Design and Validation of New Obstacle Avoidance Scheme With Noise Suppression Capability for Redundant Robot Manipulators

Obstacle avoidance is a key issue encountered in the research of redundant robot manipulators. To avoid obstacle avoidance failure affected by noise, a novel velocity-level obstacle avoidance scheme with noise suppression capability is proposed. Based on an equality criterion that is robust against noise, with the incorporation of inequality constraint for obstacle avoidance and the joint physical constraints, an enhanced minimum-velocity-norm (EMVN) scheme is proposed and investigated for robotic redundancy resolution. The proposed EMVN scheme is further transformed and described as a quadratic programming (QP) problem analyzed at the joint velocity level and resolved using a recurrent neural network. Simulation results based on PA10 robot manipulator with window-shaped obstacle in the existence of zero, constant, and time-varying bounded noises demonstrate the effectiveness, robustness, and superiority of such a scheme for obstacle avoidance with noise suppression capability of redundant robot manipulators.