A vibration based micro-generator is an energy harvesting device that couples a certain transduction mechanism to the ambient vibration and converts mechanical energy to electrical energy. In order to maximize available power, micro-generators are typically inertial devices that operate at a single resonant frequency. The maximum output power is generated when the resonant frequency of the generator matches the ambient vibration frequency. The output power drops significantly if these two frequencies do not match due to the high Q-factor of the generator. This thesis addresses possible methods to overcome this limit of vibration based micro-generators, in particular, method of tuning the resonant frequency of the generator to match the ambient vibration frequency. This thesis highlights mechanical and electrical methods of resonant frequency tuning of a vibration based micro-generator. The mechanical frequency tuning is realized by applying an axial tensile force to strain the cantilever structure of the generator. A tunable micro-generator with a tuning range from 67.6 Hz to 98Hz and a maximum output power of 156.6μW at a constant low vibration acceleration level of 0.59m·s-2 was designed and tested. The tuning mechanism was found not to affect the damping of the generator. A closed loop frequency tuning system as well as the frequency searching algorithms has been developed to realize automatic frequency tuning using the proposed mechanical tuning method. The model of duty cycle of the system was established and it was proved theoretically that a reasonable duty cycle can be achieved if the generator and tuning system is designed properly. The electrical tuning method is realized by changing the load capacitance of the generator. Models of piezoelectric and electromagnetic generators using electrical tuning methods were derived. The model of the electromagnetic generator has also been experimentally verified. The electrically tunable generator tested has a maximum 3dB bandwidth of 4.2Hz. In conclusion, resonant frequency tuning using mechanical methods presented in the thesis have larger tuning range than that using electrical methods. However, frequency tuning using electrical tuning methods consumes less power than that using mechanical methods for the same amount of tuning range.

Increasingly, complex real-world problems (including distributed sensing, air-traffic control, disaster response, network routing, space exploration and unmanned aerial vehicles) are being tackled by teams of software agents, rather than the more traditional centralised systems. Whilst this approach has many benefits in terms of creating robust solutions, it creates a new challenge — how to flexibly coordinate the actions of the agent teams to solve the problem efficiently. In more detail, coordination is here viewed as the problem of managing the interactions of these autonomous entities so that they do not disrupt each other, can take proactive actions to help each other, and take multiple actions at the same time when this is required to solve the problem. In this context, communication underpins most solutions to the coordination problem. That is, if the agents communicate their state and intentions to each other then they can coordinate their actions. Unfortunately, however, in many real-world problems, communication is a scarce resource. Specifically, communication has limited bandwidth, is not always available and may be expensive to utilise. In such circumstances, typical coordination mechanisms break down because the agents can no longer accurately model the state of the other agents. Given this, in this thesis, we consider how to coordinate when communication is a restricted resource. Specifically, we argue for a rational approach to communication. Since communication has a cost then, similarly, we should be able to calculate a value of sending any given communication. Once we have these costs and values, we can use standard decision theoretic models to choose whether to send a communication, and in fact, generate a plan which utilises communications and other actions efficiently. In this research we explore ways to value communications in several contexts. Within the framework of decentralised Partially Observable Markov Decision Process (POMDP) we develop a simple information theoretic valuation function (based on Kullback–Leibler (KL) Divergence). This techniques allows agents to coordinate in large problems such as RoboCupRescue, where teams of ambulances must save as many civilians as possible after an earthquake. We found that, in this task, valuing communications before deciding whether to send them results in a level of performance which is higher than not communicating, and close to a model which utilises a free communication medium to communicate all the time. Furthermore, this model is robust to increasing communication restrictions, whereas simple communication policies are not. We then extend this framework to value communications based on a technique from the field of Machine Learning, namely Reward Shaping, which allows the decentralised POMDP to be transformed into individual agent POMDPs that can be solved more easily. This approach can use a heuristic transformation to allow the approach to work in large problems like RobocupRescue or the Multi–Agent Tiger problem, where it outperforms the current state of the art. Further to this, this approach can also use an exact reward shaping function in order to generate a bounded approximation of the intractable optimal decentralised solution in slightly smaller problems. Finally, we show how, if we restrict our attention to relatively static (i.e. the problem does not change without an agent doing something) problems than those which the reward shaping technique was designed for, we can generate an optimal solution to decentralised control based on communication valuations. In more detail, we extend the class of Bayesian coordination games to include explicit observation and communication actions. By so doing, the value of observation and exchange can be derived using the concept of opportunity costs. This is a natural way of measuring the relationship between communication and information gathering on an agent’s utility, and removes the need to introduce arbitrary penalties for communicating (which is what most existing approaches do). Furthermore, this approach allows us to show that the optimal communication policy is a Nash equilibrium, and to exploit the fact that there exist many efficient algorithms for finding such equilibria in a local fashion. Specifically, we provide a complete analysis of this model for two–state problems, and illustrate how the analysis can be carried out for larger domains making use of explicit information gathering strategies. Finally, we develop a procedure for finding the optimal communication and search policy as a function of the partial observability of the state and payoffs of the underlying game (which we demonstrate in the canonical Multi–Agent Tiger problem). In performing all of this work, we demonstrate how communication can be managed locally by accurately placing a value on the cost and benefit of using a restricted communication resource. This allows agents to coordinate efficiently in many interesting problem domains, where existing approaches perform badly. We contribute to the field of rational communication by providing several algorithms for utilising costly communication under different domain conditions. Our reward shaping approaches are highly scalable in problems with large state spaces and come with sound theoretical guarantees on the optimality of the solution they find.

