In a future intelligent factory, a robotic manipulator must work efficiently and safely in a Human–Robot collaborative and dynamic unstructured environment. Autonomous path planning is the most important issue which must be resolved first in the process of improving robotic manipulator intelligence. Among the path-planning methods, the Rapidly Exploring Random Tree (RRT) algorithm based on random sampling has been widely applied in dynamic path planning for a high-dimensional robotic manipulator, especially in a complex environment because of its probability completeness, perfect expansion, and fast exploring speed over other planning methods. However, the existing RRT algorithm has a limitation in path planning for a robotic manipulator in a dynamic unstructured environment. Therefore, an autonomous obstacle avoidance dynamic path-planning method for a robotic manipulator based on an improved RRT algorithm, called Smoothly RRT (S-RRT), is proposed. This method that targets a directional node extends and can increase the sampling speed and efficiency of RRT dramatically. A path optimization strategy based on the maximum curvature constraint is presented to generate a smooth and curved continuous executable path for a robotic manipulator. Finally, the correctness, effectiveness, and practicability of the proposed method are demonstrated and validated via a MATLAB static simulation and a Robot Operating System (ROS) dynamic simulation environment as well as a real autonomous obstacle avoidance experiment in a dynamic unstructured environment for a robotic manipulator. The proposed method not only provides great practical engineering significance for a robotic manipulator’s obstacle avoidance in an intelligent factory, but also theoretical reference value for other type of robots’ path planning.
To address challenges in agent path planning within complex environments—particularly slow convergence speed, high path redundancy, and insufficient smoothness—this paper proposes KDB-RRT*, a novel algorithm built upon RRT.* This method integrates a bidirectional search strategy with a three-layer optimization framework: ① accelerated node retrieval via KD-tree indexing to reduce computational complexity; ② enhanced exploration efficiency through goal-biased dynamic circle sampling and a bidirectional gravitational field guidance model, coupled with adaptive step size adjustment using a Sigmoid function for directional expansion and obstacle avoidance; and ③ trajectory optimization employing DP algorithm pruning and cubic B-spline smoothing to generate curvature-continuous paths. Additionally, a multi-level collision detection framework integrating Separating Axis Theorem (SAT) pre-judgment, R-tree spatial indexing, and active obstacle avoidance strategies is incorporated, ensuring robust collision resistance. Extensive experiments in complex environments (Z-shaped map, loop-shaped map, and multi-obstacle settings) demonstrate KDB-RRT’s superiority over state-of-the-art methods (Optimized RRT*, RRT*-Connect, and Informed-RRT*), reducing average planning time by up to 97.9%, shortening path length by 5.5–21.4%, and decreasing inflection points by 40–90.5%. Finally, the feasibility of the algorithm’s practical application was further verified based on the ROS platform. The research results provide a new method for efficient path planning of intelligent agents in unstructured environments, and its three-layer optimization framework has important reference value for mobile robot navigation systems.
In Vehicle-to-everything (V2X)-enabled autonomous driving systems, collaborative perception allows multiple agents to exchange complementary information, thereby enlarging the perceptual range and alleviating occlusion. However, in real-world deployments, pose estimation errors and measurement uncertainties can undermine cross-agent feature consistency. Even after global geometric alignment, the fused bird’s-eye-view (BEV) representation may still suffer from residual local spatial misalignment, contextual inconsistency, and unstable feature responses, which degrade the robustness and localization accuracy of collaborative 3D object detection under complex perturbations. To address this issue, we propose a pose-aware BEV feature refinement method for post-fusion BEV representations. The proposed method progressively improves fused BEV features through three stages: local spatial compensation, multi-resolution contextual enhancement, and reliability-aware response filtering. In this way, residual post-fusion errors can be further suppressed without introducing additional supervision. Built upon the OpenCOOD framework and the PointPillars detector, the proposed method is evaluated on the DAIR-V2X dataset under both noise-free and perturbed settings with Gaussian and Laplacian pose noise. Compared with the strong baseline CoAlign, it improves AP@0.5 and AP@0.7 by 3.5 and 3.8 percentage points, respectively, under noise-free conditions. Under severe pose perturbations (0.6 m, 0.6°), it further improves AP@0.7 by 1.3 and 1.4 percentage points under Gaussian and Laplacian noise, respectively. Ablation studies and qualitative results further verify the effectiveness of the proposed design. These results demonstrate that refining post-fusion BEV features is an effective way to mitigate residual feature-level errors after global alignment and improve the detection robustness and localization accuracy of collaborative 3D object detection under pose uncertainty.
In unmanned aerial vehicle (UAV)‐based road crack detection, fine‐grained targets and complex backgrounds present a sharp trade‐off: lightweight models suffer from insufficient accuracy, whereas high‐accuracy models fail to meet real‐time requirements. To address this issue, we propose YOLO26‐GE, a lightweight detector that synergistically integrates GSConv (ghost shuffle convolution) and efficient multi‐scale attention (EMA) into the YOLO26 framework. By replacing standard convolutions with GSConv, we eliminate computational redundancy, reducing the model to only 2.5 million parameters and 5.1 GFLOPs while preserving fine‐grained crack texture. We further embed the EMA module to dynamically amplify crack‐related feature channels and suppress background noise, thereby enhancing the discriminability of weak‐feature targets. Evaluated on the public UAV PDD2023 dataset, YOLO26‐GE achieves an inference speed of 714 FPS, with an F 1 score of 0.724 ± 0.015, mAP50 of 0.722 ± 0.008, and mAP@50:95 of 0.415 ± 0.005. These results substantially surpass the baseline YOLO26 and other models. Together, they demonstrate that YOLO26‐GE effectively combines lightweight design with high accuracy in real‐time UAV‐based road crack detection.
This paper aims to address the path planning problem of autonomous mobile robots in complex environments. To achieve this goal, a hybrid DWA algorithm integrating an improved A* algorithm with fuzzy inference logic is proposed. Specifically, the improved A* algorithm is used to rapidly generate an initial feasible path, the rolling window mechanism is adopted to select sub-target points, and the evaluation weights are dynamically adjusted through a fuzzy controller. The research findings show that the improved A* algorithm has significant enhancements in search efficiency and path smoothness across different maps; meanwhile, the proposed improved hybrid algorithm exhibits notable improvements in runtime and average curvature, and can stably handle complex environments. The originality of this paper lies in the proposed DWA hybrid algorithm based on the improved A* integrated with fuzzy inference logic, which enables the local planner to flexibly adjust weight parameters in response to environmental changes relying on reasonable global path guidance.