Image processing based haptic guidance system for visually impaired runners

Abstract

People with disabilities face too many problems in their daily lives. The defeat of visual perception seriously con nes a person's daily life and activities. For example, visually impaired people have di culty to move around in unknown environment because they cannot determine exact object locations in their surroundings. It is even more challenging to do sports such as running on a track lanes. In this study, a wearable haptic guidance system has been developed which can help visually impaired individuals to navigate on a lane of athletic track. The system is based on image processing, and thus, includes a micro-controller, servo motors, a camera and a battery. The camera is used to feed lane detection and steering angle calculation pipeline. The skin stretch method is used to transfer calculated steering angles to the user. In order to evaluate the proposed algorithm and the feedback system, several tests have been conducted. According to the experiment results, the algorithm can detect lanes with an accuracy of 86.5 %, and the feedback system helps the user distinguish di erent feedback directions and magnitudes with an accuracy of 88.3 %. Considering these results, we can conclude that the designed haptic system is a promising tool for visually impaired runners navigating in an athletic track.