Tracking Control of Wheeled Mobile Robot Based on RBF Network Supervisory Control

Wheeled mobile robots have broad application prospects in the industry, agriculture, national defense, and other fields, but there are still many problems in their tracking control. At present, the dynamic modeling of wheeled mobile robots is mostly based on nonholonomic constraints, this means that the dynamics model of the robot cannot be accurately constructed; The work scenes of a robot are complex and diverse, and the changes in the external environment place high demands on the robustness of the robot control system. The parameter tuning by the traditional PID control method often lacks stability margin when faced with complex environments. Also, the control method based on model learning cannot guarantee system stability when accurate modeling is not possible. Aiming at solving these two problems, this paper proposes a two-wheel differential mobile robot control system based on RBF (Radial basis function) network supervisory control. By combining the control characteristics of an RBF neural network and traditional PD controller, the instability problem in model learning is overcome. Improve the ability to adapt to the dynamic environment and enhance the robustness of the control system. The proposed control system is verified by the simulation experiment, and the results show that the proposed system can completely eliminate steady-state errors when tracking a dynamic target that moves regularly. The proposed system has strong robustness against various environmental changes and global progressive stability.