Research results indicate that it is possible to predict falls (fall injuries) in the elderly using artificial intelligence (AI). By analyzing behavioral patterns such as sleep and activity levels in the elderly, as well as environmental information like temperature and humidity, it can provide real-time alerts about the risk of falls.
Sungkyunkwan University announced on the 7th that "Professor Jeong Jo-woon of the Department of Electrical and Electronics Engineering has designed a multimodal deep learning AI that predicts fall accidents with 98% accuracy by analyzing behavioral patterns and environmental information of the elderly using sensors." The research findings were published in the Journal of Biomedical and Health Informatics (JBHI), issued by the Institute of Electrical and Electronics Engineers (IEEE), in May.
Falls are a common ailment among the elderly. According to data released by the Korea Disease Control and Prevention Agency in 2021, over 80% of patients hospitalized due to falls are aged 65 and older. As people age, muscles and bones weaken, resulting in long-lasting aftereffects from falls. Especially, the incidence of fall accidents is significantly higher indoors than outdoors, with most falls occurring during daily activities.
The AI developed by the research team visually demonstrates how changes in activity levels and indoor environment increase the risk of falls at specific locations based on information obtained from non-contact sensors installed indoors. It also provides fall prevention strategies tailored to individual living characteristics, which can be utilized by caregivers or care institutions.
This study was conducted as part of an industry-academic collaboration. The research team utilized life log data amounting to 900 million records monthly from the smart care corporation RibbonCare. It also incorporated the environmental and health design expertise of Professor Lim Ri-sa from the Korea Advanced Institute of Science and Technology (KAIST).
Sungkyunkwan University plans to expand research into various fields such as elderly care using people-centered AI technology. Professor Jeong Jo-woon noted, "A new paradigm for elderly care based on smart homes has been proposed."
References
IEEE JBHI (2025), DOI : https://doi.org/10.1109/JBHI.2025.3569891