Search

High accuracy localisation in a complicated and varied indoor environment such as a warehouse or a factory is one of the most important requirement and it is still challenging. In this dissertation, this problem is solved by designing a system which combines radio frequency identification (RFID) technology with a mobile platform and development of localisation methods. The system designed in this dissertation is composed of ultra-high frequency (UHF) RFID passive tags, a mobile platform and a computer. UHF RFID passive tags are either placed as reference tags or attached to objects as targets to locate. The mobile platform is comprised of a UHF RFID reader, a Raspberry Pi board, multiple antennas, a robot, and batteries. The computer is used to control the mobile platform remotely to collect information of tags. Both the low-cost of passive tags and the mobility of the mobile platform make the system suitable for indoor localisation in a warehouse or a factory. After designing the system, a novel inverse synthetic aperture radar (ISAR)synthetic aperture radar (SAR) localisation method is proposed and experimentally tested. In order to reduce the cost of devices for trajectory measurement, reference tags with known locations are used to estimate the trajectory of the mobile platform by the ISAR algorithm. A novel ISAR-SAR loop is introduced to find out the optimal estimated trajectory, which will be used to perform SAR algorithm and locate target tags. Experimental results of 2D localisation show that the ISAR-SAR method using a straight-line trajectory can achieve a mean absolute localisation error of 15 cm, which is similar to that using a traditional SAR algorithm and by the ISAR-SAR method using an L-shape trajectory, the error can be reduced to 8 cm, which is slightly smaller than light detection and ranging (LiDAR)-SAR method. The SAR-based algorithms can achieve high localisation accuracy but require the calculation of the probability function over a fine grid and this results in a high computational load particularly for 3D localisation, which is one of the main drawbacks of the SAR-based algorithms. This thesis also proposes and demonstrates a high accuracy localisation method with reduced computational burden based on the received signal strength indicator (RSSI) and the unwrapped phase profile. After measuring the phase and RSSI, a valid dataset can be obtained by analysing the received RSSI, the strength of which can indicate whether the signal is stable or not. The stationary point of the unwrapped phase profile combined with the known trajectory of the moving platform is used to estimate the distance along the trajectory, which is termed as cross-range. The distance perpendicular to the trajectory, which is termed as down-range, can be estimated by estimating the integer number (k-parameter) of wavelengths which fits the cross-range location and phase profile. For 2D localisation, only a single straight-line trajectory is required while in 3D space, multiple antennas at various heights are used and after obtaining xand y-coordinate of the tag by cross-range estimation with a L-shape trajectory, a possible range for the height of the tag will be estimated by received RSSI values and the accurate height can be calculated by the k-parameter estimation method. Experimental results demonstrates that the mean 2D localization error is around 12 cm and mean 3D localization error is around 14 cm. SAR-based methods typically require LiDAR sensors or high quality optical cameras to measure the trajectory so the localisation accuracy by these methods is affected by the accuracy of the measurement for trajectory while fingerprint methods require deployment of reference tags so the accuracy depends on the density of reference tags, which requires a lot of reference tags. This dissertation also propose a method that aims to reduce the number of required reference tags and reduce the requirement for devices to measure the trajectory by further exploiting the phase profile and analysing the relationship between tags and the trajectory of the mobile platform. To achieve 2D localisation, three reference tags with known locations and two non-collinear straight-line trajectories are required. The main idea of the proposed method is analysing the geometric relationship between the trajectory and tags using the minima of the unwrapped received phase profile. The direction of the trajectory relative to the reference tags is firstly determined and direction of two trajectories is used to calculate the location of target tags. Experiments show the mean 2D localization error is around 12 cm.

With increasing data throughput requirements, copper network access over twisted pairs is reaching data rate limitations. Fibre-to-premises deployment is however much slower, costly and time consuming than previously anticipated [1]. In the following research, we are exploring the potential of using surface wave mode transmission over existing cable networks, which may allow a significant increase in data rates whilst keeping the installation time and cost low. A characterisation of surface wave transmission was carried in comparison with the widely deployed G.fast standard. Surface waves, derived by Arnold Sommerfeld, have been defined as a low attenuation, low dispersion and very wideband electromagnetic mode transmission. In order to test surface waves' potential, a proof-of-concept experimental setup was built. The setup consisted in data generation and processing, surface wave launchers design as well as choosing the wired link (in this case DW11 cable). Several experimental transmission parameters were investigated such as the modulation, equalisation, bandwidth and carrier frequency with the aim of achieving high data throughput. Furthermore, capacity models were built to predict surface wave transmission performance in vari ous scenarios. Because of the lab dimension restrictions, a novel surface wave capacity calculation was developed to estimate data rates over an extrapolated channel response (using two length experimental responses), and obtain longer link realistic capacity estimates. Data rates of over 12 Gb/s were experimentally demonstrated on a 6.1 m DW11 surface wave link, with a BER of 2.25 ×10−5 , the signal was an OFDM modulated 64QAM transmission with 2080 MHz bandwidth and 2.1 GHz carrier frequency. In the extrapolation simulation, a DW11 surface wave link of 100 m predicted up to 14 Gb/s and at 50 m up to 35 Gb/s for high modulation and bandwidth transmission. Compared to the G.fast data rates (up to 2 Gb/s up to 50 m) [2], the results obtained tend to indicate that surface wave communications may have potential for future twisted pair access network technologies.

