A team at the University of California San Diego has developed CARMEN (Cognitively Assistive Robot for Motivation and Neurorehabilitation), a tabletop robot designed to assist individuals with mild cognitive impairment (MCI). This robot aims to help users enhance their memory, attention, and executive function skills through interactive training exercises.
CARMEN, unique in its field, was created in collaboration with clinicians, MCI patients, and their care partners. It focuses on teaching compensatory cognitive strategies, a distinction from other assistive robots. According to Professor Laurel Riek, senior author of the work, the goal was to develop practical tools that make a meaningful impact on users’ daily lives.
MCI is a condition that sits between normal aging and dementia, affecting memory, attention, and executive functioning. It affects about 20% of individuals over the age of 65, with up to 15% progressing to dementia each year. Current pharmacological treatments have not been successful in preventing this progression, but behavioral treatments offer some benefit.
CARMEN provides cognitive training exercises such as creating routine places for important objects and using note-taking strategies. These exercises are delivered through interactive games and activities, designed to be user-friendly and functional without requiring constant internet connectivity or extensive maintenance. The robot was also designed to communicate clearly, express empathy, and provide breaks to maintain user engagement.
The research team tested CARMEN in the homes of individuals with MCI and clinicians experienced with MCI patients. Participants engaged with the robot over a week, performing tasks like organizing household items and using calendars for memory aid. Feedback indicated that users found the robot easy to use and that it boosted their confidence in applying cognitive strategies in daily life. Participants expressed a desire for more interaction with CARMEN.
The findings, which were presented at the ACM/IEEE Human Robot Interaction (HRI) conference in March 2024, earned a best paper award nomination. The next phase of research involves deploying CARMEN in more homes and enhancing its capabilities to include conversational interactions while maintaining user privacy. Researchers are also exploring the robot’s potential applications for conditions such as ADHD.
CARMEN was built using the FLEXI robot from the University of Washington as a base, with significant hardware modifications and entirely new software developed by the UC San Diego team using the Robot Operating System (ROS).
Photo: credit David Baillot/University of California San Diego
