Control Design of GOOGOL’s Mobile Two-Wheeled Self-Balancing Robot Based on TP Model Transformation and Non-fixed-time Step Sampling Method

The use of tensor product model transformation based parallel distributed compensator control method increases the calculation burden exponentially with the increase of dimensionality, especially for high-dimensional linear variable parameter models, the limitations limits the potential usage of the method. To reduce the amount of calculation, two new control methods, Hammersley sampling method and uniform design method are proposed for self-balancing robot, and are furtherly optimized by the minimum volume simplex. The tensor product model transformation design method of minimum volume simplex combined with Hammersley sampling method and the tensor product model transformation design method of minimum volume simplex combined with uniform design method are proposed in this paper. The Googol’s mobile two-wheel self-balancing robot is used as the test object of the two controllers, and the design of fixed-point tracking is proposed. In order to achieve a more realistic control effect, external disturbances are considered to added to the left and right wheels, and different external disturbances are considered. Finally, three experiments are compared and verified. The controller proposed in this paper has the advantages of strong anti-disturbances ability and fast response speed.