Personalized Velocity Control System of Autonomous Wheelchairs Considering Human–robot Interaction–based Human Factor

This study proposes a personalized velocity control system for an autonomous wheelchair by deriving human factors from the subjective opinions and requirements of users obtained through Human–Robot Interaction (HRI)–based usability evaluations. Using Bézier curve-based velocity profiles with different average deceleration levels, five deceleration modes are designed and tested with 38 participants in an indoor environment. The usability evaluation employs a seven-point Likert scale, and the collected responses are probabilistically analyzed to compute performance scores for each deceleration mode, leading to final performance indices defined as human factors. An optimal velocity profile that satisfies user requirements is generated based on these human factors, and a personalized control system is implemented by integrating a velocity controller with tracking error-based and feedforward compensation. In experiments with the same participants from the usability evaluation, the integrated controller closely tracks the optimal velocity profile, and the tracking error distribution is concentrated near zero, remaining within ±0.2 km/h. These results demonstrate that the proposed personalized control system improves driving safety and ride comfort while maintaining a high tracking performance by embedding quantified human factors into velocity planning and the integrated control framework.