The YOLO series of algorithms are commonly used in the field of object detection, and YOLOX, as one of them, is an improvement on YOLOv5, which improves the detection accuracy but decreases the inference speed. In order to further improve the detection accuracy of YOLOX and solve the problem of slow inference speed, the YOLOX-DBB algorithm is proposed on the basis of YOLOX-S with reference to the idea of structural re-parameterization. Experimental results on the VOC dataset showed that YOLOX-DBB achieved 62.88% mAP0.5:0.95, a 3.47% improvement over YOLOX-S; meanwhile, YOLOX-DBB achieved 82.46% mAP0.5, an improvement of 1.02%. In inference speed, YOLOX-DBB takes only 10.36(ms) to infer a picture, while YOLOX-S takes 12.55(ms) in the same environment. Therefore, the new network has improved detection accuracy, and the inference speed is also faster than the original network. Finally, YOLOX-DBB was applied to the art craft surface defect detection task, and the results showed that it still had better indicators than YOLOX-S.
Aiming at the problems of difficult hard-constraint enforcement and weak environmental generalization ability in the safe trajectory planning of manipulators in complex environments, a Policy-Guided Model Predictive Path Integral (PG-MPPI) planning framework is proposed. This framework integrates the advantages of reinforcement learning and model predictive control to construct a global prior guidance, local real-time optimization and hard-constraint safety assurance: a Constraint-Discounted Soft Actor–Critic (CD-SAC) offline learning policy is designed, which incorporates the configuration-space distance field as a safety guidance term to realize the learning of obstacle avoidance behavior; the offline policy is used to guide the online sampling and optimization of MPPI, improving sampling efficiency and planning quality; and a Control Barrier Function (CBF) safety filter is introduced to revise control commands in real time, ensuring the strict satisfaction of constraints. Taking the SIASUN T12B manipulator as the research object, simulation comparison experiments are carried out in multi-obstacle scenarios. The results show that the PG-MPPI algorithm outperforms the comparison algorithms in the success rate of collision-free target reaching, ensures the smoothness and feasibility of the trajectory, and has a good adaptive capacity to complex environments with unknown obstacle configurations, thus providing an efficient solution for the autonomous and safe operation of manipulators.
Aiming at the problems of difficult hard constraint enforcement, weak environmental generalization ability in the safe trajectory planning of manipulators in complex environments, a Policy-Guided Model Predictive Path Integral (PG-MPPI) planning framework is proposed. This framework integrates the advantages of reinforcement learning and model predictive control to construct a global prior guidance, local real-time optimization and hard constraint safety assurance: a Constraint-Discounted Soft Actor-Critic (CD-SAC) offline learning policy is designed, which incorporates the configuration-space distance field as a safety guidance term to realize the learning of obstacle avoidance behavior; the offline policy is used to guide the online sampling and optimization of MPPI, improving sampling efficiency and planning quality; a Control Barrier Function (CBF) safety filter is introduced to revise control commands in real time, ensuring the strict satisfaction of constraints. Taking the SIASUN T12B manipulator as the research object, simulation comparison experiments are carried out in multi-obstacle scenarios. The results show that the PG-MPPI algorithm outperforms the comparison algorithms in the success rate of collision-free target reaching, ensure the smoothness and feasibility of the trajectory, and has a good adaptive capacity to dynamic environments, thus providing an efficient solution for the autonomous and safe operation of manipulators.