Using the Numerical Optimization Method for Tuning the Regulator Coefficients of the Two-Wheeled Balancing Robot

Abstract

This article is devoted to tuning the regulator coefficients of the automatic stabilization system of a two-wheeled balancing robot. When calculating the regulator coefficients analytically, it is necessary to set the desired distribution of the system roots. This distribution can be defined based on general requirements for the quality of the system transient processes such as the maximum level of overshoot and the maximum duration of the transient process that define the sector in the complex plane within which the roots should be placed. But the issue of choosing the distribution within this sector remains open. This task becomes even more complicated if there are nonlinearities in real devices. It is shown that the root distribution in the form of Newton binomial gives not a good enough result because it is not possible to correct the transient process towards more speed or less oscillation on this case. The technique that allows tuning the coefficients of the regulator of the stabilization system by performing the numerical optimization procedure is proposed in the paper. In this case, it is possible to change the form of the transient process depending on the specific requirements for the system by changing the form of the cost function. The results of the work are confirmed by numerical simulation and experimentally.