Helping Blind People Navigate | The Economist
FOR centuries, canes have served blind and partially sighted people well by giving them a way to negotiate the world around them. The only serious improvement they experienced was in 1921, when a Briton named James Biggs, who had recently lost his sight, painted his own cane white to make it easily visible and alert others to the presence of someone unable to see. obstacles nearby. In the opinion of Daniela Rus of the Massachusetts Institute of Technology (MIT), however, the white cane has had its day. Dr Rus would like to replace it with a system that scans the user’s environment and communicates what he sees.
Dr Rus’ device, a prototype of which she presented on June 1 at the International Robotics and Automation Conference in Singapore, consists of a camera worn on a lanyard around the neck and a belt. A computer inside the camera creates a three-dimensional image of the area in front of the wearer, processes it to extract relevant information, and uses the results to transmit the appropriate signals through the belt.
Dr Rus knew from previous attempts to construct devices of this type that what might seem like the obvious way to manifest these signals, namely as sounds with specific meanings, was in fact not a good approach. Blind people depend a lot on their hearing and don’t like latest generation devices to interfere with this sense with beeps and clicks. Hence the belt, which has five vibrating motors installed in it. One sits in the center of the wearer’s abdomen. The others flank this central engine, with two spaced on either side of it.
This configuration allows the computer to warn a carrier when he is on a collision course with an obstacle. It does this by telling the motor pointing closest in the direction of the obstacle to vibrate. If the wearer walks towards a wall, for example, the central motor vibrates gently as it approaches it within a few meters. If he ignores it, maybe because he really wants to reach the wall, the computer increases the amplitude the closer he gets, giving him a good idea of the exact distance he’s at. Likewise, if he risks banging his right shoulder against a door frame while walking from room to room, the belt’s right-most motor warns him of the impending collision. And it works. Compared to white cane navigation in one of MIT’s crowded hallways, this reduced collisions between blind students and others by 86%.
The new system can do more than just help someone get around without a collision, however, as the belt incorporates a touchpad with braille instructions written on it. This allows the user to program it to perform specific tasks.
For example, Dr Rus knew that blind students often have difficulty finding a free seat in a crowded lecture hall. Adding an appropriate algorithm to the computer software helps to get around this problem by allowing it to recognize chairs, and also whether a chair is occupied or not. In this case, the motors are used to indicate a direction to go rather than a direction to avoid. Activating the algorithm using the touchpad causes the motor pointing closest to an empty chair to vibrate when the system detects one.
In trials involving a room containing an empty chair, an occupied chair, and also a recycling bin, the algorithm directed the belt wearer directly to the empty chair 80% of the time. Cane users shown with the same arrangement always ended up finding the chair empty, but in doing so, they came into contact with objects other than their target more than five times more often than those using the camera and belt.
It remains to be seen whether a camera (ideally smaller than that of the prototype) and a belt could completely replace a cane. In particular, Dr. Rus ‘system lacks an important feature of Biggs’ innovation. A white cane not only helps a blind person navigate, it also signals their condition to the rest of the world, allowing others to adjust their behavior accordingly. As an additional aid, however, his approach looks the most promising.
This article appeared in the Science & Technology section of the print edition under the title “Canne blanche 2.0”