Study Of Trajectory Tracking Control Methods On Nonholonomic Mobile Robot

Nonholonomic mobile robots are considered as the control object and control problems are studied focusing on trajectory tracking, which is an important issue on nonholonomic mobile robot motion control. Wheeled mobile robots come in a number of different mechanical structures. Typical robots have two co-axial wheels that independently or synchronously actuated to achieve forward/backward and rotational motion. The synchronous-drive type of mobile robots has wheels that can steer and rotate together so that the wheels can parallel all the time while the differential-driven robots is simplicity of its mechanical structure.For synchronous-drive robots, firstly the finite time control method is introduced and an algorithm by combining the neural dynamics model with finite time control for the trajectory tracking problem of a nonholonomic mobile robot. The finite time control method is easily designed and the resulting control law is continuous. It is improved by neural dynamics model which has stability, bounded and smoothly response character, can deal with the speed sharp jump on initial time. Then the integration of a kinemetic controller and a torque controller for the dynamic model of a nonholonomic mobile robot has been presented. Adaptive controls are derived for dynamics system of nonholonomic uncertain mobile robots tracking a reference trajectory using backstepping technique. The designed controller ensures the asymptotic zeroing of the stabilization error.For differential-driven robots, a speed controller is proposed. In addition to the independent control loop for each wheel, an angular speed compensator is included to enhance synchronization of the motions of the wheels. The controller is still designed around each wheel to keep the simple structure of independent control. Then a model-based adaptive control is presented for differential-driven robots. The controller takes into account of the coupling between the motions of the two wheels as well as the uncertainties of the system. It modifies the traditional adaptive control inputs, less computationally and keeps the motor moment outputs in a sound bound smoothly.