Joint diagonalization in ACMA and its application to multiple-tag RFID systems

This paper discusses the separation of signals in multiple-tag radio frequency identification (RFID) systems. First, a model for the RFID system in both single and multiple tag environments is presented. Then, an analytical constant modulus algorithm (ACMA) for the blind source separation problem is reviewed. An alternative approach to the traditional ACMA using joint diagonalization is considered. Finally, both the ACMAs are applied to the multiple-tag RFID environment and performance of the system is studied. Simulations are carried out for 4-QAM and 16-QAM modulations at tag and analyses of the simulation results reveal that the ACMA with joint diagonalization takes lesser CPU time to execute compared to the traditional ACMA, but with higher average modulus error (AME). The variation of system performance with the number of measurements, SNR (signal to noise ratio), number of tags and number of antennas at the reader is also studied. Based on these, some design guidelines are presented. Interestingly both the ACMA algorithms work for the 16-QAM case and yield trends similar to those of 8-PSK and 4-QAM.