Fig. 8 illustrates the localization accuracy of the proposed
system when the standard deviation (STD) of the RSSI varied
from 3 to 8 dB.With the STD variation of the RSSI, the iLocate
system showed a slow uptrend, whereas both Dash7 and ZigBee
exhibited a fast uptrend. Moreover, the maximum of the mean
localization error of the iLocate system was less than 0.2, indicating
that iLocate achieved higher levels of localization accuracy
compared with previous efforts. The reason is that iLocate
considerably eliminated the RFID RSSI noise by scheduling
communication using coordinators, and it obtained the finegrained
localization accuracy by introducing virtual reference
tags. In addition, a prototype project was implemented in a carton
packing factory, where each paper pillar was over 2500 kg
and was bigger than 3 m × 1.5 m × 2.5 m. The pillars were
placed in a layered manner. Fig. 9 illustrates the real deployment
of using the iLocate system (see Fig. 10 for the snapshot
of using the iLocate system for indoor positioning). Both
Alien 9900+ passive readers and Impinj R420 passive readers,
together with ultrahigh frequency tags with omnidirectional
antennas, have been tested. However, their recognition rates
were quite low. The Ekahau RTLS system has been also tested,