Research outputs are growing in number and frequency, assisted by a greater number of publication mediums and platforms via which material can be disseminated. At the same time, the requirement to find acceptable, timely, objective measurements of research "quality" has become more important. Historically, citations have been used as an independent indication of the significance of scholarly material. However, citations are very slow to accrue since they can only be made by subsequently published material. This enforces a delay of a number of years before the citation impact of a publication can be accurately judged. By contrast, each new citation establishes a large number of co-citation relationships between that publication and older material whose citation impact is already well established. By taking advantage of this co-citation property, this thesis investigates the possibility of developing a metric that can provide an earlier indicator of a publication's citation impact. This thesis proposes a new family of cocitation based impact measures, describes a system to evaluate their effectiveness against a large citation database, and justifies the results of this evaluation against an analysis of a diverse range of research metrics

One of the dominant properties of a global computing network is the incomplete information available to principals about each other. This was the motivation of using the notion of probabilistic trust as an approach to security sensitive decision making in modern open and global computing systems. In such systems any principal A uses the outcomes of past interactions with another principal B to construct a probabilistic model approximating the behaviour of B. Using this model, the principal A can take decisions regarding interactions with B by estimating its future actions. Many existing frameworks adopt the so-called ‘Beta model’. The main limitation of these frameworks is that they assume the behaviour of any principal to be fixed, which is not realistic in many cases. In this thesis, we first address the application of probabilistic trust to optimise security protocols, and specifically give an example where the Crowds anonymity protocol is extended to use trust information. We then address the problem of evaluating probabilistic trust in principals exhibiting dynamic behaviours. In this respect, we formally analyse the ‘exponential decay’ technique as an approach to coping with principals’ dynamic behaviours. Given the identified limitations of this technique, a more general framework for trust and reputation is introduced. In this framework, Hidden Markov Models (HMMs) are used for modelling the dynamic behaviours of principals. This framework is formally analysed in terms of a notion of ‘estimation error’. Using an experimental approach based on Monte-Carlo methods to evaluate the expected estimation error, the introduced HMM-based framework for trust and reputation is compared to the existing Beta framework. The results show in general that the latter is getting more promising in evaluating trust in principals (‘trustees’) having dynamic behaviours as longer sequences of observations are available about such trustees.

Combinations of analytical and experimental results indicate that deep level doping of Czochralski grown silicon wafers is capable of providing very high resistivity wafers suitable for silicon-on-insulator (SOI), integrated passive devices (IPD) and 3D integration configurations. Deep level doping involves adding trace elements to silicon that compensate for background free carriers introduced by impurities in the silicon and pin the Fermi level near the mid bandgap intrinsic level. Starting from n-type Czochralski-silicon wafers with a nominal resistivity of 50 Ωcm, gold ion implantation and subsequent annealing were used to increase the resistivity of silicon wafers by up to 3 orders of magnitude, to values as high as 93 kΩcm. Hall measurements performed over a large temperature range show that the increase in resistivity is solely due to a decrease in carrier concentration and not a decrease in mobility. The carrier concentration is only one order of magnitude larger than that of intrinsic silicon over a temperature range of 200-360 K. Hall results also show that the resistivity of the compensated material remains up to two orders of magnitude larger than that of the uncompensated material at near operating temperatures. High frequency attenuation measurements in the 1-67 GHz range for coplanar waveguides show attenuation reductions of up to 76% from 0.76 dB/mm to 0.18 dB/mm at 10 GHz for those fabricated on uncompensated and compensated silicon respectively. Spiral inductors fabricated on both compensated and uncompensated silicon show up to a factor of 10 increase in the maximum quality factor from 0.3 to 3.1 for inductors on uncompensated and compensated silicon respectively. A 70% increase in maximum quality factor from 9 to 15.2 is exhibited by inductors commercially fabricated on compensated silicon when compared to those on float-zone silicon. The coplanar waveguide and spiral inductor results provide clear evidence that deep level dopant compensation is effective in improving the performance of passive devices in the GHz frequency range.

