Nonlinear Dynamic Modeling and LQR Optimal Control of Wheel-Bipedal Robots

This is the authors' abstract. We don't add key points for this paper.

The wheel-bipedal robot integrates the advantages of both wheeled and legged systems, enabling efficient locomotion. However, many control approaches simplify this system to a wheeled linear inverted pendulum (WLIP), treating the legs and body as a unified entity and overlooking the distinct dynamics and effects of the legs. This paper derives a simplified model of the leg mechanism that captures the interactions between the robot’s leg and body dynamics. A linear quadratic regulator (LQR) controller is further designed to maintain overall balance, while a dedicated leg controller governs leg-specific movements, establishing a dynamic composite control framework. Additionally, this paper details the system architecture and development of controllers for various scenarios. Experimental results from prototype testing in real-world environments demonstrate the effectiveness and robustness of the proposed control strategies.