A Unified Control Framework for Self-Balancing Robots: Addressing Model Variations in Wheel-Legged Platforms and Human-Carrying Wheelchairs

Abstract

Self-balancing robots, with their compact size, are capable of achieving high agility. Small wheel-legged self-balancing robots have demonstrated significant potential across various applications. However, expanding small self-balancing robots to larger sizes to serve as personal transport tools is a more attractive and impactful direction than further miniaturization or confinement to niche laboratory demonstrations. This paper presents the development of a small self-balancing robot, which is then scaled up to a larger version designed to carry human passengers as a self-balancing wheelchair. A unified control framework, built around a shared core of online model-updating LQR for balance and PD for steering, is applied to both robots. This core is supplemented with platform-specific modules, such as a dedicated leg controller for the wheel-legged robot, to handle distinct dynamic maneuvers. The LQR controller is implemented for balance control in both robots. Additionally, a dedicated leg controller is applied exclusively to the small wheel-legged robot to enable dynamic maneuvers, such as jumping. A series of experiments conducted with the final prototypes validate the effectiveness of the control systems and highlight the robots' application potential.