Vision-based moving target interception with a mobile robot based on motion prediction and online planning

Abstract

In this paper, a vision-based moving target interception approach is proposed based on motion prediction and online path and velocity planning. For motion prediction, a fusion method of two motion prediction approaches with historical motion data is proposed to forecast the robot motion. With the predicted target pose as the subgoal, an optimal path is selected from a set of the third order Bezier curves based on their length and curvature cost. Finally, the control input can be calculated through the curvature function from the generated path. Simulation results show the effectiveness of the proposed method. The seamless integration of motion prediction and online planning makes the moving target interception more efficient and predictable.