Estimation-based disturbance adaptive model predictive control for wheeled biped robots

Enhancing the motion performance of wheeled biped robots amidst uncertain disturbances remains a challenge due to their under-actuated and inherently unstable nature. Aiming to address this issue, this paper proposes a disturbance adaptive control framework for such robots. The framework introduces a disturbance variable to describe the comprehensive effect of disturbances due to environmental interactions on the robotic system. A Kalman filter is also employed to estimate the robot’s center of mass (CoM) state and the uncertain disturbances by leveraging the dynamic coupling intrinsic to the robots. Estimated results are then integrated into a nominal model predictive control framework to generate an optimal CoM trajectory over a finite time horizon. This approach enables the robot to adapt to various types of external disturbances in the sagittal plane while maintaining accurate velocity tracking. The efficacy of the proposed approach is validated by conducting experimental evaluations on a hydraulically driven wheeled biped robot.