Online Trajectory Generation for Multiple Robots Using Discrete Reference Points in Cluttered Environment

This paper presents a new online trajectory generation method that enables multiple robots to navigate through the cluttered environment. The proposed method first employs a centralized discrete planner to obtain all initial discrete path points with superior deadlock-free performance. And then each robot iteratively updates its subgoals online based on the update strategy and optimizes the trajectory segment by segment until all robots reach the target configuration. The online intermediate subgoal update strategy balances robot body safety and task execution efficiency based on discrete trajectory information and communication exchanges, providing better adaptation to cluttered industrial environments and faster navigation. In addition, the robot's discrete path point information can effectively avoid unsolvable deadlocks and enable the robot to perform transport tasks more efficiently. Finally, simulation experiments validate the effectiveness of the proposed method.