A Novel Time-varying Path Planning Method for Dynamic Obstacles

Abstract

In the research of snake-like manipulator and mobile robot obstacle avoidance, most researchers only plan the path for static obstacles. However, there are many dynamic obstacles in the real world. Therefore, this paper proposes a general method that can plan the real-time changing path for the dynamic obstacle avoidance of moving objects such as quasi-continuous manipulators. The size, position, and other information of the obstacle are extracted by point cloud or image processing. The bounding box method is selected according to the shape of the obstacle so that the obstacle is equivalent to some basic shapes. For the dynamically updated bounding box, the corresponding continuous time-varying path/function is formulated, so that real-time obstacle avoidance can be realized. By analyzing the constraints that the continuous function that characterizes the time-varying path needs to be satisfied, the ultimate continuous function is selected by formulating the optimization objective. For various types of dynamic obstacles, the corresponding time-varying path can be planned, and the quasi-continuous manipulator moving along this path can realize real-time obstacle avoidance. Simulation results demonstrate the effectiveness of the proposed method.