Abstract:

Radio frequency identification (RFID) location based on received signal strength is a low cost, easy to implement indoor positioning method. In view of the small sample problem existing in the RFID indoor positioning system when using reference tags, we propose a RFID indoor localization algorithm based on support vector regression (SVR). Combined with the principle of passive ultra high frequency RFID system, the performance of the proposed algorithm is tested in MATLAB and then compared with that of the classical LANDMARC method. And the influence of mutual coupling effect and multipath effect on the positioning results of the algorithm is also tested. Simulation results show that the proposed method can improve positioning accuracy by at least 25% in comparison with the LANDMARC method while not increasing the number of reference tags.

Key words: