With the rapid increase of size and complexity of modem electrical power systems, 1) the simulation accuracy and 2) the capability of simulating large power systems have become two conflicting objectives. This thesis proposes a novel FPGA-RTDS co-simulator to meet these two objectives. As the basis of the co-simulator, a library of power system components is developed in FPGA, including the most commonly used power system elements and control systems. The proposed cosimulator combines the advantages of 1) the paralleled architecture, high logic density and high clock speed from FPGA and 2) better modelling flexibility and user-friendly GUI from RTDS together. Multi-FPGA structure is introduced to further improve the simulation capability for large power systems. The use of detailed EMT models in the whole system guarantees the accuracy of simulation and eliminates the potential interface error. Deeply pipelined and massively paralleled algorithms have been designed to maximize time and hardware efficiency. The modular design significantly improves the system expandability. Case studies including large scale power system with more than 4000 nodes are presented to demonstrate the simulation capability. Comparisons are made with SIMULINK and RTDS to verify the accuracy of the proposed co-simulator.

The massive growth of computer usage has led to an increase in the related security concerns. Malware, such as Viruses, Worms, and Trojans, have become a major issue due to the serious damages they cause. Since the first malware emerged, there has been a continuous battle between security researchers and malware writers, where the latter are constantly trying to evade detection by adopting new functionalities and malicious techniques. This thesis focuses on addressing some of the concerns and challenges encountered when detecting malware, based on their behavioural features observed; for each identified challenge, an approach that addresses the problem is proposed and evaluated. Firstly, the thesis provides an in-depth analysis of the underlying causes of malware misclassification when using machine learning-based malware detectors. Such causes need to be determined, so that the right mitigation can be adopted. The analysis shows that the misclassification is mostly due to changes in several malware variants without the family membership or the year of discovery being a factor. In addition, the thesis proposes a probabilistic approach for optimising the scanning performance of Forensic Virtual Machines (FVMs); which are cloud-based lightweight scanners that perform distributed monitoring of the cloud’s Virtual Machines (VMs). Finally, a market-inspired prioritisation approach is proposed to balance the trade-off between the consumption of VMs’ resources and accuracy when detecting malware on the cloud’s VMs using Virtual Machine Introspection-based lightweight monitoring approaches (e.g. FVMs). The thesis concludes by highlighting future work and new directions that have emerged from the work presented.

The recent “resurrection” of interest in Virtual Reality has stimulated interest in the quest for true “immersion” in computer-generated worlds. True immersion may only ever be achieved through advanced BCI systems, but, until that day arrives, it is important to understand how it may be possible to measure human engagement and emotions within virtual worlds using psychophysiological techniques. This study aims to design an affective computing system, capable of responding to human emotions, within virtual environments. Based on the development of a Valence/Arousal model, a controllable affective VR, capable of evoking multiple emotions, has been constructed. Multiple variations of the VR have been evaluated subjectively using over 68 participants. More objective, physiologically-based experiments have been executed, in which the EEG, GSR and heart rates of 45 participants have been recorded during exposure to the most powerful affective environments, identified in the earlier study. Multiple affective recognition systems have been trained and crossvalidated against 30 participants and evaluated using the other 15 individuals. The results suggested that the trained classifiers perform highly accurately in the training database, but achieve random classification accuracies in the new dataset. It was highlighted that the extreme performance attenuation is due to the high individual differences in participants’ physiological responses, in emotional experiences.

