The increasing number of senior citizens who need long-term care or rehabilitation has become a serious issue, especially considering the increasing aging population and declining birthrate. To maintain one’s motor functions and improve one’s muscle strength or range of motion, it is important to continue exercising constantly. We have developed a prototype exercise support system that aim to promote and evaluate exercise of users. This system includes a small humanoid robot presenting the exercises, amotion sensor, a touch panel, a laptop PC to control other devices, and other parts. We propose a method of detecting the specific poses of the users from the skeleton data of the motion sensor. We use a questionnaire to evaluate the system’s effect on the promotion of exercise and the users’ understanding of the movements of the robot, and we use a motion sensor to evaluate motion recognition of the users during the exercises. The system is tested by young people and also elderly staying in a geriatric health service facility. The questionnaire results indicate that the elderly subjects react positively to the exercises. The pose detection method shows a correct answer rate of 94% for the young subjects and 87% for the elderly subjects. It is confirmed that the prototype system can be put into practice use.
DEVELOPMENT OF A COMMUNITY SUPPORT SYSTEM BY ACTION RESEARCH TO SUPPORT THE ELDERLY IN FRAILTY