Smart sensors of RF and backscatter signals with localization

Abstract

Information may be more valuable when the location of its source is known. This thesis develops localization algorithms based on received signal strength (RSS) measurements for unknown radio frequency (RF) sources and bistatic scatter radio tags (sensors). This thesis demonstrates RF source location es- timation utilizing RSS measurements by a community of smartphone users, within 800m (or more) from the source. The location estimation algorithm incorporates careful modeling of the time-varying source transmission power, source antenna directionality (even with a single 4-parameter model) and different path loss exponents among the various source-user links. More im- portantly, a vast number of measurements is collected and exploited through an automated community of smartphones. Location estimation error on the order of 50m is achieved, even when users are within 800m or more from the RF source. Furthermore, multiple input single output RSS localization for bistatic scatter radio is also considered. The RF scatter radio tag is illuminated by multiple low-cost carrier emitters, operating consecutively. Experimental validation of the proposed algorithm reports localization error on the order of 3m for tag and emitters placed at an area of 70m x 70m. Both estimation algorithms on real-world data exploit non-parametric estimation based on particle filtering.