GNN data association based multi-target localization and vital signs monitoring using SIMO-UWB sensor network

Abstract

This paper presents a method for multi-target localization and vital signs monitoring using impulse ultra-wideband (UWB) technology. A single-input multiple-output (SIMO) sensor network is employed to enable simultaneous signal reception, which introduces challenges in signal separation and precise target positioning. The primary difficulty arises from the data association problem, where measurements must be correctly assigned to their respective targets, particularly in cases of overlapping signals or closely spaced targets. To address this issue, a time of arrival (TOA) algorithm is applied to estimate target ranges, while all possible measurement–target associations are evaluated. Subsequently, true target positions are obtained through trilateration combined with the global nearest neighbor (GNN) method. For vital signs monitoring, the continuous-time Fourier transform(CTFT) is utilized to estimate respiratory and cardiac rates. Experimental results demonstrate high accuracy, with relative errors of 1.42% in distance estimation, 2.34% in breathing rate, and 0.73% in heart rate estimation.