Abstract:

To overcome the difficulty of relative position estimation between RFID tag and RFID antenna, an RFID tag localization method was proposed for RFID based object localization in intelligent space. A kind of multipower RFID tag recognition rate model was established and the upper and lower bounds of statistical multipower RFID tag recognition rate were obtained. Then the RFID tag localization problem with multipower RFID information was converted into uncertain information fusion problem. Bayesian estimation was used to construct basic belief assignment function, and DSmT generalized fusion machine was used for multi-power RFID tag localization information fusion. The random particles in the identifiable area, which indicated the position of RFID tag, were weighted with the information fusion to estimate the position of RFID tag. The experimental results demonstrated that the RFID tag localization errors could be less than 025m, and efficienty of object searching was increased by 60％ comparing with conventional methods, which met the demand of object localization.

Key words:  Service robot, intelligent space, RFID, localization, DSmT