Balancing Wheeled Robot: Effective Modelling, Sensory Processing and Simplified Control

This paper deals with the modelling and control of balanced wheeled autonomous mobile robot. For the MBS dynamics modelling software tool Matlab-SimMechanics is used. The model derived automatically from geometric-topological description of MBS is used for the control purposes (local linearization for state space control, testing of nonlinear system controlled by LQR) and also as a reference during the analytical model formulation for global feedback linearization. The dual accelerometer is used as a tilt sensor and the proposed method for sensory processing is described in this paper. The approach is based on iterative solution of nonlinear equation. Control using the state space (LQR) and the feedback linearization is compared. Also, the influence of sensor noises and delays implemented into the model are discussed. Finaly, the solution is verified on real physical model controlled by means of hardware in the loop (HIL).