Research On Two-wheeled Self-balancing Robot Based On Adaptive Fuzzy Neural Network Control

Two-wheeled self-balancing robot is a kind of nonholonomic andunderactuated non-stationary system which is a complex nonlinear system. Underthe meaning of experimental robotics, two-wheeled robot is an importantscientific experimental apparatus, and valuable for scientific research andexperiment in robotics and related subjects. At the same time, the two-wheeledrobot has the characteristics of covering small occupation area and flexible.It canbe used in many special environment (such as searching and rescuing in thenarrow space or crowded office environment).It can also be used as a specialvehicles, special vehicles, special sports equipment, security system for thedisabled and many other fields of modern society.The history of two-wheeled self-balancing robot and the prospect of two-wheeled robot technology application are introduced. The basic structure of two-wheeled self-balancing robot is also introduced, then the movement principle ofthe two-wheeled robot is analyzed.The kinematics model of two-wheeled robot system is established, and alinearization method near the balance point is used and obtain a linear model.The controllability, observability and stability of the linear system are analyzed.A linear state feedback controller based on pole assignment is designed, and dothe simulation experiment. An adaptive neural fuzzy controller for robot systemis designed, do the simulation experiment, and apply it to the two-wheeled robotexperiment platform produced by Gugao company.