Recursive genetic algorithm for robot manipulator motion planning in the existence of obstacles

Abstract

Robotic Manipulators motion planning may be challenging due to the high dimensionality of configuration space, complexity of some manipulators geometry, obstacles and physical constraints. In this paper, a recursive genetic algorithm (GA) for manipulators off-line motion planning in existence of obstacles is introduced. The proposed algorithm considers the kinematics prospective of the manipulator motion planning problem. It divides the motion planning problem into smaller sub-problems. Hence, it is capable of handling efficiently complex problems. In addition, the algorithm utilizes the multi-objectives optimization feature of GA to solve implicitly the Inverse Kinematics enabling the motion task to be assigned in high level Cartesian-coordinates. Hence, the user doesn't have to specify target joint angles. Moreover, using a recursive mechanism, the algorithm can be modified to handle online-motion planning. The algorithm is verified and tested through 3D simulation. © 2012 Springer-Verlag GmbH.