Real-time and Accurate RFID Tag Localization based on Multiple Feature Fusion

We propose a new radio frequency identification (RFID) localization approach that achieves both low latency and high accuracy by fusing multiple type of signal features. Existing RFID tag localization approaches either suffer from large localization latency (e.g., approaches based on phase measurements), or cannot provide high localization accuracy (e.g., approaches based on received signal strength (RSS)). We propose a two-step approach that fuses phase measurements and RSS measurements to resolve this dilemma. First, coarse-grained RSS measurements are utilized to Figure out a small bounding box that encloses the position of the target tag. Second, fine-grained phase measurements are used to refine the position estimation of the target tag in the bounding box. Experimental results show that the proposed fusion approach achieves centimeter-level localization accuracy with less than 10 feature measurements, reducing localization latency by more than one order of magnitude when compared to state-of-the-art solutions.