Theoretical Performance Evaluation of UHF-RFID Systems With Multi-Antenna Maximum-Likelihood Decoding

Radio frequency identification (RFID) is a technique that enables autonomous identification of objects using tiny and cheap transponders (tags). Therefore, its use increases, especially in the area of logistics. However, in dense applications scenarios with many tags multiple tags may transmit their data at the same time, leading to collisions. These collisions, especially in case of UHF RFID, significantly reduce the system throughput. The use of multiple input-multiple output (MIMO) is one candidate to achieve a high throughput even in case of tag collisions. Within this paper, we therefore analyze the performance of maximum-likelihood decoding for UHF RFID with multiple reception antennas. We especially take into account the impact of the FM0 channel encoding used by the EPCglobal Class-1 Gen-2 standard, the most common UHF RFID standard in the area of logistics. The channel characteristics of passive RFID systems follow the so-called double or backscatter Rayleigh distribution, which makes the performance analysis a challenging task. We use two methodologies to analyze the performance in this. First, we derive an upper bound symbol error probability (SEP) based on the pairwise error probability (PEP). The second technique depends on the exact evaluation of the SEP by taking into account all probabilities of errors between symbols. We finally compare the accuracy of both methodologies with simulation results. These results show that the PEP analysis converges with increasing number of antennas, and the exact analysis gives accurate results even with only two receive antennas.