Real time obstacle avoidance for redundant robot

Abstract

Real-time obstacle avoidance for redundant robot is always of consequence in the field of robot research. According to the mechanism of the human arm movement in obstacle avoidance, an artificial safety zone around the obstacle in the operational space is defined. Based on the artificial safety zone, a unique technique of real-time collision avoidance for position controlled robot is proposed in this paper. The measure of obstacle avoidance in the Cartesian space is available once the joint intrudes into the region. On the basis of the mapping relationship between the Cartesian space and the joint space, an algorithm of real-time obstacle avoidance by using MoorePenrose inverse is provided. First of all, a simulation with a 3 DOF (degrees of freedom) planar manipulator is presented. Then two supportive online experiments of real-time obstacle avoidance are conducted and the experimental data are analyzed as well. The simulation and experimental results demonstrate the efficiency as well as the real-time performance of the algorithm. Furthermore, this method is also applicable to obstacle avoidance of dynamic environment, in which the target and obstacles are permitted to move.