Formation control of unmanned surface vehicles (USVs) in complex marine environments is required to contend with strongly coupled, high-dimensional disturbances. A Multi-Agent Active Disturbance Rejection Control (MAADRC) framework is developed for this purpose. The design centers on a distributed extended state observer (DESO) coupled with a dual-channel feedback structure—NEFL-GCO and LGL-FC—that collectively maintains formation geometry. Three main ideas underpin the approach. First, a bandwidth-efficient distributed observation scheme enables agents to share disturbance estimates while using substantially less communication bandwidth. Second, an adaptive consensus compensation mechanism accommodates parameter variations as formations evolve. Third, a formation-compatible obstacle avoidance algorithm enhances reliability in congested waters. To evaluate the control structure and optimize its parameters, a multi-agent reinforcement learning (MARL) method—specifically Multi-Agent Proximal Policy Optimization (MAPPO)—is employed. The MARL agent tunes two critical parameters: observer bandwidth and nonlinear feedback gain, thereby establishing a performance baseline. After ten million training steps, the MAPPO-optimized MAADRC achieves a tracking root-mean-square error (RMSE) of 1.18 m. This value lies within 3% of the manually tuned result of 1.21 m, indicating that the bandwidth parameterization is near-optimal. Extensive simulations incorporating realistic wind, wave and current disturbances demonstrate a dynamic obstacle avoidance success rate maintaining an expected level, alongside consistently low formation tracking errors. Collectively, these findings confirm the resilience and practical utility of the proposed framework in demanding maritime settings.

Autonomous obstacle avoidance is a critical capability for Autonomous Underwater Vehicles (AUVs) to operate safely in dynamic and uncertain marine environments. Traditional AUV control methods rely on precise physical modeling and preset rules, yet they struggle to adapt to multiple sources of uncertainty, such as random initial states, dynamic obstacles, and varying currents. In recent years, deep reinforcement learning has provided a new avenue for data-driven adaptive policy learning. However, it remains insufficient for handling long-horizon tasks with sparse rewards. While hierarchical reinforcement learning can mitigate reward sparsity through temporal abstraction, it often faces challenges including exploration–exploitation imbalance, slow global convergence, and insufficient safety guarantees. Furthermore, most existing studies neglect dynamic environmental disturbances and task continuity, which further limits the practical application of these algorithms. To address these challenges, this paper proposes a hierarchical curiosity-driven AUV obstacle avoidance algorithm (HDAO), designed for autonomous obstacle avoidance in dynamic and uncertain underwater environments. The core design of HDAO incorporates several key innovations. Firstly, it introduces a Collision Threat Index for dynamic obstacles, which enables explicit risk perception and quantifies collision threats, thereby enhancing the policy’s generalization and robustness. Secondly, a task-decoupled hierarchical architecture is employed to synergistically optimize global path planning and local obstacle avoidance behaviors. This approach effectively manages long-horizon navigation tasks while alleviating high-dimensional training pressure. Finally, a novel reward mechanism is designed by integrating hierarchical active exploration with curiosity-driven passive exploration. This mechanism effectively incentivizes the agent to explore unvisited areas under sparse reward conditions and dynamically balances exploration and exploitation. Experimental results demonstrate that HDAO significantly outperforms existing methods in terms of obstacle avoidance success rate, training convergence speed and robustness against external disturbances.

This paper proposes a dynamic obstacle-avoidance algorithm for unmanned surface vehicles (USVs) that combines a Fuzzy-enhanced Dynamic Window Approach (FDWA) with an Improved Bidirectional A*–Improved Grey Wolf Optimizer (IBA*–IGWO) framework. Firstly, the traditional dynamic window method (DWA) is improved by adopting an initial heading angle optimization strategy to reduce the heading deviation of unmanned vessels during cruising. Secondly, a fuzzy controller is introduced, which can adaptively adjust the weight coefficients in the cost function of the DWA algorithm based on the current position of the unmanned vessel, surrounding environmental information, etc., to improve obstacle avoidance ability and adaptability in different environments. Finally, using the global static cruise route provided by the IBA*–IGWO algorithm, key nodes are selected as local endpoints for the FDWA algorithm to ensure that the unmanned vessel can perform cruise tasks according to the optimal plan during navigation and make dynamic adjustments in case of emergencies. The simulation results demonstrate the feasibility of the proposed method in handling unknown and dynamic obstacles under the current grid-based experimental settings, while enabling the USV to return to the pre-planned global route after local obstacle avoidance. These results provide a basis for further development toward more robust and rule-aware autonomous navigation in realistic maritime environments.

