Haptic feedback to detect obstacles in multiple regions for visually impaired and blind people

Abstract

Focus on the development of assistive devices for visually impaired and blind people (VIBs) to provide assistance in their safety and mobility has increased, but making such devices portable is still a challenge. We propose a system for localized obstacle avoidance with a haptic-based interface for VIBs implemented using a robotic operating system (ROS) to improve the obstacle detection of existing assistive devices. With a depth camera sensor, an obstacle localization algorithm was developed utilizing the ROS framework to identify key regions to detect head-level, left/right torso-level, and left/right ground-level obstacles. The proposed wearable device provides a discernible array of haptic feedback to convey the perceived locations of obstacles. The system was tested by blindfolded volunteers to determine the accuracy in determining object locations in various environments. Experimental results showed the consistency of the system across different setups. The obstacle detection algorithm was optimized and evaluated to discriminate noises and concurrently detect smaller obstacles, thus making detection more robust. Subsequently, the Eulerian video magnification method was used to determine the level of vibration isolation for a prototype.