Tag size profiling in multiple reader RFID systems

In this paper, we study the tag size profiling (TSP) problem in RFID systems with multiple readers, which is to estimate the number of tags in every subregion in the system covered by different set of readers. The TSF problem is vitally important to reader scheduling and many other related operations in large scale multi-reader RFID systems. To our knowledge, however, it is not well solved in previous researches. We propose a novel approach to the TSF problem. The key idea is to treat the size of subregions as variables and construct a linear system in these variables to solve them. We theoretically prove that for any multi-reader RFID system, a linear system that can be used to uniquely solve the variables corresponding to the subregion sizes can always be constructed. We then propose a time-efficient algorithm that uses two heuristics to quickly find enough linearly independent equations to construct the linear system. Extensive simulation results show that the proposed approach achieves very high accuracy. When the estimation results of individual readers contain 5% errors, our approach achieves median estimation error of smaller than 0.02 and 90-percentile estimation error of smaller than 0.04 in large systems containing more than one hundred readers.