Joint Space Variable Impedance Control in Stair Descent of a Wheeled Humanoid Robot

This paper conducts a comprehensive investigation into joint space Variable Impedance Control (VIC) for a Wheeled Humanoid Robot (WHR) during stair descent. It starts by examining the structural characteristics, kinematics model, and equivalent dynamics model of the WHR, thereby establishing the theoretical foundation for designing the control algorithm. Subsequently, the principles of joint space VIC are elaborated, including impedance parameter selection, adjustment mechanisms, and how the control algorithm ensures joint compliance and stability. By continuously monitoring torso acceleration, roll angle, wheel-ground contact status, and dynamically modulating the joint impedance characteristics of the WHR accordingly, while simultaneously adjusting the target hip joint angles based on the torso's pitch angle, the proposed control strategy effectively manages joint torques. Simulation results demonstrate the superiority of the VIC approach compared to traditional fixed impedance control strategies in joint space. The findings clearly indicate that the VIC method enhances the stability and smoothness of the WHR's motion, enabling it to descend stairs with greater precision and confidence. This research significantly contributes to the field of WHRs by offering a more robust and adaptable control strategy for navigating complex terrains.