| Summary: | How to improve localization accuracy is a big challenge for highly dynamic and sparse industrial scenarios with active RFID tags. Since antenna of active tag is anisotropic, its emitting signal propagates damply with transmission distance and emitting orientation. In this paper, we aim at modeling anisotropic signal attenuation of active RFID tags by analyzing measurement data in real environment. As the features of signal attenuation with transmission distance on different signal-emitting orientations are the same, two basic models are regressed using experimental data firstly: 1) directional signal-distribution models for both horizontal and vertical orientation in a certain distance; 2) an attenuation model of RF signal with transmitting distance along one direction. Afterwards, an Anisotropic Signal Attenuation Model of active RFID tag (ASAM) is deduced. Furthermore, a noise filtering model in a tag-grid environment is optimized for the spatial model ASAM. Finally, the experimental results in 400-square-meter experimental field show that the average standard deviation (STD) of the optimized model reduces by 50% when the STD is bigger than 4-dB, and the probability distribution is over 70% when the deviation is less than 2.
|
|---|
