控制理论(社会学)
弹道
导线
计算机科学
控制工程
稳健性(进化)
运动控制
工程类
运动规划
粒子群优化
最优控制
再制造
轨迹优化
工作量
机器人
滑模控制
模型预测控制
鲁棒控制
控制系统
气缸
运动(物理)
控制器(灌溉)
控制(管理)
监督人
作者
Zhonglin Xia,Xiaoyan Liu,Guosheng Yang
标识
DOI:10.1016/j.jer.2026.09.025
摘要
Objective To address the severe disconnect between trajectory planning and control in flexible pneumatic robotic arms—caused by strong nonlinearity, pneumatic delays, and material flexibility—this study proposes a motion generation and sliding-mode control model based on particle swarm optimization and deep reinforcement learning. Method The model combines PSO and DRL for high-order continuous trajectory planning and introduces SBR-RRT for collision-free motion generation. Its main technical novelty is a hierarchical planning–control coupling mechanism that transfers optimized trajectory parameters as nominal references and continuously feeds back tracking, obstacle, and pneumatic states for online correction, thereby bridging trajectory generation and controller execution. Result The model achieves a 96.56% success rate in traversing static obstacles. The accuracy is 97.01% when performing dynamic target grasping tasks. The model achieves a computational workload of 125.31 MFLOPS and a control accuracy of 0.939. Conclusion The model demonstrates excellent capability in the trajectory planning and control of the manipulator. The model has smooth motion paths, as well as robustness and performance stability. The model offers a good technical solution for the stable motion of the manipulator.
科研通智能强力驱动
Strongly Powered by AbleSci AI