The aim of this research is to investigate a novel Multiple-Input-Multiple-Output (MIMO) sensor system for automotive applications. Compared to traditional phased arrays, a MIMO array can achieve the same fine angular resolution, but with a drastically less amount of sensor elements. For example, a MIMO array of 10 elements can deliver the same resolution as a phased array of 25 elements. The other highlight of this technology is that it can operate at short ranges, which is physically impossible with a phased array. Therefore, a MIMO system can potentially provide very high angular resolutions at short ranges. These properties make such a system attractive for a number of automotive applications, including parking aids, short-range cruise control, speed-over-ground estimation, pedestrian detection etc. Research started with the verification of application of conventional MIMO techniques for radar context. MIMO techniques were tested with the existing RF equipment in laboratory environment. Beamforming capabilities were verified, range and angular resolutions were compared to equivalent phased arrays and multiple-target resolving capabilities were confirmed. Nearfield focusing algorithms for MIMO arrays were developed and verified via experiments in the anechoic chamber with the same equipment. A technology demonstrator based on ultrasonic equipment was built and then tested in an anechoic chamber and the findings were compared to computed performance parameters. Further performance optimisations via aperiodic MIMO configurations were explored via use of heuristic optimisation algorithms. An optimised configuration then was tested in anechoic chamber and its performance was confirmed in experimentally. Finally using radio equipment again, an initial study on MTI applications was done. Platform motion compensation methods were developed and tested in order to make up for vehicular motion and to compensate for its possible effects. Both stationary platforms and a moving platform was used to experimentally confirm the MTI capabilities on both in an indoor setup and an outdoor setup.

The lack of references on nonlinearity issue faced in reconfigurable antennas has motivated the work described in this thesis. The nonlinear behaviour is caused by active switches introduced on the radiating structure of the reconfigurable antennas. Depending on the type of active switches deployed on the antenna, the nonlinearity could be severe, which could have serious implications for antenna operation. Thus, the issue of nonlinearity in reconfigurable antennas should not be ignored and nonlinearity measurements should be performed to ensure the nonlinear performance is within an acceptable level. A set of nonlinearity measurements has been identified and performed on the proposed reconfigurable PIFAs. Prototypes are presented with PIN diode and E-PHEMT switches. For the purpose of comparison, measurements were also made with the active switch replaced with a copper bridge for linear interconnection. The nonlinearity performance can be evaluated from the measurement values of third-order intermodulation distortion (IMD3) products, ratio of IMD3 products to carrier, IMD3 products asymmetry, third-order input intercept point (IIP3) and 1-dB gain compression point (P\(_1\)\(_-\)\(_d\)\(_B\)). The measurements are performed when the antenna is transmitting signals. All measurements are performed on the state-of-the-art, 4-port ZVA67 Rohde & Schwarz VNA. Based on the nonlinearity measurements, it can be concluded that the presence of active switches has compromised the nonlinearity of the reconfigurable antennas. This is evident from the appearance of strong IMD3 products at the frequency of interest. In addition, the power-series-based approximation of 10 dB difference between the measured P\(_1\)\(_-\)\(_d\)\(_B\) and IIP3 is shown to be reasonable. Moreover, this work has demonstrated that the ratio of the IMD3 products to carrier does not vary significantly with radiation angles.

In industry, pneumatic actuating systems have been widely employed due to their distinct advantages. However, pneumatic actuators have rather low energy efficiency, around 23-30% in general. The work described in this thesis is to study potential strategies for improvement of pneumatic actuator energy efficiency by recovering the exhaust air energy using a scroll-type air motor. The scroll-type air motor is a relatively new technology, and its unique compact structure determines that it can work at high energy efficiency. The mathematical models for vane-type air motors with arbitrary number of vanes and for scroll-type air motors are derived, which cover their thermodynamic process and responses. A complete mathematical model for the pneumatic-electrical system is developed. The simulation studies are carried for initial investigation of the working process for the electrical-pneumatic system and its dynamic responses. A closed-loop control strategy is proved to be essential. An experimental test rig for this system has been built in the laboratory, and the test results are described. The simulation study and the laboratory test results demonstrate that the proposed system is realistic for energy recovery while a proper control strategy is imposed and the overall system energy efficiency can be improved under reasonable constraint conditions.