Obstacle avoidance is a fundamental requirement for safe and reliable robot operation. Due to the lack of techniques capable of efficiently and accurately computing volumetric distances and their derivatives, many obstacle avoidance frameworks neglect the volumetric characteristics of objects by using points or spheres as the choice of object representation. To overcome this, an optimisation-based method is used in this paper to accurately compute the minimum distance between convex shapes. In order to enable integration with gradient-based obstacle avoidance algorithms, a method that calculates the gradient of the minimum distance is introduced. This information is incorporated into a Quadratic Program to demonstrate the effectiveness of the proposed approach for online obstacle avoidance. In contrast to methods that rely on closed-form expressions, the proposed framework leverages optimisation to simultaneously compute pairwise distances between multiple convex shapes. Simulation results, averaged over 1000 runs, demonstrate that the proposed approach completes tasks more accurately and consistently, improving trajectory tracking accuracy by up to 35% compared to the state-of-the-art baseline. The proposed framework is further validated on a physical robot system, confirming its practical applicability and robustness on real hardware.

Development and application of a dynamic obstacle avoidance algorithm for small fixed-wing aircraft with safety guarantees

A dynamically hybrid path planning and obstacle avoidance algorithm for mobile robots based on improved A-star and dynamic windows approach

Multi-rotor vehicles are always facing the threat of collision due to the nature of low altitude flight, common solutions are perceptual obstacle avoidance algorithms and propeller protection structures, however these also fail in the event of unavoidable collisions therefore, it is necessary to study the recovery of the vehicle after a collision. In this paper, a low-cost and fast collision recovery framework is proposed, which uses extended state observer (ESO) to estimate the wrench caused by a collision and then performs collision detection and determines the type of collision based on the vehicle state. Inspired by the recovery manoeuvre of mosquitoes hit by raindrops, three new bionic-based collision recovery strategies: rolling with the trend, admittance control and collision escape are proposed to cope with different types of collisions, which enable the vehicle to present suppleness like a mosquito after a collision and mitigate the impact of the collision, and the applicable scenarios and implementation details of each collision recovery strategy are elaborated. These strategies solve the problem of active and passive collisions for most vehicles. Finally, a series of convincing collision experiments for both under-actuated and over-actuated vehicles are designed, and the vehicles are able to perform collision detection and collision recovery reasonably, and the experimental results verify the effectiveness of the collision recovery strategies.

Abstract Path planning and tracking control play a crucial role in the safe driving of an autonomous vehicle. To enhance the safety, comfort, and energy efficiency of the vehicle during obstacle avoidance, a path planning and tracking method considering road space constraints and safety constraints is proposed. Firstly, a path planning algorithm for obstacle avoidance was designed, which takes the safety distance between vehicles and the constraints of road width into account, ensuring safe obstacle avoidance during the process, while reducing energy consumption and enhancing comfort; Second, a three-degree-of-freedom simplified vehicle dynamics model considering real-time slip rate is constructed, improving the dynamic accuracy of the vehicle simulation model under different road conditions; Finally, a nonlinear time-varying MPC is designed based on the dynamics model. The results show that in various scenarios, compared with the path planning methods based on obstacle avoidance function and artificial potential field, the safety has been improved by an average of 10.06%, the comfort has been enhanced by an average of 24.45%, and the average energy consumption has been reduced by 14.73%.

