Trajectory planning of manipulator for a hitting task with autonomous incremental learning

Abstract

A new approach based on genetic algorithm (GA) and autonomous mental development for trajectory planning of robot manipulator is presented in this paper. The trajectory for the manipulator is optimized by the GA. To make the trajectory planning used in real time application, a developmental learning algorithm is proposed to generate an incremental hierarchical discriminating regression (IHDR) tree to form the mapping from the state space to the action space. Just like the human brain from infancy to adulthood, the algorithm develops its cognitive and behavioral skills through online learning from the samples obtained from the GA-based method. When the IHDR tree is generated, it can perform the trajectory planning in real time by retrieving in its knowledge database.