A Blind Driving System Leveraging Multimodal Feedback
Ruizhe Zhang, Sha Zhao, Wei Yang, Shijian Li, Gang Pan · 2024
Driving poses a significant challenge for blind and visually impaired (BVI) individuals, due to the loss of vision. Fortunately, the metaverse presents a promising platform where BVI individuals can experience independent driving just like sighted individuals. In order to assist such individuals in driving in the metaverse, we have devised a blind driving system that leverages multimodal feedback. We utilized stereo headphones and vibration motors to convey necessary information, including current and predicted vehicle states (i.e., speed, position, and direction) during driving, through sound, voice, and vibration feedback. Meanwhile, in order to improve the user-friendliness, we reduced the user’s information burden by data combination and compression. With the help of the multimodal feedback, BVI drivers could intuitively perceive the position and direction of the vehicle, thereby improving their driving experience. To evaluate the effectiveness of the system, we designed virtual scenarios to simulate a real driving environment, and we further devised two tasks: lane-keeping and lane-changing. We conducted user studies where sighted individuals wearing blindfolds were involved in the two tasks. The results suggest that our system demonstrates a beneficial effect in assisting BVI individuals in driving. Users could keep the vehicle in the center area of the lane for over ninety percent of driving time, and completed a lane-changing task with a high success rate. This study provides a potential solution for blind driving for BVI people by employing multimodal feedback.