Security and collision in RFID systems

Abstract

Listen

Translate article

Abstract Radio Frequency Identification (RFID) is a promising technology to provide automated contactless identification of objects, people and animals. The identification process is performed as the reader receives simultaneous responses from various tags over a shared wireless channel and without no requirement of line-of-sight in the interrogation zone. The communication between the reader and tags is separated into two processes: identification and acknowledgment processes. Both processes suffer from serious drawbacks that limit the proliferation of RFID. Such drawbacks are security and privacy and collision problems. This thesis has two main parts. The first part examines the security and privacy of the existing RFID authentication protocols. We introduced a novel cryptographic scheme, Hacker Proof Authentication Protocol (HPAP) that allows mutual authentication and achieves full security by deploying tag static identifier, updated timestamp, a one-way hash function and encryption keys with randomized update using Linear Feedback Shift Register (LFSR). Cryptanalysis and simulation show that the protocol is secure against various attacks. In comparison with the various existing RFID authentication protocols, our protocol has less computation load, requires less storage, and costs less. The second part focuses on solving RFID collision arbitration imposed by the shared wireless link between a reader and the many tags distributed in the interrogation zone. In most proposed anticollision algorithms, tags reply randomly to the time slots chosen by the reader. Since more than two tags may choose the same time slot in a frame, this Random Access (RA) causes garbled data at the reader side resulting the identification process fails. Towards this challenge, two ALOHA based anti-collision algorithms that adopt a new way for tags to choose their replied time slots to enhance system efficiency are presented. In MBA and LTMBA, tags use modulo function to choose their owned time slot. The difference between the two algorithms relies on the method by which the reader estimates the next frame size. The performance evaluation of the two algorithms shows better performance than previously proposed algorithms in terms of fewer communication rounds and fewer collided/empty slots considering the limitation of the EPCglobal Class-1 Gen-2 standard.