A multi-stage shared control method for an intelligent mobility assistant

This paper presents a multi-stage shared control method (MSSC) which can be used to control the movement of a robotic mobility assistant designed to facilitate safe mobilization for people with unstable gait. The multi-stage control module consists of user intent, obstacle avoidance and fuzzy logic components. The user intent represents the person's commands. The obstacle avoidance component reads data representing any obstacles in the vicinity of the assistant and uses the vector field histogram (VFH) algorithm to select a suitable path to avoid any obstacles in the path of travel. The fuzzy logic component is responsible for merging the user intent and obstacle avoidance information such that the user's request is satisfied to the highest extent possible. When an unsafe situation presents itself the user's request(s) would be partially or wholly overridden so the assistant can return to a safe state. The system has been designed to be dynamically configurable so as to suit different users in terms of gait stability and strength, preferred speed of travel and level of control over the system. It has been tested both in a simulated environment and real-world operating conditions and has been shown to effectively avoid obstacles with minimal disruption to the user and their intent.