Model predictive control for a two wheeled self balancing robot

In this paper, a model predictive control is presented and applied to a two wheeled self balancing robot. This robot is known as a three-degree-of-freedom mobile robot with multivariable underactuated dynamics. The dynamic equations of the system are first introduced and the inputs and outputs are assigned. Then, by using a compensator block in sequential closing loop method, the original system is transformed into two subsystems. The model predictive controller is designed for maintaining robot in balance despite the external disturbances. The robustness of the method is evaluated in the presence of sinusoidal and step-like disturbances. The simulation results are also presented and discussed using Matlab software.