Research on the Control System of Two-wheel Balance Vehicles on the Basis of Fuzzy Neural Network

Abstract

A control method on the basis of fuzzy RBF neural network PID is put forward to solve the problem that balance vehicles can not better adapt to users of different body types and road surfaces of different complexity because of their fixed control parameters. Firstly, the mathematical model of balance vehicles is established with the Newtonian mechanics so as to obtain a dynamic equation of this system. Secondly, the fuzzy control and RBF neural network are combined to obtain a fuzzy neural network control system that can dynamically adjust the parameters. The control system of the balance vehicles is built on the simulink platform of MATLAB and simulated with the following methods: the fuzzy RBF neural network PID control, fuzzy PID control and traditional PID control methods. Last, the self-balancing and antijamming experiments are carried out on the balance vehicle experimental platform. After comparing the experimental results, it can be found that the first method represents better control performance than the later two ones, because it enjoys smaller overshoot, better adaptability and anti-interference, and can be applied to different users and in varied scenes.