Research On Fuzzy-PD Control Method Of Two-wheeled Self-balancing Robot

The two-wheeled self-balancing robot system is considered as a typicalresearch target with the characteristics of multiple-variables, non-linear, highorder, coupled, and instability. Therefore, the robot has become an essential andclassical experiment facility to prove various kinds of control theory and controlmethod. So it has an important prospect for theoretical and practical significanceto carry out the study of two-wheeled self-balance robot. Because of its simplestructure, stable operation, high energy efficiency and environmental adaptability,it has a very broad application prospects both in the military and in civilian areas.This paper makes a study of the GBOT1001two-wheeled self-balancing robotproduced by Googol Technology (Shenzhen) Limited.Firstly, this paper gives an introduction of the hardware structure of theobject in detail, including DSP, inclinator, gyroscope and photoelectric encipher,and then analyzes the primary of self-balancing. Secondly, based on the Newtondynamics mechanics theory, a detailed mathematical model of the robot isprovided, and then, using the reasonable method, a linear state-space equation isbuilt up. We analyze the stabilization, controllability and observability of thesystem qualitatively. Thirdly, a LQR controller is designed to control the two-wheeled self-balancing robot. After a number of simulation experiments, we getthe best state-feedback matrix and Q, R matrix, and then build the Simulinkmodel. Fourthly, a fuzzy-PD controller is designed. The inputs of the fuzzycontroller are position and speed of the robot, while the angle and angle rate ofthe robot are controlled by PD controller. Giving a method to estimate I/Oproportional factors, according to the PID control parameters tuning principle,we determine the PD parameters. The LQR controller is compared with fuzzy-PDcontroller. At last, this paper designs a real-time control platform of the two-wheeled self-balancing robot, after a number of parameter debugging, the robot is controlled well.The simulation results prove that the fuzzy-PD controller has a betterdynamic performance than the LQR controller, with low overshoot and smalleroscillation, but longer settling time, while the LQR controller has a goodperformance of shorter settling time and faster reaction.