Title: CDI-Type II: Cyber Enhancement of Spatial Cognition for the Visually Impaired

NSF grant CDI-0835689

N.A. Giudice, (UMaine-PI), K. Daniilidis, UPenn (PI), R. Manduchi, UCSC, and S. Roumeliotis, UMN.

Brief Abstract.

Indoor navigation poses significant challenges for blind and visually-impaired persons as without vision, there is often no mechanism for accessing room numbers, building maps, and other navigation-critical environmental cues. Effective wayfinding requires successful execution of several related behaviors including orientation, spatial updating, and object and place recognition, all of which necessitate accurate assessment of the surrounding environment. Most research on wayfinding aids has focused on outdoor environments, as spatial behavior is readily supported by speech enabled GPS-based navigation systems which provide access to information describing streets, addresses, and points of interest. By contrast, there is a dearth of technology for conveying such information for supporting environmental access and wayfinding behavior for indoor usage. the limited technology that is available requires significant modifications to the building infrastructure and has limited functionality, barriers which have discouraged adoption.

The lack of compensatory indoor navigation technology has led to dramatic problems in  the independence, quality of life and safety risks for blind individuals, one of the fastest growing demographics of our aging population. Guide dogs and white canes are widely used for the purpose of mobility and environmental sensing. However, where these tools are extremely effective for identification of obstacles in the path of travel, they do not provide information useful for staying oriented in the environment or building up a robust mental representation of the space (cognitive map). What is needed to solve the indoor navigation challenge is a device that conveys realtime information about the environment and which supports wayfinding behavior and cognitive map development.

This proposal adopts a multi-faceted approach for solving the indoor navigation problem for people with limited vision. It leverages expertise from machine vision, robotics and blind spatial cognition, with behavioral studies on interface design to guide the discovery of information requirements and optimal delivery methods for an indoor navigation system. The proposed cyber assistant provides position, orientation, local geometry, and object identification via the use of appropriate sensors and sensor fusion algorithms. The combination of designing perceptually guided  navigation algorithms, implemented on miniature-size commercially-available hardware, in a device which is motivated by theories and end-user needs of the target population, will lead to the development of a navigation assistant with the broadest impact to the widest range of potential users.