Path planning for redundant manipulators using metaheuristic for bilevel optimization and maximum of manipulability

In this paper, we propose a novel path planning method for a robot manipulator in the case of several obstacles. The path is generated by steps. At each step, a new position of the end effector is found in the Cartesian space. The redundancy of the manipulator, imposes that this position can be found by an infinity of configurations in the joint space. Thus, we search the best configuration which allows to avoid obstacles and singularities of the robot. This method is based on using the bilevel genetic algorithm to solve this problem. Simulation results showed the effectiveness of the proposed approach.