Artificial Memory Helps Robots Locate Lost Items

A new robotic program has been developed to help locate missing items, with particular applications for people with dementia or chronic memory loss.

The program, designed by a team of engineers from the University of Waterloo, uses an object-detection algorithm to identify items and store a memory of what time and where the object is perceived. Using this artificial memory, the robot can then advise users as to where an item was last seen.

The team used a Fetch mobile manipulator robot, which is equipped with a camera to perceive and navigate through its surroundings. This system was then integrated with a graphic interface that allows users to select particular items for tracking, with a log of these items then added to a user’s smartphone or computer. The combination of the two equipped the robot with an episodic memory that allowed it to identify and track selected items.

"The long-term impact of this is really exciting," said Ali Ayub, study co-lead. "A user can be involved not just with a companion robot but a personalized companion robot that can give them more independence."

While the team said the design was initially intended for patients suffering from dementia, they said it could be extended to anyone with memory issues or simply those prone to misplacing items.

Initial tests of the design were successful, with the next steps being to test it in real-life trials, first in non-dementia sufferers, and then patients with the condition.