A Command-Based Control Framework for Smart Walkers with Real-Time Position Estimation

This paper presents a control strategy for a smart walker designed to assist users with mobility and visual impairments. The proposed system integrates an Extended Kalman Filter-based localization algorithm and a discrete command generation strategy using auditory cues, enabling intuitive interaction and safe navigation in structured environments. The walker's behavior is modeled as a differential-drive system, with position estimation based on sensor fusion from magnetic encoders and an inertial measurement unit. A finite state machine governs the command logic based on tracking error and alignment thresholds. Experimental validation was conducted with multiple users performing T-type and step-type trajectories in a laboratory setting. Results demonstrated accurate trajectory tracking, effective command interpretation by users, and robustness of the control strategy. These findings support the applicability of the approach in real-world assistive scenarios and open paths for future enhancements through adaptive interfaces and dynamic control mechanisms.