A Mode-Switching Strategy for Robotic Gait Assistance Based on a Wearable Wireless Electromagnetic Handle

Existing gait-assistance robots are constrained by the limitations of line-of-sight perception in complex environments, static decision architectures, and the absence of disturbance-compensation mechanisms, resulting in poor environmental adaptability, limited rehabilitation modes, and low control accuracy. To address these issues, this paper proposes a mode-switching strategy that combines a wearable wireless electromagnetic handle, a Hierarchical Finite State Machine (HFSM), and a Radial Basis Function–Sliding Mode Controller (RBF–SMC). First, we design and introduce a wearable wireless electromagnetic handle as the core human–robot state-sensing module. By leveraging low-frequency alternating magnetic fields, it provides real-time 3D localization, velocity, and pose estimation of the target user; its strong penetration through the human body and common occlusions supplies key state variables, reliable mode switching. Second, we construct an HFSM: at the high-level decision layer, the system dynamically switches among forward assistance, lateral cane-style support, and rear monitoring modes to accommodate users with walking abilities at levels II–IV; at the low-level decision layer, the assistive following distance is adaptively adjusted based on the user’s walking speed and trunk inclination. Finally, we propose an adaptive sliding-mode controller based on an RBF neural network that compensates unmodeled dynamics online to effectively suppress chattering and improve pose-control accuracy. It operates in the same control cycle as a behavioral-dynamics obstacle-avoidance module, enhancing the robot’s naturalness and safety. Experimental results show that, compared with conventional PID, SMC, and LQR methods, the proposed approach improves control accuracy by at least 32.1%, and increases assistance compliance and coordination by at least 60.3% and 78.1%, respectively. By monitoring users’ gait states in real time, the HFSM autonomously switches assistance modes, improving comfort and safety. This study provides a highly adaptable and safe control framework for rehabilitation gait-assistance robots with clinical application.