Research on motion planning and real-time obstacle avoidance of cooperative robot based on model predictive control

In this paper, the motion planning and real-time obstacle avoidance method of cooperative robot (Cobot) with model predictive control (MPC) is studied. The traditional motion planning method has a large amount of calculation in complex environment, and it is difficult to meet the millisecond response requirements of dynamic obstacle avoidance, and its ability to deal with unknown or dynamic obstacles is limited. Therefore, this paper proposes a motion planning method combining MPC and real-time obstacle avoidance, and realizes efficient and safe trajectory planning in complex environment through dynamic constraint optimization and rolling time domain strategy. This method adopts hierarchical control architecture, the global layer uses improved A* algorithm to generate the initial path, and the local layer uses MPC to dynamically adjust the trajectory. At the same time, the rolling time domain truncation and parallel computing strategy are introduced, which significantly improves the real-time performance. Experiments are carried out in ROS-Gazebo simulation environment and real robot platform. The results show that the proposed method has high obstacle avoidance success rate, trajectory smoothness and computational efficiency in both low-density and high-density dynamic obstacle environments, which verifies its adaptability and practicability in complex dynamic environments.