Evaluating Robotic Walker Performance: Stability, Responsiveness, and Accuracy in User Movement Detection

Abstract

This work presents the experimental evaluation of a robotic walker following the full implementation of its sensor and motorization system. The aging population and increasing mobility impairments drive the need for assistive robotic technologies that enhance safe and independent movement. The main objective was to validate the device's behavior in real-use scenarios by assessing its stability, responsiveness, and accuracy in detecting user movement. Tests were carried out in straight-line walking and on paths involving directional changes, both with and without motor assistance, using a cohort of five test users. Principal Component Analysis (PCA) and t-SNE dimensionality reduction techniques were applied to analyze the inertial (IMU) and proximity (TOF) sensor data, complemented by motor control monitoring through wheel Hall sensors, to explore gait patterns and system performance. Additionally, synchronized measurements between the user's and walker's inertial units and Time-of-Flight sensors allowed the evaluation of spatial alignment and motion correlation. The results provide a foundation for future system adjustment and optimization, ensuring the walker offers effective, safe, and adaptive assistance tailored to the user's needs. Findings reveal that the walker successfully distinguishes individual gait patterns and adapts its behavior accordingly, demonstrating its potential for personalized mobility support.