Study On Tracking Control Of Wheeled Mobile Robot

The mobile robot, as a typical nonholonomic system, cannot be stabilized using smooth linear feedback, control of such system has been a hot topic in the research of robotic community. Many elegant control strategies have been proposed for tracking control of mobile robot. But most of the existing literature are based on the kinematic model of control or simplified dynamics model; control law, based on considering the actual parameters of the full dynamical model control, needs to be further developed; meanwhile, the mobile robot as a complex nonlinear coupling, multi - variable dynamic system, parametric unmodeling disturbance and outside disturbance, as well as detection signal of uncertainty, makes it difficult for researchers to establish a precise mathematical model, in the design process of tracking controller, consider these factors of uncertainty; furthermore, when most of the existing literature extend the kinematics model to the dynamics level, there are some disadvantage: during tracking controller design process, it is under the assumption of robot kinematics model of distance d which is equal zero; but in dynamics model of mobile robot, it assumes that the distance d is not equal zero. So, how to deal with such problems is a research topic in this thesis.Consider the above shortcoming, this thesis applies cascaded control method, neural dynamics model, feedback linearization control method, and adaptive approach, to deal with the velocity jump problem at the initial time, the reunification of handling the distance d both in kinematic model and dynamics model, and the issue of un-accurate pose information because of robot sensors un-accurate measurement with the impact of external disturbance.First, a stable structure of the tracking control algorithm has been proposed, to solve the mobile robot dynamics level tracking control; under consideration of the mobile robot nonholonomic constraints and parameters of the actual car, apply neural dynamics-improved cascaded controller combing of torque calculation method, to achieve the dynamics level expansion; Lyapunov theory ensure the overall system asymptotic stability. Then, in order to unifying the dealing with distance d both in kinematic model and dynamics model of tracking control, in the case of distance d not equal zero, design kinematics controller, further, apply the adaptive method, improve the PI filter,get the improved controller with better response performance. Finally, under the situation of un-accurate measurement of pose information, design the kinematic observer, to effectively overcome the defect of sensor measurement errors and external disturbance .In Matlab/Simulation environment, simulation results show: based on the comprehensive tracking controller for mobile robot system in the face of external disturbance circumstances, still possesses an excellent quality control performance.