Improving Group Integrity of Tags in RFID Systems

Checking the integrity of groups containing radio frequency identification (RFID) tagged objects or recovering the tag identifiers of missing objects is important in many activities. Several autonomous checking methods have been proposed for increasing the capability of recovering missing tag identifiers without external systems. This has been achieved by treating a group of tag identifiers (IDs) as packet symbols encoded and decoded in a way similar to that in binary erasure channels (BECs). Redundant data are required to be written into the limited memory space of RFID tags in order to enable the decoding process. In this thesis, the group integrity of passive tags in RFID systems is specifically targeted, with novel mechanisms being proposed to improve upon the current state of the art. Due to the sparseness property of low density parity check (LDPC) codes and the mitigation of the progressive edge-growth (PEG) method for short cycles, the research is begun with the use of the PEG method in RFID systems to construct the parity check matrix of LDPC codes in order to increase the recovery capabilities with reduced memory consumption. It is shown that the PEG-based method achieves significant recovery enhancements compared to other methods with the same or less memory overheads. The decoding complexity of the PEG-based LDPC codes is optimised using an improved hybrid iterative/Gaussian decoding algorithm which includes an early stopping criterion. The relative complexities of the improved algorithm are extensively analysed and evaluated, both in terms of decoding time and the number of operations required. It is demonstrated that the improved algorithm considerably reduces the operational complexity and thus the time of the full Gaussian decoding algorithm for small to medium amounts of missing tags. The joint use of the two decoding components is also adapted in order to avoid the iterative decoding when the missing amount is larger than a threshold. The optimum value of the threshold value is investigated through empirical analysis. It is shown that the adaptive algorithm is very efficient in decreasing the average decoding time of the improved algorithm for large amounts of missing tags where the iterative decoding fails to recover any missing tag. The recovery performances of various short-length irregular PEG-based LDPC codes constructed with different variable degree sequences are analysed and evaluated. It is demonstrated that the irregular codes exhibit significant recovery enhancements compared to the regular ones in the region where the iterative decoding is successful. However, their performances are degraded in the region where the iterative decoding can recover some missing tags. Finally, a novel protocol called the Redundant Information Collection (RIC) protocol is designed to filter and collect redundant tag information. It is based on a Bloom filter (BF) that efficiently filters the redundant tag information at the tag’s side, thereby considerably decreasing the communication cost and consequently, the collection time. It is shown that the novel protocol outperforms existing possible solutions by saving from 37% to 84% of the collection time, which is nearly four times the lower bound. This characteristic makes the RIC protocol a promising candidate for collecting redundant tag information in the group integrity of tags in RFID systems and other similar ones