Nonlinear Model Predictive Control for Ballbot systems: A benchmark with Hierarchical Sliding Mode and Linear Quadratic controls

This study presents the Nonlinear Model Predictive Control (NMPC) method for ballbot systems. Due to the complex nature of the underactuated system, ballbot requires a controller capable of ensuring both goals: moving to the desired position and maintaining the balance of the body throughout the moving process. NMPC is an optimal control method based on a predictive model to calculate the states of the ballbot after a certain period of time in the future. Moreover, NMPC can optimize the position, tilt angle and control signal through the cost function and handle special constraints of the system. Simulation results show that NMPC is more effective than some previous techniques such as LQR and HSMC in terms of safety constraints of tilt angle and input-output signals.