Research On Balance And Path Tracking Control Of Two-wheeled Mobile Robot

Based on posture balance of two-wheeled mobile robot, the article focuses on theory research and experimental analysis of path tracking control. Meanwhile, RBF neural network algorithm is introduced to optimize the gain of path tracking controller which improves the performance of path tracking of the robot. Based on a large number of relevant reference materials, the article mainly makes some analysis and research from the following parts.(1) The construction and analysis of the mathematical model, this paper establishes the kinematics model and dynamic models of two-wheeled mobile robot respectively. It establishes kinematic model by analyzing the kinematic constraints and the dynamic model by using Lagrange analyzing method. Besides, it makes some analysis for the control performance of the system.(2) The design of hierarchical sliding mode control algorithm and chattering suppression. Although the robustness of hierarchical sliding mode controller is well, the controller’s chattering will reduce the performance of the system. In order to suppress chattering of hierarchical sliding mode controller, a non-linear interference disturbance observer is introduced to observe the interference of the system, and the results is returned to the hierarchical sliding mode controller to make some compensation for the control system. So the magnitude of the gain of switching term is determined by the prediction value of disturbance observer and error value of the actual value, which suppress the chattering of the controller effectively. Therefore, the control system has good robustness and dynamic characteristics.(3) The design and optimization of path tracking controller based on dynamics model. In order to solve the problem of path tracking of two-wheeled mobile robot based on dynamics model, the control system is divided into two parts which contains kinematics and dynamics control firstly. Then the path tracking controller is designed based on dynamics model. Finally, RBF neural network is used to correct the gain of the controller in order to improve the performance of path tracking of the system under the condition of presence of interference.