Hybrid Obstacle Avoidance and Repetitive Motion Scheme of Mobile Robots Synthesized by Varying-Parameter Dynamic Learning Networks

In order to effectively avoid obstacles and resolve non-repetitive problems of mobile redundant manipulators, a hybrid obstacle avoidance and repetitive motion (HOARM) scheme is proposed, analyzed, and applied, which is synthesized by varying-parameter dynamic learning networks (VPDLN). Specifically, to guarantee the efficient and accurate tasks, a periodic relaxation constrained repetitive motion scheme and an obstacle avoidance approach are integrated as a quadratic criterion, then a bicriteria scheme is constructed by combining the schemes and kinematic models of mobile redundant manipulators, and is further deduced as a quadratic programming (QP) problem through a dynamic weighting method. Then a varying-parameter dynamic learning networks method is applied to solve the proposed QP problem. Furthermore, a path-tracking task is conducted to demonstrate the accuracy and efficiency of the proposed HOARM scheme.