A comprehensive multi-factor analysis on RFID localization capability

The concept of using radio frequency identification (RFID) devices to obtain the location information of objects is novel, and it has great potentials for supply chain applications. The capability of RFID localization, reflected by localization accuracy, is a fundamental issue. This paper presents a comprehensive analysis on how the accuracy is affected by multiple factors, which include region geometry, target distribution, ranging-error distribution, and landmark layout. The performance metrics are the expected mean squared error (MSE) and the expected Cramer-Rao lower bound (CRLB) over the entire localization region. The optimal landmark layouts were first obtained for all the combinations of other factors, and then the impact of individual factors on the localization accuracy was analyzed and discussed. By combining the theoretical analysis and Monte Carlo simulation, it was discovered that (1) the optimal landmark layouts follow simple empirical deployment rules; (2) the performance generally improves as the target distribution is more centering in the region; (3) the performance for rectangular geometries decreases as the aspect-ratio increases; (4) a higher landmark number can improve localization accuracy, but the beneficial result becomes negligible if more than 8 landmarks are used. Moreover, it was demonstrated that the results from a preliminary RFID localization experiment are in agreement with the findings from the simulation.