Inertial-based indoor backtracking and wayfinding aid for blind travelers

点击翻译按钮获取中文摘要

! Project Summary This application is concerned with technology that supports independent travel by people with visual impairment. Specifically, this project is primarily concerned with indoor navigation, a necessary activity for independent living. We will develop a system, in the form of an iPhone app, that will support two modes of operation: path backtracking, which enables one to re-trace a path previously traveled, but in the reverse direction, without the need for a map of the environment; and map-based wayfinding, which supports full-blown wayfinding functionalities, and is enabled when a map of the building is available in digital form. Our system will track the location of a walker using the inertial and magnetic sensors on the smartphone, which is kept in the walker’s pants pocket. Unlike most other proposed wayfinding systems, it will require no external infrastructure (such as radiofrequency or infrared beacons). The system will be specifically designed to take into account the peculiar characters of walking without sight. The algorithms developed in this project will be tested on WeAllWalk, a large data set of inertial data collected from multiple blind walkers as part of a prior project. Three user studies with 14 blind subjects will be conducted to assess the benefit provided by the proposed system. The Specific Aims of this project are briefly summarized below. Specific Aim 1: Robust path backtracking assistance in a corridor network. Work under this aim will implement a backtracking strategy for buildings characterized by a simple topology of interconnecting corridors, with special care given to challenging real-world situations typically encountered when walking without sight. These include: moving with irregular gait (e.g. when avoiding an obstacle); bumping into obstacles, walls, or other people; involuntary or intentional motion of the phone (e.g. to pick up a call). Specific Aim 2: Extend backtracking support to open space environments. Navigating through open spaces such as halls and courtyards is particularly challenging for blind travelers. We will implement and test with a system that facilitates backtracking in open spaces. Our approach is based on a simple dead-reckoning strategy, built on a robust step counting module coupled with heading direction measurements, with special provisions to ensure correct alignment at the beginning of the path. Specific Aim 3: Self-localization in a map and wayfinding. This project component will build a complete wayfinding system, relying on a floor plan that has been converted into an appropriate digital format (GeoJSON). Functionalities will include self-localization and generation of directions to a destination. Much of the work in this aim will be based on technology developed in the prior two aims. !