Passive multi-target indoor localization based on maximum likelihood estimation of RFID interference

Passive multi-target indoor localization is one of the important fields recently and could contribute to many applications including storage management or somatosensory systems. In this paper, a passive multi-target indoor localization system was proposed using ultra-high frequency RFID. The reference tags were placed as UCA to detect the interference of the targets. The signal propagation was analyzed and the channel information of targets was extracted from complicated received signal. To distinguish the coherent signals, maximum likelihood algorithm was used to estimate the channel parameters of different targets. In order to obtain the estimation of the number of targets and reduce the cost of multi-dimensional optimization, weighted spatial averaging method was introduced and a virtual antenna array was constructed by antenna movement. The average of object function of different element of the antenna array was utilized to enhance the true weak positions and eliminate the false estimated positions. The targets tested in this paper is instant noodles and colas. The accuracy of target number estimation is 97.5%, the spatial resolution is about 50cm, and the median error of 2-Dimensional multi-target localization is about 5.5 cm.