Prediction System for Bed-Leaving Based on Center Position of Body Pressure on Bed

Nighttime patrols for patients with dementia are important for the prevention and early detection of wandering and bed falls. On the other hand, the nighttime patrol of patients requires a sense of tension and places a heavy psychological burden on facility staff. For this reason, systems have been developed to detect bed abandonment. However, since alerts are not triggered until after the patient has left the bed, there is a need to develop support devices that can determine the risk of leaving the bed at an early stage before the patient leaves the bed. Previous research analyzing the getting up of inpatients in long-term care wards suggests that there are two major patterns of movements leading to bed-withdrawal. The purpose of this study is to develop a system that can detect movements that are included in the two patterns and have a high risk of leaving the bed and predict the early stage of leaving the bed for patients with dementia. Eight pressure sensors are installed under the mattress of the bed to measure the position of the center of body pressure, and the system detects movements on the bed from the range where the center of body pressure is located. Based on a musculoskeletal model of the human body structure and kinematics, we formulated the system by considering the range of positions of the center of body pressure during each movement. This enables alerts to be issued in the pre-bed release phase based on the trajectory of the center of body pressure in the process leading up to bed release. Comparative experiments with existing products have shown the effectiveness of the proposed system in alerting the user to the danger of leaving the bed before the patient does. (Abstract)