Adaptive Navigation Assistance for Visually-Impaired Wheelchair Users

It is estimated that approximately 10% of people who are legally blind require wheelchairs [1]. Wheelchair users with visual impairments face difficulties in avoiding obstacles as well as identifying visual cues in the environment, thus making independent navigation challenging, and in some cases, impossible. The authors in [1] suggest that intelligent wheelchairs capable of collision avoidance and path planning would greatly benefit wheelchair users with visual impairment. Although several intelligent wheelchairs have been developed recently [2-4], these wheelchairs navigate autonomously, thus taking control away from the user. On the other hand, wheelchairs that only provide collision avoidance support [5] are not appropriate for drivers who are unable to determine their location and want to navigate to a specific location. We thus present a novel, real-time, vision-based intelligent wheelchair system that avoids collisions and provides adaptive audio prompts to help blindfolded users navigate to specified destinations. Existing intelligent wheelchairs have used various active sensors (acoustic, sonar, infrared, laser, etc.) [6]. We rely solely on a stereovision camera due to its low power consumption, ability to perform in natural environments, and relatively low cost. Most outdoor wayfinding systems rely on GPS, which is unreliable in indoor settings, while indoor wayfinding systems typically use beacon and RFID technology, which require modifications to the environment. By using vision-based techniques we can achieve accurate localization, while reducing/eliminating the need for environment modifications. In addition, cameras capture and provide a richer dataset than can be used for high-level scene understanding to build maps and determine what type of room the wheelchair is in.