User-Centered Obstacle-Avoidance-based Shared Control Algorithm for an Active Assistive Walker

Smart walkers (SWs) have emerged as vital assistive technologies for enhancing mobility and independence among the elderly and individuals with gait impairments. However, effectively interpreting user intent without relying on expensive, fragile force sensors remains a significant challenge. This paper proposes a novel control framework for a SW that integrates a Sensorless Force Estimation (SFE) observer with a sector-based shared control strategy. The SFE algorithm utilizes intrinsic motor currents and kinematic data to estimate user interaction forces. It compensates for system dynamics and ground inclination to provide a natural admittance interface. To ensure safety without compromising user agency, a discrete finite state machine controls obstacle negotiation is performed. When an obstacle is detected within a safety threshold, the system halts motion and enters a safety interlock state, requiring a deliberate backward unlock gesture from the user to resume movement. Once unlocked, the shared controller evaluates environmental clearance using LiDAR sectors and modulates the turning gain, amplifying user intent toward clear paths while providing haptic resistance against obstacles. Experimental validation demonstrates that the proposed system is able to prevent collisions, interpret user intent, and provide intuitive guidance, offering a cost-effective and robust solution for assistive mobility.