Real-time wireless UWB sensor network for person monitoring

Abstract

Ultra-wide band (UWB) radar is a very perspective technology for short-range localization and tracking of moving persons. UWB radar sensor network (UWB-SN) employing the centralized data fusion method can significantly improve tracking capabilities of more people in complex building environments. In this paper we present real-time performing wireless UWB-SN hardware demonstrator. Its sensor nodes use M-sequence UWB radars front-end and low-cost ARM based quad-core microcomputer (ARM-MC) as a main signal processing block. The ARM-MC based on Raspberry Pi provides processing power for the preprocessing of received raw radar signals, algorithms for detection and estimation of target's coordinates, and finally compression of data sent to the data fusion center. Low-rate data streams (3600–6000 bits/s/node) of compressed target coordinates are sent from each sensor node to the data fusion center in the central node by using RF transceivers with integrated ARM microcontroller. The proposed UWB-SN uses wireless RF communication by using FSK modulation in free short range devices (SRD) 868–870 MHz frequency band concurrently with operation of the UWB radar front-end. Experimental testing confirmed real-time performance of developed UWB-SN hardware and acceptable precision of software algorithms implemented in 32-bit arithmetic. The introduced modular architecture of UWB-SN can be used for fast development and testing of new real-time people tracking applications.