In just over a decade the optical frequency comb technique has completely transformed the field of frequency metrology. These devices have made the measurement of the frequency of light a much easier and affordable task when compared to with earlier techniques. With both research and technology development on these devices becoming more mature, optical frequency combs have been affecting other science areas. Applications are already found in spectroscopy, attosecond physics and astrophysics and more science and engineering areas can be expected to be affected in the near future. The dissemination over optical fibre of optical frequency combs between research labs, or between research labs and industry, could facilitate and accelerate this process. In particular, with optical frequency standards currently exhibiting a fractional accuracy better than 10^-17 and optical frequency combs making this accuracy available across a wide spectrum, new experiments could be devised in a wide range of research fields if ultra accurate microwave and optical frequencies were to be made available beyond the walls of metrology laboratories. However, before the work presented in this thesis, limited research was undertaken to test how accurately an optical frequency comb could be transferred over optical fibre. Environmentally-induced noise in the fibre, dispersion issues and other processes taking place during the propagation and detection of the optical signal could all degrade its quality to a level incompatible with the desired applications. The experiments described here demonstrate that optical frequency combs can be disseminated over optical links, from several-km to many tens of km-long, whilst preserving the stability and accuracy of its mode frequency spacing and mode frequency to a level compatible with the majority of the most demanding frequency metrology applications.

We formalise the notion of enriched coalgebraic modal logic, and determine conditions on the category V (over which we enrich), that allow an enriched logical connection to be extended to a framework for enriched coalgebraic modal logic. Our framework uses V-functors L: A → A and T: X → X, where L determines the modalities of the resulting modal logics, and T determines the coalgebras that provide the semantics. We introduce the V-category Mod(A, α) of models for an L-algebra (A, α), and show that the forgetful V-functor from Mod(A, α) to X creates conical colimits. The concepts of bisimulation, simulation, and behavioural metrics (behavioural approximations),are generalised to a notion of behavioural questions that can be asked of pairs of states in a model. These behavioural questions are shown to arise through choosing the category V to be constructed through enrichment over a commutative unital quantale (Q, Ⓧ, I) in the style of Lawvere (1973). Corresponding generalisations of logical equivalence and expressivity are also introduced,and expressivity of an L-algebra (A, α) is shown to have an abstract category theoretic characterisation in terms of the existence of a so-called behavioural skeleton in the category Mod(A, α). In the resulting framework every model carries the means to compare the behaviour of its states, and we argue that this implies a class of systems is not fully defined until it is specified how states are to be compared or related.

In this thesis, we propose a suite of buffer-aided transmission schemes designed for a multihop link or for a three-node network by exploiting the characteristics of buffer-aided transmissions. Our objective is to improve the end-to-end BER, outage probability, throughput and energy dissipation. Specifically, we firstly proposed and studied a buffer-aided multihop link (MHL), where all the relay nodes (RNs) are assumed to have buffers for temporarily storing their received packets. Hence, the RNs are operated under the so-called store-and-forward (SF) relaying scheme. As a benefit of storing packets at the RNs, during each time-slot (TS), the best hop having the highest signal-to-noise ratio (SNR) can be activated from the set of those hops that have packets awaiting transmission in the buffer. A packet is then transmitted over the best hop. This hop-selection procedure is reminiscent of selection (SC) diversity, which is referred to here as multi-hop diversity (MHD), when assuming that each hop experiences both propagation pathloss and independent identically distributed (i.i.d) flat Rayleigh fading. In order to make the channel activation practical, a MAC layer implementation is proposed and several closed-form formulas are derived for its characterization. Then we studied the buffer-aided multihop link, when assuming that each hop experiences both propagation pathloss and independent non-identically distributed (i.n.i.d) at Nakagami-m fading. Both BPSK as well as M-ary quadrature amplitude modulation (MQAM) are employed. During each TS, the MHD scheme activates the specific hop's transmission, whose signal-to-noise ratio (SNR) cumulative distribution function (CDF) gives the highest ordinate value amongst all the available hops. The next packet is then transmitted over the selected hop. This CDF-aware MHD scheme is suitable for operation in the scenarios, where the different hops may have different length, hence resulting in different average SNRs, and/or experience different types of fading. This MHD scheme is also capable of achieving the maximum attainable diversity gain provided by the independent fading experienced by the different hops. Then the benefits of adaptive modulation are exploited, where the number of bits transmitted in each TS is affected both by the channel quality and the buffer fullness. During each TS, the criterion used for activating a specific hop is that of transmitting the highest number of bits (packets). When more than one hops are capable of transmitting the same number of bits, the particular hop having the highest channel quality (reliability) is activated. Hence we refer to this regime as the Maximum Throughput Adaptive Rate Transmission (MTART) scheme. Additionally, a new MAC layer protocol is proposed for implementing our MTART management. Finally, we propose and study a routing scheme, namely the Buffer-aided Opportunistic Routing (BOR) scheme, which combines the benefits both opportunistic routing and MHD transmission. It is conceived for the transmission of information in a Buffer-aided Three-node Network (B3NN) composed of a Source Node (SN), a buffer-aided Relay Node (RN) and a Destination Node (DN). When applying opportunistic routing, each packet is transmitted from SN to DN either directly or indirectly via a RN based on the instantaneous channel quality. When applying MHD transmission, the RN is capable of temporarily storing the received packets, which facilitates transmission over three links, namely the SR, RD and SD links. In this network, the three channels define a 3D channel probability space (CPS), which is divided into four regions representing the activation-region of the three channels and an outage region. Then the instantaneous channel quality values map to a specific point in this 3D channel space. The BOR scheme relies on the position of this point to select the most appropriate channel in the 3D CPS for its transmission. In comparison to the benchmark schemes considered in the literature, the BER, the OP, throughput and/or energy dissipation of our proposed systems are significantly improved.

