Hyperbolic Positioning and Tracking of Moving UHF-RFID Tags by Exploiting Neural Networks

In this paper we propose a novel real-time tracking method of a moving UHF-RFID tag. Two antenna pairs fixed at predefined positions monitor the moving tag collecting phase measurements. Phase differences are calculated for each pair and then mapped to distance-differences of the two antennas from the target tag. The latter corresponds to a hyperbola for each pair of antennas. The intersection of the two hyperbolas denotes the position of the tag. Since solving the system of hyperbolas to find the intersection point is not feasible with standard practices, a neural network is exploited to approximate the solution. Compared to prior art, the proposed method does not require knowledge of the tag's initial position or the trace followed (e.g. conveyor belt). Experiments were conducted by placing a tag on a moving robot capable of performing SLAM (Simultaneous Localization and Mapping) to ensure knowledge of the ground truth. The results showed a mean error under 0.5m throughout the experimental campaign.