James M. Keller, Professor of Electrical and Computer Engineering, has been engaged in interdisciplinary and collaborative research throughout his career. Currently, he is working on a project that draws upon the latest technological advances to improve elder care with a team led by fellow electrical and computer engineer Marjorie Skubic and a group of people from MU’s Schools of Nursing, Social Work, Health Management and Informatics, Physical Therapy, and Engineering, along with colleagues from the Medical Automation Research Center (MARC) at the University of Virginia. The project is being developed and tested at TigerPlace in Columbia, an assisted living facility designed by the School of Nursing and Americare Corp. to provide independence and quality of life for elders while offering necessary assistance, monitoring, and care. The research team is exploring the challenge of how seniors can be both independent and yet safe, and how technology might be applied to support that effort.
Some of the strategies the team has devised include installing passive sensors in residential units at TigerPlace to determine whether the residents are engaging in normal activities or whether subtle signs of decompensation call for early intervention. Types of sensors include motion detectors, vibration sensors on the floors to detect falls, heat sensors around the stove area and refrigerator door, sensors in the bed to measure respiration, heartbeat, and restlessness, and even silhouette video sensors that can be used to build models of normal versus deviation-from-normal behaviors.
As it turns out, Keller’s expertise in image processing and pattern recognition, as well as his background in employing tools such as "fuzzy logic" and neural networks, is useful when applied to elder care. Invented by Lotfi Zadeh in 1965, "fuzzy logic" is a precise mathematical theory that deals with vague events, taking into account varying degrees. When applied, fuzzy logic essentially makes computers behave more like humans, for instance, acknowledging "shades of gray" in activity and in decision-making. This becomes important in elder care because seniors do not generally function well one day and suddenly decompensate the next, but rather tend to fail gradually. Instead of waiting for a catastrophic event such as a fall, fuzzy logic may be able to detect how well a person is doing (relative to her or his normal state) and predict when a fall appears to be imminent by watching the "grey areas" over time. This innovative collaboration promises to work simultaneously to improve the care of seniors and to prolong the quality of their lives as long as possible.