Kinematic performance-based path planning for cable-driven parallel robots using modified adaptive RRT*

Abstract

Cable-driven parallel robots (CDPRs) have demonstrated a remarkable potential for a wide range of applications over the past few decades. However, multiple cables introduce various interferences, which is a challenging task for the path planning of CDPRs. In this study, an adaptively optimal rapidly exploring random tree (RRT*) is proposed for CDPRs to generate collision-free paths in cluttered environments. The proposed method considers the kinematic performance of CDPRs by maximizing the feasible wrench volume and optimizing the dexterity of the robot. Furthermore, a sampling method was developed to guide the tree growth efficiently, by adaptively adjusting the sampling space. In the sampling space, the adaptive forward and backward areas are defined for biased sampling, and the probability weight of both areas is determined by the failure rate of the current node expansion. Finally, a post-processing method was adopted to obtain a shorter and smoother path. Simulations and experiments were carried out to verify the proposed path planning method on a self-built 8-6 (8 cables with 6 degrees of freedom) CDPR. The results indicate that the proposed method has much better efficiency than the conventional RRT* method, and the kinematic performance of the CDPR can be guaranteed.