Abstract
The assistive robotics field works with technologies aimed at assisting people with physical impairments. Simple tasks like passing through doorways can be a significant challenge for motored wheelchairs users. Besides, depending on the disability degree, the task may be impossible to complete without help. Therefore, mobile robotics technologies have been implemented in wheelchairs to give them intelligence, and a supervisory interface inserts the user in the control loop to complete the objectives. Thus, this research proposes an innovative method to assist the user to cross narrow passages automatically. The methodology works in known and unknown environments. Besides, it does not require mapping and global localization algorithms. In this technique, a depth camera recognizes the edges between a narrow path. A Kalman filter method is used to improve the precision of these points even when the passage leaves the camera field of view. Finally, a nonlinear controller performs the narrow passage crossing. This methodology was analyzed using a MATLAB model, Gazebo simulation platform, and a real mobile robot Pioneer-P3DX. The results demonstrated promising applicability.
