This interesting project out of MIT aims to use technology to help visually impaired people navigate through the use of a haptic feedback belt, chest-mounted sensors, and a braille display.
The belt consists of a vibration motors controlled by what appears to be a Raspberry Pi (for the prototype anyway) with a distance sensor and camera connected as well. The core algorithm is designed to take input from the camera and distance sensors to compute the distance to obstacles, and to buzz the right motor to alert the user — fairly expected stuff. However, the project has a higher goal: to assist in identifying and using chairs.
Aiming to detect the seat and arms, the algorithm looks for three horizontal surfaces near each other, taking extra care to ensure the chair isn’t occupied. The study found that, used in conjunction with a cane, the system noticeably helped users navigate through realistic environments, as measured by minor and major collisions. Users recorded dramatically fewer collisions as compared to using the system alone or the cane alone. The project also calls for a belt-mounted braille display to relay more complicated information to the user.
[via MIT News]