Low cost sensor network for obstacle avoidance in share-controlled smart wheelchairs under daily scenarios.

Intelligent wheelchairs working in various environments need to sense the neighborhood around itself and prevent dangerous situation such as collision and falling down from stairs. In this paper, we develop a low cost and real time intelligent wheelchair shared control system for people with disabilities at the rehabilitation center or at home. Our system is composed of an electric wheelchair, an RGB camera, an infrared camera, 4 ultrasonic sensors, a laser LiDAR (light detection and ranging equipment), and a personal computer. Based on the intelligent sensor network and sensor priority control algorithms, it can detect different obstacles with various dangerous levels, send out voice and graphical alarm, and suggest a safe path to avoid them. The combination usage of the RGB camera and the IR camera enables our system to work under normal lighting conditions and under low lighting conditions. The reliability of individual type of sensors and the whole system was examined by simulated scenarios and user questionnaire. Experiment results reveal that the system is reliable and effective. The key components in our system are low cost and easy to be installed on various electric wheelchairs and common PCs. It will be a useful tool for further developing smart wheelchairs with shared control frameworks.