Tackling the Duality of Obstacles and Targets in Shared Control Systems: A Smart Wheelchair Table-Docking Example

Many studies have shown that a smart wheelchair could improve the quality of life of people with restricted mobility by providing them with more freedom in the daily activities they can undertake independently. In addition to enhancing independent mobility, it is important to ensure safety for wheelchair users and those around them. To date, previous studies have mostly focused on (semi-)autonomous navigation or obstacle avoidance. By contrast, in this study, we tackle the challenging, but important problem of safely docking to tables. We propose a robotic navigation assistance, applied to electric powered wheelchairs using Time-of-Flight (ToF) sensors to facilitate table-docking for users. To meet this objective, we designed a low-cost sensor system that was integrated into our smart wheelchair prototype, which can detect a table and accurately estimate its height. We then developed a robust algorithm to deliver the manoeuvring assistance. First, we simulated the smart wheelchair system within Unity3D to find the best positions for the ToF sensors and evaluate the accuracy of the docking system, employing different table styles. Then, we experimentally validate the system on our physical wheelchair, using varying angles of approach, which demonstrate its feasibility.