Adaptive Neural Fuzzy Controlapplied In Two-wheeled Self-balancing Robot

Two-wheeled self-balancing robot as a kind of the wheeled robot has abroad development prospects, and it will be applied in service, industry, business,for its low cost, small volume, stronger ability to adapt to the terrain andflexibility, so the two-wheeled self-balancing robot is full of balancing researchand commercial value.Two-wheeled self-balancing robot which is designed based on the invertedpendulum is a multi-variable, nonlinear, nonstable system. Therefore it is a idealplatform which can be used to test various laboratory control theory and controlmethod in the laboratory.The two-wheeled self-balancing robot in this paper was studied andanalyzed from several aspects below:1.The working principle, software and hardware component and systemoutside power suppliment of the two-wheeled self-balancing robot were brieflyintroduced,so do the performance and parameters of the main hardware.2.Mathematical model of two-wheeled self-balancing robot was establishedbased on Newton mechanics analysis and Lagrange equation. Linear systemmodel was made based on linear method around the system equilibrium point.The controllability, observability and stability of the system were all analyzed.3.Two controllers were designed. One is the system’s linear state feedbackcontroller based on pole-placement and another is adaptive neural fuzzycontroller. The comparative analysis of simulation was made.4.The adaptive neural fuzzy controller is applied to two-wheeled self-balancing robot which is a production of Googol Company. The real-time controlachieved a good control effect, from which we verified the feasibility of thecontroller.