Research On Phase Information Based Passive UHF RFID Localization Algorithms

Passive UHF RFID tags were highly concerned for the advantages of small size, low price and easy deployment, especially that tag based indoor localization technology was a hot spot in current research. With the complexity of indoor environments, RFID radio signal measurement information has great errors caused by the influence of noise and multipath from the enviroments, and eventually the localization accuracy of passive UHF RFID system was affected severely. To improve the accuracy of indoor localization, the studies of phase based passive UHF RFID localization algorithms are carried out in this thesis. The main work and innovations of this thesis are summarized as follows:A ranging algorithm named Refer-All-phase-FFT, based on the phase information of passive reference tags is proposed. By calculating the similarity between all reference tags and target tag about the phase information when communicating with one reader, the most nearest neighbor reference tag is selected and the line of sight distance between the reader and the reference tag is chosen as the distance between target tag and reader. After that, the final localizaiton estimation is implemented by linear least square algorithm. The simulation results show that the final localization error of Refer-All-phase-FFT is smaller than that of LANDMARC system, and the error is further reduced with an appropriate increase of the number of reference tags.A ranging localization algorithm based on improved weight factor of k-NN algorithm named Phase(improved k-NN) algorithm is proposed. In Phase(improved k-NN) algorithm, statistic theory is introduced into the selection of nearest neighbor reference tags. Then weight factors of k-NN algorithm in LANDMARC are calibrated by the calculation of the expectation of phase similarity. At last, the line of sight distance between reader and tag is obtained. The simulation results show that Phase(improved k-NN) algorithm is superior to k-NN algorithm for reducing the localization error remarkably.A new deployment of reference tag for ranging named arc deployment is proposed. In this new deployment, the distance similarity between reference tags and target tag is increased, and the interference between tags is reduced. The ranging error is thus reduced and the coverage of reference tag is also improved. Thereby, reference tag based ranging algorithm is more effective. Simulation results show that compared with traditional grid uniform deployment, the localization error of arc deployment is reduced, and the number of reference tag required is decreased. These results illustrate that arc deployment is superior to traditional grid uniform deployment.A double tag based two step ranging localization approach is proposed. When the side length of target object is larger than the wavelength of carrier signal, a passive tag deployment that double tags with a distance of the half wavelength are attached on the target object is introduced. During the communication with one reader, the triangular geometry relationship between the reader and two tags is utilized for localization. In the approach, the utilization of two tags can adjust the ranging information to reduce the localization error, as well as implement the localization estimation with one reader which decreases the hardware cost. Simulation results show that the approach can further reduce the localization error without reference tags.