Adaptive Sampling-Based Moving Obstacle Avoidance for Cable-Driven Parallel Robots

Abstract

The ability to avoid the moving obstacle is critical to cable-driven parallel robots (CDPRs) operating in real-world environments. However, the moving obstacle raises extraordinary challenges for CDPRs due to various constraints introduced by cables. In this article, an adaptive sampling-based path planning method is presented for CDPRs to find collision-free trajectories under dynamic scenarios. In detail, the suggested method is based on the rapidly exploring random tree algorithm; it allows the robot to deal with complex constraints, such as cable collision and dynamic feasible workspace. To handle dynamic environments, partial motion planning with forward simulation is employed to construct an optimal time-based tree, and the predicted collision is checked in a closed-loop system. In particular, the swept volume caused by the time interval is considered, and the Gilbert–Johnson–Keerthi algorithm is applied for collision detection. Finally, a vision system with double webcams is designed to track the moving obstacle; the image processing based on morphological operations is introduced to segment the moving objects and automatically initialize a tracking box. Simulation results show that the proposed method can efficiently avoid the moving obstacle; these results are, then, validated experimentally using a built eight-cable-driven parallel robot with a drone as the moving obstacle.