Researchers discussed the potential of robots to serve as coaches and companions to the elderly, during an Annual Meeting session. | USC Viterbi School of Engineering
Even as robots are increasingly able to drive our cars, clean our homes and help us care for our children – the goal for three researchers focused on the elderly is to design robots and cyber systems that do not do tasks for people, but instead keep the elderly active, safe and independent.
“What is talked about in robotics almost exclusively in terms of automation, of replacing human work,” said Maja Mataric, a professor at the University of Southern California and co-director of its robotics lab, during a press briefing at the 2017 AAAS Annual Meeting on 17 February.
Mataric’s lab has developed a robotic assistant that can serve as a companion or a rehabilitation coach to allow the elderly and disabled to remain in their homes. “The issue of motivating people to do their own work is critical. When we do our own work, we live longer, we feel better.”
Long-term care providers served about 9 million people in the United States in 2014, the majority of which were age 65 and older, according to the Center for Disease Control. The panelists are working to introduce robotic and cyber systems to help adults avoid such facilities.
The population of Americans over 65 is also projected to more than double in the next few decades, reaching over 98 million by 2060. In an aging country, innovative robotic systems offer the elderly an opportunity to “age in place.” Given the significant costs of hospitalization and long-term care, such an option can be particularly advantageous at a time when financial resources can be limited.
“Research has shown that if you can improve your cardiovascular health and keep physically active and socially engaged… it has great implications for how you’re going to age,” said Marjorie Skubic, a professor in electrical engineering and computer science at the University of Missouri. “And the question becomes, how can robots help?”
Skubic said clinical practitioners often see aging in a “step stair” fashion: elderly people remain on a plateau of functionality until an event — like a stroke, or a fall — causes a significant decline in health, dropping them to a lower plateau and the process then repeats.
Skubic’s intervention takes the form of a sensor system installed in homes to monitor a patient’s normal movements. The system recognizes shifts in behavior, such as a shorter gait – a sign that can predict a potential fall three weeks before it occurs – and issues a health alert to the patient’s doctor or social worker. In addition to testing the system at a local elder care facility, Skubic recently installed the system in her parents’ home for her mother’s 93rd birthday.
“If you can introduce a treatment at the right time, you can keep people up at this high level of functionality in the end,” Skubic said. “It’s not about extending someone’s lifetime, but extending the amount of time they have good health and independence.”
For Ruzena Bajcsy, a half century of studying robotics, computer vision and artificial intelligence has transformed her work into using such systems to help the elderly. Bajcsy, 83, and still teaching at the University of California at Berkeley’s Electrical Engineering and Computer Sciences school, is helping develop robotics to facilitate what, physically, a person cannot do any longer.
Bajcsy uses non-invasive observations of movement and devises algorithms to model the features and properties of patients’ movements, known as kinematics. Mechanical engineers can then use these models to design assistive devices.
“I want to know, how far can you reach? Can you close your bra? Can you wash your face?” Bajcsy said. “It is a diagnostic procedure that I developed. If you can’t open a jar, you need more torque. We will design a device to give you a little more oomph.” Like Skubic, one of Bajcsy’s goals is to detect patients’ limitations to help prevent injuries and permit subjects to get physical therapy.
For patients who have already suffered an injury or health condition, Mataric’s robotic companion encourages patients to work on regaining lost mobility. The programming used to direct assistive robots is based on a decade of social science research on human behavior. It uses jokes and encouragement to inspire patients. Studies on wearable technologies have shown that data alone usually is not enough to encourage people to change their behavior.
“Feeling responsible to another entity — a friend, a buddy, or a robot — makes all the difference,” Mataric said. “If you look at what drives us as humans, the social component is the only reliable thing that makes people change behavior. But there aren’t enough people for people, so that is where we need to introduce technologies and technologies that are fundamentally social.”
[Associated Image: USC Viterbi School of Engineering]