In most of the existing localization approaches based on RFID phase measurements, the model of the phase of the received signal accounts for the distance but not for the orientation between the reader's and tag's antennas. This lack is not very relevant if both antennas are circular polarized and/or if they are not reciprocally rotating. However, in other conditions, like in robotic localization with antennas having different polarization, the dependence on the relative orientation becomes important. In this paper we analytically characterize this phenomenon and provide some experimental data to support the model. Then, to give an idea of the importance of modeling this phenomenon, we compare a localization algorithm based on an Extended Kalman filter, where this factor is included in the model, with an Extended Kalman filter where, on the contrary, this phenomenon is ignored. In this second case, the algorithm tries to cope with the offset variability by including the offset among the variables estimated but the performance of the approach where the phenomenon is considered remains significantly higher, as illustrated through numerical results. © 2022 IEEE.
Considering polarization mismatch in modeling the RFID phase offset variability for tag localization
Di Giampaolo E.;
2022-01-01
Abstract
In most of the existing localization approaches based on RFID phase measurements, the model of the phase of the received signal accounts for the distance but not for the orientation between the reader's and tag's antennas. This lack is not very relevant if both antennas are circular polarized and/or if they are not reciprocally rotating. However, in other conditions, like in robotic localization with antennas having different polarization, the dependence on the relative orientation becomes important. In this paper we analytically characterize this phenomenon and provide some experimental data to support the model. Then, to give an idea of the importance of modeling this phenomenon, we compare a localization algorithm based on an Extended Kalman filter, where this factor is included in the model, with an Extended Kalman filter where, on the contrary, this phenomenon is ignored. In this second case, the algorithm tries to cope with the offset variability by including the offset among the variables estimated but the performance of the approach where the phenomenon is considered remains significantly higher, as illustrated through numerical results. © 2022 IEEE.Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.