SleepPal: A Novel System for Elderly Sleep Monitoring and Bed Falls Detection

Long term sleep monitoring has become an important mean for healthcare professionals to investigate the correlation between nocturnal activities and sleep disorders which may induce incidents such as falling from bed. In fact, bed falls can cause several psychological and physical problems such as injuries, trauma, and fear of falling. Early detection of falls enables prompt patient rescue and preventing more severe injuries. This paper introduces SleepPal_a novel and unobtrusive sleep monitoring system for body movement, sleep posture detection, and bed fall detection. The system consists of a wearable compact device that retrieves motion data from an inertial measurement sensor and transmits it to a remote monitoring station. The identification of body movements, posture recognition, and bed falls detection was accomplished through the utilization of a machine learning algorithms. The experiments involved 15 subjects, and the results demonstrated an average accuracy of 98.7%, an average sensitivity of 98.5%, an average specificity of 98.9%, and an average precision of 98.9%.