In this paper, a novel framework is proposed to incorporate task assignment, path planning, and tracking control of a multi-robot system. The dynamic task assignment of multi-robots is achieved using a self-organizing map based feature. The real-time collision-free robot path is generated from a neuro-dynamics network through sensor measurement and responding immediately to dynamic elements in the environment including the robot, the target, and obstacles. The tracking control is accomplished by a neuro-dynamics and back-stepping based model. This type of control is able to generate smooth, bounded acceleration control signals for a non-holonomic mobile robot to track the reference path generated by the path planner. Experiments under various situations demonstrated the effectiveness of this integrated system.
展开▼
