Research On Methods Of The Two-wheeled Self-balancing Vehicle's Wide-range Stability

This paper studies the two-wheeled self-balancing vehicle, which is a special kind of unstable, wheeled mobile vehicle. Through analysis we find that the system is characteristic with multi-variables, nonlinear, strong coupling, uncertain parameters, and is a complex experimental device. The research of two-wheeled self-balancing vehicle can effectively reflects many typical problems existing in control field, such as non-linearity, robustness, stabilization, follow-up and tracking. Because of the widespread application background and the research value, it has become an ideal platform to prove different kinds of control theory and control methods. This article mainly focus on two aspects as following: firstly, research how to enlarge the controllable angle range. Secondly, research how to simplify these control algorithm,and make them control the real vehicle stable eventually.Through analyzing and profiting the composition of the two-wheeled self-balancing vehicle from the domestic and foreign, the hardware structure, circuit and software of two-wheeled self-balanced vehicle are designed. With deep analysis in dynamics and kinematic, the non-linear model of the vehicle is obtained. In order to make the vehicle stable, firstly we linearize the nonlinear model through Taylor Series Expansion, and design the state feedback controller and linear quadratic regulator based on the linear model. On the basis of considering the limited range of leaning angle and the analysis of controllability, we draw the conclusion that these two linear controllers can only control the vehicle in a small angle range, and is unable to realize the wider range controllability in the practical work. Based on the consideration, we proposed two kinds of control method to seek for wider range of the two-wheeled self-balanced vehicle's stable angle. They are fuzzy linear controller and adaptive neuron-fuzzy inference controller based on the multipoint linearization. These two methods make the controller obtain a wider controllable angle range for the nonlinear model, moreover improved system's dynamic performance. Finally ,by reducing two state variables of vehicle's nonlinear model, we obtained a two-dimensional fuzzy control rules table about the angle,velocity and input signal u. Then we make out the fuzzy control algorithm with assembly language and C language,and use it onto the vehicle to make it stable eventually.