An Integrated LQR and Sliding Mode Controller for Balance and Motion Control of a Five-Link Two-Wheel Legged Robot

Two-wheel legged robots combine the mechanical structures of both wheeled and legged robots, achieving higher mobility speeds than legged-only robots while offering superior obstacle-crossing capabilities compared to wheeled-only robots. However, balance control remain challenges due to the inaccuracies of system model and external disturbances, particularly when operating under changes of robot height, mass and load. This paper proposes a novel fusion controller combining a single layer State-Fused Sliding Mode Controller (SF-SMC) and Linear Quadratic Regulator (LQR). The proposed approach retains the optimality of LQR while incorporating the robustness of SMC, thereby reduces the reliance on model accuracy and simplifies the complexity of SMC design. Meanwhile, the proposed fusion control algorithm only requires a simplified wheeled inverted pendulum model rather than a complex and accurate model, making it well-suited for low-cost robots with limited sensors and low computational capacity. The balance performance of the proposed method is extensively tested on a low-cost five link two-wheel legged robot. The results demonstrate that, compared to the standalone LQR, the proposed fusion control algorithm could reduce system overshoot, expand the stability region, and enhance robustness against external disturbances.