A Precision Navigation System for People with a Visual Impairment

Atlanta Research & Education Fdn, Inc., Decatur GA

Investigators

Abstract

? DESCRIPTION (provided by applicant): The purpose of this proposal is to develop a precision navigation system for people with visual impairment. There is a technological gap between what can now be provided and what is needed. What is needed is cane-touch precision. The person needs to be guided close enough to specific locations to find it with a cane. Existing navigation systems employing the Global Positioning System (GPS) can only provide the name of the intersection being approached. The proposed research design is constructed around the hypothesis that the goal of precision guidance can be achieved by merging two types of navigation systems: (1) GPS and (2) Inertial Navigation. Inertial navigation uses sensors now contained in smart phones to track a person's location as s/he is walking toward a destination. Given the location of a destination, it can precisely guide the user there. But it is a short-range solution that loses precision past distances of 50 to 100 meters. GPS often has an offset bias error. In a particular locale it may be off by 10 meters to the north of the actual location. This offset bias may be stable for several blocks, and then shift over a fairly short walking distance. The working hypothesis is that cane-touch precision can be achieved by fusing inertial navigation data with GPS data using a data fusion algorithm. Such an algorithm uses statistical methods to merge from two sources to provide more accurate and reliable information than can be obtained from data source separately. There is also an information gap. Research has shown that current inertial navigation systems do not work well for persons with a visual impairment. The reason is that people walking with a cane have an atypical walking style that makes it difficult to track their footsteps. People with a visual impairment tend to walk in a tentative fashion, making sure there is solid ground beneath their foot before placing it down. Various types of body motion characteristic of this population have been shown to be indicative of a footstep. To this end, a motion analysis study for persons with visual impairment is proposed. The hypothesis is that data from this study will provide what is needed to track the movement of their feet, and thus the person's movement. Also needed is map information providing the exact location of accessible walking paths and crosswalks, etc. We will develop a crowd-source paradigm to populate a public database with such map data. The resultant navigation app will be implemented as an application (app) for a smart phone and evaluated by subjects who are visually impaired. The expected outcome is an app for smart phones that provides precision (cane-touch) navigational guidance to persons with visual impairment. This will be invaluable, providing them the ability to independently and safely navigate outdoor environments and find important environmental elements such as pedestrian push buttons, which are never in a standard location. It will also enhance the gait literature for this population.