environments. Full article

Quadrupedal wheel-legged robots possess exceptional mobility in complex terrains, but their robust locomotion control is severely hindered by the difficulty of accurate state estimation without external sensors. Existing reinforcement learning methods relying on two-stage imitation often suffer from representation collapse and information loss during sim-to-real transfer. To address these challenges, this paper proposes a novel end-to-end reinforcement learning framework for implicit state estimation, incorporating terrain and external force features. Inspired by internal model control, the proposed method leverages a history of purely proprioceptive observations to extract explicit kinematic responses, as well as implicit environmental and external force representations via prototypical contrastive learning, completely circumventing explicit terrain regression and the need for physical force sensors. Furthermore, a tailored composite reward function and a progressive curriculum training strategy with large-scale domain randomization are integrated to ensure dynamic stability and hardware safety. Extensive cross-simulator validations and real-world deployments demonstrate that the approach achieves highly agile and robust locomotion, including adaptive traversal over diverse terrains. Experiments show that the method significantly enhances robustness under external disturbances, notably reducing the lateral linear velocity tracking error from 0.2421 m/s to 0.1319 m/s. The proposed method realizes zero-shot sim-to-real transfer with superior sample efficiency, providing a reliable and universal control paradigm for wheel-legged robots in unstructured environments. Full article

South Africa is highly susceptible to climate variability and long-term climatic shifts, necessitating a comprehensive understanding of changing extreme precipitation patterns to guide effective mitigation and adaptation responses. This study examined variations in extreme precipitation indices from 1981 to 2023 across the eastern Free State Province using daily rainfall records derived from the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS). Ten extreme precipitation indices were evaluated, with trend detection conducted through the Innovative Trend Analysis (ITA) technique. Findings indicate that the majority of municipalities exhibited statistically significant declining trends (p < 0.05) in total wet-day precipitation (PRCPTOT), R99P, R95P, the Simple Daily Intensity Index (SDII), CDD, RX5day, R20mm, and R10mm, suggesting an overall reduction in both heavy and moderate rainfall occurrences. In contrast, significant upward trends (p < 0.05) were identified in CWD, and RX1day, reflecting a shift toward prolonged wet periods and more intense short-duration rainfall events. Taken together, these divergent patterns point to the simultaneous emergence of heightened drought vulnerability driven by reduced cumulative rainfall and increased flood risk linked to intensified precipitation extremes. These results underscore the importance of forward-looking, climate-resilient water resource management and context-specific adaptation strategies suited to the eastern Free State’s complex mountainous terrain. Full article

Vegetative restoration in degraded landscapes requires scalable deployment strategies capable of adapting to heterogeneous terrain conditions. Conventional aerial seeding methods typically operate in open-loop mode, distributing seeds uniformly without considering terrain suitability. This study presents a modular, vision-guided unmanned aerial vehicle (UAV) system for real-time adaptive seed deployment based on the closed-loop integration of onboard perception and actuation under embedded computational constraints. The proposed system combines RGB-based terrain classification, embedded processing, and altitude-adaptive seed dispensing within a unified perception–decision–actuation framework, enabling selective and context-aware seed deployment during flight. Terrain suitability is evaluated onboard using three convolutional neural network (CNN) models and a color-based baseline to distinguish sowable and non-sowable areas. A confidence-based decision strategy with temporal filtering improves reliability, while an altitude-adaptive control mechanism regulates seed distribution across varying flight heights. Field experiments conducted in semi-arid environments demonstrate classification accuracy above 85% with inference latency below 100 ms on a Jetson Nano platform. Additional offline evaluation under varying altitude, speed, illumination, and terrain conditions confirms the robustness of the perception module. The results demonstrate the feasibility of integrating real-time perception with adaptive actuation, enabling UAVs to transition from passive sensing platforms to active agents for environmental intervention. The proposed system provides a practical and scalable approach for precision vegetative restoration in heterogeneous environments. Full article

Post-disaster rescue scenarios often involve complex and variable terrains, imposing heterogeneous mobility requirements on different transport modes. Single-type vehicles face challenges in independently completing comprehensive rescue tasks. This study addresses the critical problem of coordinating heterogeneous aerial and ground vehicles to collaboratively plan relay rescue routes. To tackle the NP hard multi-terrain, multi-vehicle, and multi-route path planning problem, we propose a New Adaptive Multi-Strategy Particle Swarm Optimization algorithm (AMS-PSO-NEW). The algorithm features a synergistic integration of differential evolution’s multi-strategy mutation, SHADE-based adaptive parameter control, population diversity monitoring with restart mechanisms, and multi-level local search. A sequential hybrid mechanism is designed in which DE-generated trial vectors serve as reference positions for PSO velocity updates, enabling balanced global exploration and local exploitation. By leveraging adaptive parameter tuning, success history memory, and diverse population maintenance, AMS-PSO-NEW effectively overcomes premature convergence and low accuracy issues typical in discrete combinatorial optimization using traditional PSO, achieving a balanced global exploration and local exploitation. Performance validation is conducted over six rescue scenarios varying in scale and complexity, benchmarking AMS-PSO-NEW against nine algorithms: PSO, GA, NSGA-II, GWO, DE, ABC, CS, Q-learning, and MIP. Results demonstrate superior performance across four metrics (rescue success rate, average rescue time, total cost, and fairness), with significant improvements in high-complexity environments. Full article