Image-Based Indoor Navigation for the Visually Impaired

Visual impairment and blindness preclude many essential activities of daily living. Among these is independent navigation in unfamiliar indoor spaces. Promising research to date has yielded several new assistive technologies for indoor navigation that require the use of tags and/or access points deployed in the environment, which can be impractical in many settings, thus rendering the indoor navigation system inaccessible. 
Incorporating users’ feedback and leveraging our experience with the design of indoor navigation systems for the blind, we are developing a new CPS technology for PERCEPT-V, our proposed organic vision-driven, smartphone-based indoor navigation system in which the user can navigate in open spaces without requiring retrofit of the environment. Using the PERCEPT-V application on an iPhone, the user simply takes a picture of the environment and is provided with navigation instructions to their selected destination. While the user is taking a picture of the environment, the PERCEPT-V application records observations from multiple onboard sensors (camera, magnetometers, antennas) in order to perform user localization. Once the location and orientation of the user are estimated, they are used to calculate the coordinates of the navigation landmarks surrounding the user. The application then provides accessible directions to the chosen destination, as well as an optional description of the landmarks around the user.  As the user moves throughout the space the onboard sensors are used to continuously track the user and provide seamless orientation to their selected destination. The PERCEPT-V system was deployed at the University of Massachusetts Amherst Campus Center, where we completed an initial phase of six usability trials with three BVI participants. Throughout the usability trials, the PERCEPT-V system was enhanced based on participants’ feedback.  The trials demonstrate that BVI participants were able to independently navigate the environment and next steps are to expand the test-site to introduce more complex navigation tasks and to test the system on a larger sample of BVI participants.