fuzzy model reference learning control for underactuated segway robot

The problem of the Segway robot and wheeled inverse pendulum is one of the most important classic issue and well-known example of nonlinear control problem in engineering control. The main purpose of the control in this paper is achieving the cart body to the desired angle and the speed of the body as quickly as demanded velocity by controlling of the applied torque. This paper presents an intelligent control method for Segway robot, in which the downstream fuzzy controller using the calculated error from the plant output, provides the proper input to the dynamic system. But, there is a learning mechanism in the upstream controller that compares the output of the controller VS the reference model and then improves it using fuzzy inverse controller. it makes method more robust and efficient. Dynamic equations of the system are extracted by the Kane method. The design of the fuzzy controller, based on the experience of the expert human for such a system, can perfectly overcome the instability and nonlinear nature of the system. Therefore, the use of the superviser controller, due to automatic and intelligent adjustment, can help simpler design of this controller. The performance of the proposed control algorithm on Segway robot case study is evaluated and demonstrated by simulation in MATLAB. Also, considering the system's uncertainty in design and disturbances, the controllers are robust against the uncertainties and disturbances in the problem. Figures and results are evidence of this issue.