In this thesis we study systems with switching dynamics and we propose new mathematical tools to analyse them. We show that the postulation of a global state space structure in current frameworks is restrictive and lead to potential difficulties that limit its use for the analysis of new emerging applications. In order to overcome such shortcomings, we reformulate the foundations in the study of switched systems by developing a trajectory-based approach, where we allow the use of models that are most suitable for the analysis of a each system. These models can involve sets of higher-order differential equations whose state space does not necessarily coincide. Based on this new approach, we first study closed switched systems, and we provide sufficient conditions for stability based on LMIs using the concept of multiple higher order Lyapunov function. We also study the role of positive-realness in stability of bimodal systems and we introduce the concept of positive-real completion. Furthermore, we study open switched systems by developing a dissipativity theory. We give necessary and sufficient conditions for dissipativity in terms of LMIs constructed from the coefficient matrices of the differential equations describing the modes. The relationship between dissipativity and stability is also discussed. Finally, we study the dynamics of energy distribution networks. We develop parsimonious models that deal effectively with the variant complexity of the network and the inherent switching phenomena induced by power converters. We also present the solution to instability problems caused by devices with negative impedance characteristics such as constant power loads, using tools developed in our framework.

Cloud computing is a paradigm that promises to move IT another step towards the age of computing utility. Traditionally, Clouds employ dedicated resources located in data centres to provide services to clients. The resources in such Cloud systems are known to be highly reliable with a low probability of failure. Desktop Cloud computing is a new type of Cloud computing that aims to provide Cloud services at little or no cost. This ambition can be achieved by combining Cloud computing and Volunteer computing into Desktop Clouds, harnessing non-dedicated resources when idle. The resources can be any type of computing machine, for example a standard PC, but such computing resources are renowned for their volatility; failures can happen at any time without warning. In Cloud computing, tasks are submitted by Cloud users or brokers to be processed and executed by virtual machines (VMs), and virtual mechanisms are hosted by physical machines (PMs). In this context, throughput is defined as the proportion of the total number of tasks that are successfully processed, so the failure of a PM can have a negative impact on this measure of a Desktop Cloud system by causing the destruction of all hosted VMs, leading to the loss of submitted tasks currently being processed. The aim of this research is to design a VM allocation mechanism for Desktop Cloud systems that is tolerant to node failure. VM allocation mechanisms are responsible for allocating VMs to PMs and migrating them during runtime with the objective of optimisation, yet those available pay little attention to node failure events. The contribution of this research is to propose a Fault-Tolerant VM allocation mechanism that handles failure events in PMs in Desktop Clouds to ensure that the throughput of Desktop Cloud system remains within acceptable levels by employing a replication technique. Since doing so causes an increase of power consumption in PMs, the mechanism is enhanced with a migration policy to minimise this effect, evaluated using three metrics: throughput of tasks; power consumption of PMs; and service availability. The evaluation is conducted using DesktopCloudSim, a tool developed for the purpose by this study as an extension to CloudSim, the well-known Cloud simulation tool, to simulate node failure events in Cloud systems, analysing node failure with real data sets of collected from Failure Trace Archives. The experiments demonstrate that the FT mechanism improves the throughput of Cloud systems statistically significantly compared with traditional mechanisms (First Come First Serve, Greedy and RoundRobin) in the presence of node failures. The FT mechanism reduces power consumption statistically significantly when its migration policy is employed.