A Multiple Baseline Approach to Face Multipath

The positioning of tagged objects by means of the phase of the backscattered RFID signal is a challenging problem. Several approaches have been conceived in the years to face the problem of the narrow bandwidth of the RFID signal, which produces a periodic, and hence ambiguous, measurement of the tag-reader distance. One possible strategy is to exploit the movement of either the tag or the reader to collect a set of phase measurements which allow to solve the aforementioned ambiguity: this is, e.g., the approach pursued by SAR (Synthetic Aperture Radar) methods. In this article we propose an alternative approach which elaborates the sequence of phase samples in order to discover additional indicators that allow to improve the effectiveness of synthetic aperture methods. In particular we develop a strategy that permits to identify, among a set of measurements, the ones more strongly affected by the multipath. If a pattern emerges among the measurements without significant multipath (and this occurs if a low to moderate level of multipath characterizes the environment), by exploiting the good subset of measurements, it is possible to achieve an improvement in the positioning estimation. The method has been developed for a straight line trajectory: numerical results are reported in this article to illustrate the approach, while experimental results confirm its effectiveness.