The low signal coverage problem has long been one of the most crucial problems in UHF RFID systems. In this thesis, the problem is tackled using a combination of antenna design, mathematical modelling and algorithm development. A novel UHF RFID reader antenna with 2D beam-scanning ability, a wide axial ratio (AR) beamwidth and tunable polarisation performance is designed, simulated, measured and tested in Chapter 3 of this thesis. It is experimentally verified that the antenna's AR minima follow its gain maxima, generating an ultra-wide 3 dB AR beamwidth of 136°. It is demonstrated that the ratio of inventoried tags within 120 seconds increased from 65% with a static beam configuration to 96% with 1D beam-scanning and further to 100% with 2D beam-scanning. By manually degrading the proposed antenna's AR to 4 dB without affecting other parameters, a maximum 18% degradation in the coverage ratio is observed, verifying the coverage contribution from the proposed antenna's improved AR performance. To further improve the signal coverage of RFID systems, an optimised reader antenna deployment is necessary. A 3D ray-tracing model tailored for RFID applications is proposed, constructed and verified in Chapter 4 of this thesis. Compared with ray-tracing models in the literature, the proposed model considers practical antennas' irregular beam patterns, polarisations, arbitrary orientations and variable AR. The proposed model is verified against commercial MoM simulation software and has shown comparable accuracy but orders of magnitude less time and memory consumption. It is also verified using experiments and shows a good match between the prediction and measurement. In the case study presented, the proposed model demonstrates the ability to consider practical antennas' variable AR, predicting a 13% coverage difference between the AR-optimised antenna and its AR-degraded version, matching the measured value at 10%. Due to the higher power requirement of sensor tags over conventional passive tags, a fast and effective beam-forming algorithm is desirable. A scalable, computationally-light beam-forming algorithm for multi-antenna RFID systems is proposed and experimentally verified in Chapter 5. The proposed algorithm does not rely on iterative phase searching methods, avoiding slow convergence speed and can be implemented with low computational complexity. It is experimentally demonstrated that the proposed algorithm can achieve an RSSI improvement of up to 18 dB using three transmitting antennas. For the signal coverage problem in near-field RFID applications, a simple yet effective near-field antenna design is proposed and verified in Chapter 6 using full-wave simulations. The proposed antenna's signal nulls can be dynamically moved using two sources with suitable phase differences. Moreover, an almost uniform field distribution along the strip-line can be achieved by setting the phase difference of two sources to be 90° or 270°. Furthermore, by adding a new track of strip-line and setting a proper phase difference, a side-power leakage suppression of up to 20 dB can be achieved with the vertical detection range extended from 43 cm to 65 cm at 30 dBm input power. This theory is further verified by changing the phase difference of S1 and S2 to π, in which the side-leakage gets greatly strengthened as predicted by theory.

Ultra High Frequency (UHF) Radio Frequency Identification (RFID) has gained importance over the past two decades in many applications such as stock management, asset tracking and access control. For wide area applications, Distributed Antenna Systems (DAS) have been used to obtain good coverage with few antennas by making use of multiple spatially distributed antennas and phase dithering. This implements a far-field beamforming that maximises the instantaneous power at a tag. Separately, phased array antennas have also been used to increase the read range by increasing the effective field of view of an antenna and overcoming multipath fading through beam steering. This dissertation explores a combination of both approaches to improve RFID read ranges in wide interrogation zones. Distributed antenna arrays are explored in the context of delivering high tag detection probabilities in a multi-cell RFID system, while maximising inter-antenna separations. A Distributed Antenna Array System (DAAS) is designed and shown to be capable of providing comparable performance to a fixed DAS system with fewer antennas. The properties of the system are further studied and its upper performance limit is explored by modelling a hypothetical perfectly steerable antenna array. The concept of using perfectly steerable arrays is further explored to propose a cell-less RFID system, in which cell allocation in wide area RFID is replaced with a tag location-based interrogation requiring the global reader antenna population to be used for interrogation of all tags, leading to significant potential increases in inter-antenna separation, and consequently good coverage with fewer antennas. It is also argued that this system leads to the avoidance of complex reader anti-collision policies, since only a single central reader is now required. Finally, the design of a wide-scan-angle antenna array is presented as a compromise solution for perfectly steerable antennas, whist still keeping the desired property of being flat panel. A 3D RFID multi-antenna model is presented and used for simulating and analysing the various described systems and for system planning.