LAUSR

Accurate tag quantity estimation is a prerequisite to maximize the throughput of radio frequency identification (RFID) systems. Previous estimators, mainly designed for single-antenna RFID systems, often suffer from performance degradation in low signal-to-noise-ratio (SNR) regimes, making them inappropriate for multi-antenna RFID systems where received tag signals are likely to overlap. In this regard, a high-dimensional tag quantity estimator is proposed in the multi-antenna context by exploiting the spatial diversity at receive antennas. We first show that the collided tag signals can be rearranged as high-dimensional vectors, whereby the tag quantity estimation problem can be modeled as a high-dimensional data clustering one. We next prove that when the SNR on each backscattering subchannel is greater than 3 dB, the distance incrementation between clusters offered by the modeling advantage benefits their separation. This finding encourages us to integrate the density-based spatial clustering of applications with noise (DBSCAN) algorithm with this high-dimensional space for tag quantity estimation, and its superiority over several existing approaches are supported by both synthetic and real-world case studies.