Error control for radio frequency identification

This work proposes to use Forward Error Correction (FEC) for the tag-to-reader communication in long-range Radio Frequency Identification (RFID) systems, specifically in the widely used EPCglocal Class-1 Generation-2 (EPCc1g2) systems, to overcome some of the limitations that occur at extended communication distances. In a long-range RFID system, the reader must be able to identify a great number of tagged items in a reliable fashion, and in a limited amount of time. Cluttered propagation paths, difficult channel conditions, and the low power of the backscattered tag signals, make this task difficult. Especially, when the communication distances are high, as for battery-assisted tags, the signal-tonoise ratio at the reader’s receiver is low, and the bit-error probability is high. Nevertheless, the widely used standard protocols offer only rudimentary errorcontrol mechanisms. The use of FEC is not yet established in practical systems and has not been widely researched in the context of RFID. A model was developed, to describe the impact of errors on the duration of the tag identification, and it is subsequently used, to evaluate the use of convolutional codes and their benefits in terms of the improved identification time, related to the tag-reader distance. For a chosen code, it was shown that the range of the system can be extended by approximately 4m without increasing the identification time. Furthermore, the application of an iterative decoder architecture is presented that is used to decode the concatenated code, which is formed by the proposed FEC code and the standard EPCc1g2 baseband codes. An improvement of the coding gain of up to 2.8 dB could be shown in simulations, comparing the iterative decoder to a non-iterative one. Additionally, the convergence behavior of the decoder is analyzed and its limitations are discussed. Since a great number of RFID systems have already been deployed worldwide, standards and standard compatibility should be a great concern, when