The Adaptive Path Planning Research for a Shape-Shifting Robot Using Particle Swarm Optimization

A shape-shifting robot with diverse configurations, named "AMOEBA-I" has been developed for search and rescue tasks. The accessibility of this robot to uneven environments was efficiently enlarged by changing its configurations. In this paper, for optimizing the path of the robot through an environment containing static polygonal obstacles, we present an adaptive path planning method for the AMOEBA-I, which integrates the reconfigurable ability of the robot into the particle swarm optimization (PSO). The unique accessibility of the AMOEBA-I is thus fully displayed. The modified PSO method can reduce the computational load and has certain degree of robustness on selecting dimension of particles. It can generate smooth paths by string of cubic splines. Experiment results show that the robot can change its configurations to perform the adaptive path planning corresponding to the environmental variation and current reconfigurable ability of AMOEBA-I. As a result, the path length has been reduced successfully and effectively.