Signal processing methods for chipless RFID

Radio frequency identification (RFID) is a technology that automates routine procedures of data extraction, identification, tracking and surveillance in applications such as inventory control and logistics. The unit cost of conventional RFID tags is too high for them to be used in large item level tagging applications. This is because of the expensive electronic integrated circuits (ICs) used in the tags. As a solution to further reduce the cost of RFID tags, chipless RFID tags have been developed. A chipless RFID tag does not require an IC for its operation. Current research on chipless RFID technology has been focused on, the development of tag designs with enhanced data capacity, the development of tags with sensing capabilities, and the development of RFID reader architectures and signal processing algorithms. Despite current research efforts, further work is required in the area of signal processing for chipless RFID. For this purpose, three novel signal processing methods are introduced in this thesis, (i) development of a robust multidimensional detection algorithm for detecting data bits encoded in a chipless RFID tag, (ii) time and frequency domain analysis of backscattered tag signals for the removal of interference, and (iii) a new systematic calibration procedure for single antenna RFID readers. These methods enhance the performance of chipless RFID systems in terms of the data bit detection and reading range. Existing algorithms used for detecting data bits encoded in a frequency signature of a chipless RFID tag use a one dimensional approach to detection. The one dimensional approach to detection does not consider all the characteristics of the spectral features that encode data bits in a frequency signature. Therefore, the detection performance achieved is suboptimal. In order to enhance the detection performance, a new multidimensional detection method is introduced. The new method utilizes a set of orthonormal basis functions to fully describe the characteristics of a frequency signature. Using these orthonormal basis functions a frequency signature is represented as a signal point in a multidimensional signal space. The detection of data bits contained in an unknown frequency signature is performed using minimum distance detection. It is shown that the performance achieved by the new method exceeds the performance of existing one dimensional threshold based detection of tag data bits. The second method proposed in the thesis focuses on improving the reading range of an RFID reader beyond proximity based reading. For this purpose, the total received signal at an RFID reader is analysed in the time domain as well as the frequency domain to identify the essential signal component that contains the tag data. It is shown that the useful data is contained in the antenna mode of the backscattered tag response. The antenna mode backscatter is separated from the rest of the received signal using a time window. The separated antenna mode is then analysed in the frequency domain to estimate the tag’s frequency signature. Through this time and frequency domain analysis, non-proximity based reading is achieved. It is shown that the tag can be read in non-proximity reading conditions using simulation results and measurements taken in an anechoic chamber environment. The final method introduced in the thesis is a systematic calibration procedure for single antenna based chipless RFID readers. The calibration procedure takes into account practical conditions prevailing in a real world application environment. The calibration allows the RFID reader to accurately estimate the frequency signature of a chipless RFID tag in a cluttered environment. It also addresses the limitations of existing calibration methods used for chipless RFID systems such as the need for repeated calibration and antenna alignment.