Ultrasound user-identification for wireless sensor networks

Indoor localization and asset tracking is an area of research which is finding increasing application in monitoring the wellbeing of elderly subjects in hospital or residential care environments. Radio frequency (RF) based localization methods tend to show poor performance in indoor environments, due to multipath scatter, and simple room-level misassociations, resulting from the relative transparency of some walls to RF waves. Ultrasound (US) localization has shown excellent room-level accuracy, and as US will not easily penetrate walls, room-level misassociations do not often occur. For normal asset tracking systems, the association between an asset and a room is not of critical importance; for clinical applications, which make inferences based on a patient and carer occupying the same room, this association is paramount. This paper outlines a proof of concept for an US-based indoor localization system with sub-room-level accuracy. The proposed design uses a US-based technique, with transmitter beacons placed throughout the environment. The subject wears the receiver module and is free to move throughout the environment. Transmission collisions are alleviated by assigning a unique prime transmission interval to each beacon; meaning the beacons do not require synchronization or coordination and operate independently of one another. Preliminary results show accurate detection at a range of over 5 m at a 0° angle and at approximately 3.5 m at an angle of 30°. Testing of two collocated transmitters, transmitting in the same plane and separated by 2.5 m results in accurate reception of both signals, even within the region of overlap.