Design of accelerating techniques for performance issues in RFID systems

Abstract

This thesis examines the performance issues of Radio Frequency Identification (RFID) Technology. The RFID technology has recently received an immense attention as an augmentation technology in the field of ubiquitous computing. This upcoming technology holds a potential to make phenomenal impact on many industries. It is widely in adoption due to its lucrative benefits over other existing identification technologies, since it does not require alignment for line-of-sight, read through obstacles and read in real time. Apart from the positive aspects offered by the technology, there are some practical issues associated with RFID systems which are preventing this technology to be accepted and adopted in a phenomenal manner. These issues affect and hinder the performance of RFID systems. This thesis examines such performance issues faced by RFID technology. The issues investigated in this thesis include uncertainty in RFID data, waiting time in RFID reader systems and tag-tag collision. This thesis proposes a fuzzy technique for handling the uncertainty issue for monitoring and controlling the uncertainty occurring in the raw RFID data based on local RFID system environment knowledge. This thesis also proposes a genetic algorithm namely RFGeni for the waiting time issue which optimizes the overall time required including the pre-scheduling of the RFID reader commands. Further, the thesis examines the anti-collision issue and proposes a Code Division Multiple Access based anti-collision deterministic algorithm for passive RFID tags (because of their most wide usage and low cost) and investigates its performance comparison with the ISO 18000-3 mode 1 and ISO 18000-3 mode 1 “extended mode” protocols. The data load impact on the existing RFID protocols stated earlier has also been evaluated on identification time for varied number of RFID tags. This thesis also proposes architecture for RFID systems accommodating the proposed accelerating techniques. Also, the proposed architecture has been implemented for one of these accelerating techniques namely uncertainty in RFID data as working RFID model at application layer. This architecture assumes inputs from an RFID emulator in the RFID sensing scenario to appreciate the reduction of cost and time. The data processing layer of this architecture inside the application layer is embedded with intelligence and optimizations as a part of this work. This thesis also presents the implementation results of the proposed architecture in order to support the existence and further scope of implementations in RFID systems. Also, in the proposed architecture the accommodation and validation of the uncertainty issue has been achieved. The proposed architecture is also compared with the existing architectures as a part of this work.