RF signal sensing and source localisation systems using Software Defined Radios

Abstract

Radio frequency (RF) source localisation is a critical technology in numerous location-based military and civilian applications. In this thesis, the problem of RF source localisation has been studied from the perspective of the system implementation for real-world applications. Commercial off-the-shelf Software Defined Radio (SDR) devices are used to demonstrate the practical RF source localisation systems. Compared to the conventional localisation systems, which rely on dedicated hardware, the SDR-based system is developed using general-purpose hardware and software-defined components, offering great flexibility and cost efficiency in system design and implementation. In this thesis, the theoretical results of source localisation are evaluated and put into practice. To be specific, the practical localisation systems using different measurement techniques, including received-signal-strength-indication (RSSI) measurements, time-difference-of-arrival (TDOA) measurements and joint TDOA and frequency-difference-of-arrival (FDOA) measurements, are demonstrated to localise the stationary RF signal sources using the SDRs. The RSSI-based localisation system is demonstrated in small indoor and outdoor areas with a range of several metres using the SDR-based transceivers. Furthermore, interests from the defence area motivated us to implement the time-based localisation systems. The TDOA-based source localisation system is implemented using multiple spatially distributed SDRs in a large outdoor area with the sensor-target range of several kilometres. Moreover, they are implemented in a fully passive way without prior knowledge of the signal emitter, so the solutions can be applied in the localisation of non-cooperative signal sources provided that emitters are distant. To further reduce the system cost, and more importantly, to deal with the situation when the deployment of multiple SDRs, due to geographical restrictions, is not feasible, a joint TDOA and FDOA-based localisation system is also demonstrated using only one stationary SDR and one mobile SDR. To improve the localisation accuracy, the methods that can reduce measurement error and obtain accurate location estimates are studied. Firstly, to obtain a better understanding of the measurement error, the error sources that affect the measurement accuracy are systematically analysed from three aspects: the hardware precision, the accuracy of signal processing methods, and the environmental impact. Furthermore, the approaches to reduce the measurement error are proposed and verified in the experiments. Secondly, during the process of the location estimation, the theoretical results on the pre-existing localisation algorithms which can achieve a good trade-off between the accuracy of location estimation and the computational cost are evaluated, including the weight least-squares (WLS)-based solution and the Extended Kalman Filter (EKF)-based solution. In order to use the pre-existing algorithms in the practical source localisation, the proper adjustments are implemented. Overall, the SDR-based platforms are able to achieve low-cost and universal localisation solutions in the real-world environment. The RSSI-based localisation system shows tens of centimetres of accuracy in a range of several metres, which provides a useful tool for the verification of the range-based localisation algorithms. The localisation accuracy of the TDOA-based localisation system and the joint TDOA and FDOA-based localisation system is several tens of metres in a range of several kilometres, which offers potential in the low-cost localisation solutions in the defence area.