Fall Detection Utilizing Frequency Distribution Trajectory by Microwave Doppler Sensor

Falls are serious issues encountered in the lives of the elderly living alone. Since the elderly cannot stand up without support from a caregiver, or because they may lose consciousness after falling, they may remain on the floor for an extended period of time after a fall; this leads to serious complications including hypothermia, dehydration, and sometimes, even death. Therefore, immediate detection of falls is necessary. In this paper, we propose a fall detection system based on a microwave Doppler sensor. In the proposed system, we apply the frequency distribution trajectories corresponding to the velocities of the movements while falling, to a hidden Markov model. In order to evaluate the proposed system, we carry out verification experiments for three types of fall events (tripping, slipping, and fainting) and four types of non-fall events (walking, bending, sitting, and standing). From the results, the accuracy, positive predictive value, and negative predictive value are found to be 0.95, 0.94, and 0.97, respectively.