A Dynamic Obstacle Avoidance algorithm for unmanned aerial vehicles based on Predictive Velocity Obstacles

Research and application of obstacle avoidance algorithm for hydropower station inspection robot based on spatio-temporal networks

Introduction To address the core challenges of inaccurate fruit occlusion localization and inefficient robotic arm dynamic obstacle avoidance in complex, unstructured agricultural environments, this study proposes an integrated algorithm for harvesting. Methods The proposed algorithm is built upon an improved YOLOv8 model and the BIT* planner. The YOLOv8 model was enhanced by introducing the Swin Transformer module to improve multi-scale feature fusion and global context modeling. The BIT* planner was integrated with a BiLSTM network to endow it with dynamic obstacle prediction capabilities, thereby constructing a unified architecture for visual perception and motion planning. Results Experimental results demonstrated that the algorithm achieved real-time performance with a processing frame rate of 32.7 fps and an inference time of 32.6 ms for target localization, with a localization error standard deviation as low as 1.70 mm. In obstacle avoidance planning, it achieved a balance with manipulator energy consumption of 124.58 J, while controlling the computational load and memory resource consumption per task to 22.7 GFlops and 187 MB, respectively. Discussion This approach provides a high-precision, low-energy-consumption cooperative control solution for agricultural harvesting robots, advancing the practical application of automated fruit and vegetable harvesting.

This study presents a Tau-theory-based guidance methodology for obstacle avoidance in low-altitude, high-speed rotorcraft operations, especially within obstacle-dense and degraded visual environments (DVEs). A geometric approach is employed to develop the obstacle avoidance algorithms. The methodology considers both inner-loop and outer-loop guidance with a decision logic that determines the appropriate maneuver (turn, climb, or deceleration) based on real-time analysis of the environment and the helicopter’s operational limits. Extensive desktop simulations conducted in the MATLAB/Simulink environment, using FLIGHTLAB® high-fidelity nonlinear models and different pilot models, demonstrate the method’s ability to guide pilots with safe and efficient trajectories for obstacle field navigation. These findings lay the groundwork for potential real-world implementations in both manned and autonomous rotorcraft missions.

Dynamic obstacles in high-speed scenarios bring a challenge to the effectiveness and adaptation of active obstacle avoidance algorithm due to the time-varying characteristics. To realize safe collision avoidance in real time, a lateral path planning and tracking method is presented based on the parabola tentacle algorithm and Model Prediction Control (MPC) method. The merits of the proposed method lie in its ability to overcoming the lack of dynamics constraints in tentacle algorithm and avoiding path replanning facing the decoupling geometric and dynamic constraints. Firstly, the improved tentacles with parabola structure fan out with different curvatures discretizing the basic driving options of the vehicle. Additionally, a transition path from the tentacle path segment to road centerline is calculated by quintic polynomial to improve the smoothness of the reference path. To remedy the deficiency of the tentacle algorithm failing to ensure vehicle stability in high-speed scenarios, dynamics constraints are considered by the incorporation of MPC in the path execution level. The tentacle satisfying the safety and comfort constraints is executed by the MPC-based controller. Finally, the superiority of parabola tentacle is highlighted by the comparison with circle and clothoid tentacles. Furthermore, hardware-in-the-loop (HIL) experiments are carried out in two scenarios to verify the effectiveness of the proposed lateral obstacle avoidance algorithm. The results show that the proposed method can achieve safe and real-time dynamic obstacles avoidance as well as favorable path tracking accuracy.

Abstract Small-scale magnetic robots that can assemble, disassemble, and propel under globally applied magnetic fields can be versatile modular subunits for manufacturing and in vivo operations. This paper presents a magnetic cuboid robot that contains assembled cubes with encapsulated, freely-rotating permanent magnets. This minimalistic and scalable design enables magnetic cubes to assemble under magnetic fields into a cube chain that can propel using pivot-walking locomotion. The magnets for propulsion are evenly distributed between the cubes, but individual cubes can only move when joined with at least one other. A vision-based closed-loop controller that modulates the cuboid robot’s position and orientation during pivot walking is presented. The controller is simulated to navigate cuboid robots to user-selected goal locations. A Breadth-First Search (BFS) path-planning algorithm for obstacle avoidance is used to generate optimal paths for closed-loop pivot walking. Two physical workspaces are tested, one with a large free space and the other with a maze. Experiments and simulations demonstrate that magnetic cuboid robots can navigate in complex mazes and selectively self-assemble into cube chains while following the optimal path generated by the motion planner with visual feedback control.

The fuzzy inference system for obstacle avoidance developed in this paper is designed for NAO humanoid robot. The fuzzy obstacle avoidance (Fuzzy OA) has been tested in Webots virtual environment and the results showed that this method is almost two times faster than the Naoqi framework obstacle avoidance (Naoqi OA) while the robot is much more stable. Because the fuzzy inference system is a method that relies on trial an error and experience, the obstacle avoidance algorithm is subject to improvements. Future developments will take into account these results and will add other fuzzy inference systems for navigation, in order to get more autonomy for Nao robot.

