Abstract
Lower-Limb Fractures (LLF) are a major health concern for older adults, oftenleading to reduced mobility and prolonged recovery, potentially impairing dailyactivities and independence. During recovery, older adults frequently facesocial isolation and functional decline, complicating rehabilitation andadversely affecting physical and mental health. Multi-modal sensor platformsthat continuously collect data and analyze it using machine-learning algorithmscan remotely monitor this population and infer health outcomes. They can alsoalert clinicians to individuals at risk of isolation and decline. This paperpresents a new publicly available multi-modal sensor dataset, MAISON-LLF,collected from older adults recovering from LLF in community settings. Thedataset includes data from smartphone and smartwatch sensors, motion detectors,sleep-tracking mattresses, and clinical questionnaires on isolation anddecline. The dataset was collected from ten older adults living alone at homefor eight weeks each, totaling 560 days of 24-hour sensor data. For technicalvalidation, supervised machine-learning and deep-learning models were developedusing the sensor and clinical questionnaire data, providing a foundationalcomparison for the research community.