Georgia Institute of Technology
Mild Cognitive Impairment (MCI) is an early-stage decline in memory, language perception, visual perception, and/or other cognitive abilities that is greater than expected for an individual’s age and education level, but that does not greatly impact their ability to perform everyday activities. Research has shown that this condition is present in 3-19% of people over the age of 65. One common symptom of this condition is misplacing objects due to the inability to recall where the item was last kept. This could cause individuals living with MCI to feel distressed, anxious, and even create tension between them and their care partner/giver. To address this issue, we are developing a robot to keep track of commonly misplaced objects and recall their location upon request.
Our preliminary implementation requires the robot to routinely patrol the home based on a pre-generated map and capture images of pre-set waypoints. The objects in these images are classified with an object detection model, and their names, timestamps, and locations are stored in a JSON database. Upon request by the user, the database is scanned for all instances of the missing object, and the associated images are displayed on a screen for the user to confirm that it is indeed the object they are searching for. Finally, the robot may either lead the user to the object or provide oral directions.
We are currently working on validating our current prototype through user studies, and on improving the capabilities of the robot - in terms of learning user patterns from the data stored and recognizing partially occluded objects.
The goal of this research project is to develop a system that will improve the quality of living of MCI patients by restoring their sense of independence through collaboration between them and the robot. In summary, this project utilizes the decision-making skills of the human and the robot's perfect memory to find misplaced objects.