This paper addresses the issues of target unreachability and local optima in traditional artificial potential field (APF) methods for UAV swarm path planning by proposing an improved collaborative obstacle avoidance algorithm. By introducing a virtual target position function to reconstruct the repulsive field model, the repulsive force exponentially decays as the UAV approaches the target, effectively resolving the problem where excessive obstacle repulsion prevents UAVs from reaching the goal. Additionally, we design a dynamic virtual target point generation mechanism based on mechanical state detection to automatically create temporary target points when UAVs are trapped in local optima, thereby breaking force equilibrium. For multi-UAV collaboration, intra-formation UAVs are treated as dynamic obstacles, and a 3D repulsive field model is established to avoid local optima in planar scenarios. Combined with a leader–follower control strategy, a hybrid potential field position controller is designed to enable rapid formation reconfiguration post-obstacle avoidance. Simulation results demonstrate that the proposed improved APF method ensures safe obstacle avoidance and formation maintenance for UAV swarms in complex environments, significantly enhancing path planning reliability and effectiveness.

Path planning generates a shorter path from source to destination based on sensor information acquired from an environment. An obstacle avoidance is an important task in robotics within path planning since the automatic functioning of robots requires reaching the destination without collisions. Moreover, obstacle avoidance algorithms have an important part in robotics. The existing algorithms did not enable robots to navigate their environments effectively, lessening the threat of collisions and preventing obstacles. Here, an Adaptive Spider Wasp Optimizer (ASWO) is introduced for path planning in mobile multi-robots. Initially, the simulation of an environment utilizing multiple robots and targets along with obstacles is accomplished. Thereafter, multi-objectives namely path smoothness, obstacle avoidance, and path length are considered. Lastly, path planning is conducted employing ASWO by considering fitness parameters such as path smoothness, obstacle avoidance, and path length. However, ASWO is designed by integrating adaptive concept with Spider Wasp Optimizer (SWO). In addition, ASWO achieved maximal value of fitness and path smoothness about 1.795 and 91.121% as well as minimal value of path length about 897.883 km.

This paper proposes a novel framework for autonomous unmanned aerial vehicle (UAV) navigation in complex environments, seamlessly integrating Theta* for global path planning with a simplified modulated velocity obstacle avoidance (MVOA) algorithm for local obstacle avoidance. Theta* generates optimal, smooth paths, while MVOA processes 2D LiDAR data as a single obstacle block to compute modulated velocities, enabling efficient avoidance of static and dynamic obstacles with minimal computational overhead. Compared to MVOA-only navigation, the integration of Theta* and MVOA produced shorter trajectories and faster mission completion with smoother velocity adjustments, demonstrating clear improvements in efficiency and stability. Simulation results show the framework maintains a 0.6 m safety distance and operates at 10 Hz, underscoring its robustness and reliability. The resulting control velocity is transmitted to an ArduPilot-based flight controller via MAVLink, ensuring precise, real-time execution. The current implementation focuses on 2D navigation in a planar environment as a foundation for future 3D expansion, with all results obtained through high-fidelity simulation. Building on these findings, the framework shows strong potential for real-time applications such as swarm UAV coordination, terrain surveying, and indoor navigation, offering a scalable solution for autonomous systems in dynamic settings.

Adaptive obstacle avoidance algorithm for wave gliders in dynamic marine environments based on improved DAPF with multi-model prediction

This paper introduces a spatiotemporally coupled coordination framework to address the limitations of conventional navigation strategies for UAV swarm operations, particularly with respect to dynamic obstacle avoidance and formation stability. The approach integrates an improved A* algorithm with dynamic potential field fusion, thereby ensuring global path optimality and local control feasibility. A sliding mode predictive control (SMPC) scheme is developed to guarantee exponential convergence of formation errors. Extensive experimental evaluations demonstrate an 81.5% reduction in path redundancy, a 100% success rate in dynamic obstacle avoidance, and an 86.5% enhancement in trajectory smoothness. These findings validate the superior effectiveness of the proposed method in complex and dynamic environments.