Learning Motion Trajectories and Visual Commands of a Nonholonomic Mobile Robot Using Metaheuristic Technique

Abstract

The hybrid mobile robot control algorithm consists of two independent control loops. By developing two control phases, the transportation task is separated into two parts: movement from the initial position to a position at a great distance from the machine tool (global control) and movement from this position to the machine tool/intermediate point (local control). The original control system based on epipolar geometry as well as on learning of motion trajectories and visual commands was implemented on the Khepera II nonholonomic mobile robot (with additional equipment: KheGrip gripper and Prestigio PWC 2 WEB camera) by using metaheuristic technique in a laboratory model of the technological environment.