The advantages of DC transmission over the traditional AC transmission, particularly under medium and high voltage level, are widely recognised in recent years. In order to utilise DC power transmission in a more efficient manner, a construction of multi-terminal DC grids is put forward in previous literature. However, the development of large-scale DC grids faces various technical challenges, such as the power flow management, the voltage conversion and regulation, and the fault current management. One substantial and promising solution against these technical challenges whilst enhancing the controllability and flexibility of DC grids is by the inclusion and control of DC-DC converters. In this thesis, three potential applications of the DC-DC converter in future DC grids are investigated: 1) acting as a DC current/power flow controller for current/power flow management in meshed DC grids; 2) acting as a DC transformer for the interconnection of DC grids; 3) acting as a DC interface for integrating energy storage systems into DC grids.

This thesis introduces an alternative type of radar system for airborne surveillance. The system comprises of a co-located Multiple-Input-Multiple-Output (MIMO) radar array that can be towed by an airborne platform. This could allow the realisation of a long antenna, giving a narrow beam-width, whilst optimising the trade-off between size/weight of a radar array and its length. As new concept, a system study of it operating onboard a high-altitudepseudo-satellite (HAPS) platform was done. Operating at 400MHz, it needed a 25m array having a weight of only 2.2kg, and had detection ground range of 40km for a target with a RCS of 10\(m\)\(^2\) . A fundamental problem with the concept, is that the towing line will deviate due to turbulence during flight. The MIMO structure of the array could allow compensation of the errors caused by its deviation. The key contribution in this thesis are motion correction methods and proof of concept experimental results which shows that compensation is possible. A simulation tool for the proposed radar concept was created, whose results were confirmed with the experimental results, which was used to simulate the concept onboard the HAPS platform. Finally, an experimental MIMO radar system, with a towable 5m antenna array, was built and verified.

There are several methods to get individualized HRTFs such as measurements, syntheses, or selection. The study first aims to select the fittest HRTFs in an existing database for listeners. The idea is developed based on the connection between anthropometric parameters and auditory localisation. By means of machine learning, the neural network designed for HRTF ranking produces a reliable prediction. The new approach has been verified by the anthropometry and listening perception of 24 subjects. The final selected dataset is used to synthesize a virtual audio scene. Two binaural ambisonic decoders are proposed to overcome the dilemma of improving sound localisation or enhancing audio quality. The first one is the equalization decoding, equalizing the root mean square (RMS) power in each 1/3 octave frequency bank, especially compensating the low-pass filtered components in a high-density speaker array. Therefore, the energy distribution of the treated signal is nearly uniform. The other proposed method is split-band decoding, selecting and then mixing the better reconstructed frequency components from different speaker arrays. Through several experiments and listening tests, there are no click noises in the real-time system, when the virtual auditory space is rapidly rotated. The split-band decoder presents comparable performance to a pure HRTF decoder.

One of the most important modules of the current and next generation of the wireless communications is the antenna. The coexistence of the machine and human in the next communication system will open-up a vast range of new applications and communication services which need to be supported by the antenna. Moreover, the forthcoming 5G technology vision is prognosticated on the use of multiple communication bands and standards in a seamless fashion. This can force the mobile devices to have multiple antennas on a single device which will add significant complexity or using an antenna with wideband reconfiguration capability. On the other hand, switched-off analogue communications provides the opportunity for re-using prime spectrum in UHF bands. Considering the size of hand-held devices, this will need strong miniaturization. To address these requirements, electrically small, tunable, wideband and highly efficient antenna technology is strongly desired. In this thesis a new area of research in antenna design is introduced which has been unexplored by the other researchers. A new theory called Injection Matching Theory (IMT) is proposed which uses multi-port configuration. This will enable to control current distribution on the antenna structure at its extremities and couple a wave length, much larger than what the antenna dimensions naturally allow. Apart from electrically small operation this can be used for improving band width and efficiency, and providing reconfiguration capability. To illustrate the versatility of the proposed theory, for every feature mentioned above a chapter is provided which demonstrates the potential capability of the proposed theory via simulation and fabrication of the prototyped examples.