Advances in artificial intelligence (AI) can be intimidating, but this development gives hope for greater independence for the blind or visually impaired (BVI). Engineers at the University of Colorado Boulder are using advances in AI to develop a new kind of walking stick. In a study published in October 2022, Shivendra Agrawal, a doctoral student in the Department of Computer Science and his colleagues in the Collaborative Artificial Intelligence and Robotics Lab think that tasks like grocery shopping or finding a seat in a cafe are solvable problems for the BVI. (University of Colorado at Boulder. “Smart walking stick could help visually impaired with groceries, finding a seat,” ScienceDaily, 19 January 2023, www.sciencedaily.com/releases/2023/01/230119185812.htm)

The team’s prototype walking stick resembles the white-and-red canes we know but includes other features. Using a camera near the handle and computer vision technology similar to that in self-driving cars, the walking stick maps and catalogs its surroundings. It then guides users by using vibrations in the handle and spoken directions.

To find a seat in a crowded cafe, sighted persons will enter, scan the area and choose a seat. The BVI prefer seats close to walls and not face to face with a stranger. To test the walking stick, researchers set up a model cafe in their lab with several chairs, patrons and obstacles. Study subjects wore blindfolds and a backpack with a laptop. Using the smart walking stick, they turned to survey the room with the camera. Algorithms running inside the laptop identified the various features in the room, then calculated the route to a preferred seat. Subjects were able to find the right chair in 10 out of 12 trials with varying levels of difficulty.

Next came grocery shopping. While the variety of products with similar packaging can be confusing for the sighted, it’s an enormous challenge for the BVI. The team designed a grocery shelf stocked with several different kinds of cereal. The researchers then created a database of photos of cereal boxes into their software. Study subjects used the walking stick to scan the shelf, searching for the product they wanted. The system selected the most likely product, then issued commands such as “move a little bit to your left.” The group now wants to make the system more compact to run off a standard smartphone attached to a cane. Researchers hope their preliminary results will inspire other engineers to rethink what robotics and AI can do. As Agrawal said, “We think assistive robotics has the potential to change the world.”

Linda Conlin
Pro to Pro Managing Editor
[email protected]