Real-time torque control of nonholonomic mobile robots with obstacle avoidance

In this paper, a novel torque controller is presented for nonholonomic mobile robots with obstacle avoidance. In the proposed controller, based on the artificial potential fields technique, an obstacle torque is introduced in the controller, which acts locally to push the robot away from the obstacles. The environment is initially assumed to be completely unknown, except the target location. Environment information is obtained from onboard robot sensors that have limited visibility range only. The neural network assumes a single layer structure, by taking advantage of the robot regressor dynamics that express the highly nonlinear robot dynamics in a linear form in terms of the known and unknown robot dynamic parameters. System stability and convergence are rigorously proved using a Lyapunov theory, subject to unmodeled disturbance and bounded unstructured dynamics. The real-time fine control of mobile robots is achieved through on-line learning of the neural network without any off-line learning procedures. A series of simulation results show that the proposed controller can be successfully applied to both static and dynamic environments, as well as a multi-robot system.