Extraction, representation, organisation and application of metadata about audio recordings are in the concern of semantic audio analysis. Our broad interpretation, aligned with recent developments in the field, includes methodological aspects of semantic audio, such as those related to information management, knowledge representation and applications of the extracted information. In particular, we look at how Semantic Web technologies may be used to enhance information management practices in two audio related areas: music informatics and music production. In the first area, we are concerned with music information retrieval (MIR) and related research. We examine how structured data may be used to support reproducibility and provenance of extracted information, and aim to support multi-modality and context adaptation in the analysis. In creative music production, our goals can be summarised as follows: O↵-the-shelf sound editors do not hold appropriately structured information about the edited material, thus human-computer interaction is inefficient. We believe that recent developments in sound analysis and music understanding are capable of bringing about significant improvements in the music production workflow. Providing visual cues related to music structure can serve as an example of intelligent, context-dependent functionality. The central contributions of this work are a Semantic Web ontology for describing recording studios, including a model of technological artefacts used in music production, methodologies for collecting data about music production workflows and describing the work of audio engineers which facilitates capturing their contribution to music production, and finally a framework for creating Web-based applications for automated audio analysis. This has applications demonstrating how Semantic Web technologies and ontologies can facilitate interoperability between music research tools, and the creation of semantic audio software, for instance, for music recommendation, temperament estimation or multi-modal music tutoring.

Cooperation is seen as a means to improve the signal in OFDMA wireless networks by overcoming the inter-cell interference. Such co-operation can be deployed in both the physical layer and the MAC layer. In this thesis, a cross-layer cooperation strategy is considered. Firstly, in the physical layer, a cooperative coding scheme with private information sharing is proposed based on dirty paper coding; this is analyzed in a scenario with two transmitters and two receivers. To implement the cooperation, a rate limited link is deployed at the transmitters’ side in order to share the information. A new achievable rate region is established in both strong interference regime and weak interference regime. Secondly, in the MAC layer, a graph-based dynamic coordinated clustering scheme is proposed. An interference weighted graph is constructed to assist dynamic coordinated clustering for inter-cell interference mitigation and to improve the cell-edge user performance. Only 2 bits are allowed for the signalling exchange between transmitters and this reduces the overhead of the approach. The system throughput and cell-edge throughput with different user distributions are used to evaluate the performance. Thirdly, a transmit antenna selection algorithm is presented to optimize system performance with the constraint of fairness. A graph is generated by using the channel condition between the transmit antennas and Mobile Stations. Based on the graph, a heuristic algorithm is proposed to choose the transmit antenna for each user in order to improve the system performance and guarantee the user fairness. Finally, combining the cooperative coding scheme and cooperative clustering scheme, a cross-layer cooperation scheme is presented. In the physical layer, the cooperation coding scheme mitigates the interference and increases the transmission rate; in the MAC layer, the cooperative clustering scheme provides efficient cooperative transmission. Simulation results show that the proposed scheme can effectively increase both the system throughput and cell-edge throughput.

The publish/subscribe paradigm is used to support a many-to-many model that allows an efficient dissemination of messages across a distributed system. Message Oriented Middleware (MOM) is a middleware that provides an asynchronous method of passing information between networked applications. MOMs can be based on a publish/subscribe model, which offers a robust paradigm for message delivery. This research is concerned with this specific type of MOM. Recently, systems using MOMs have been used to integrate enterprise systems over geographically distributed areas, like the ones used in financial services, telecommunication applications, transportation and health-care systems. However, the reliability of a MOM system must be verified and consideration given to reachability to all intended destinations typically with to guarantees of delivery. The research in this thesis provides an automated means of checking the (re)configuration of a publish/subscribe MOM system by building a model and using Linear-time Temporal Logic and Computation Tree Logic rules to verify certain constraints. The verification includes the checking of the reachability of different topics, the rules for regulating the working of the system, and checking the configuration and reconfiguration after a failure. The novelty of this work is the creation and the optimization of a symbolic model checker that abstracts the end-to-end network configuration and reconfiguration behaviour and using it to verify reachability and loop detection. In addition a GUI interface, a code generator and a sub-paths detector are implemented to make the system checking more user-friendly and efficient. The research then explores another aspect of reliability. The requirements of mission critical service delivery over a MOM infrastructure is considered and we propose a new way of supporting rapid recovery from failures using pre-calculated routing Abstract tables and coloured flows that can operate across multiple Autonomous System domains. The approach is critically appraised in relation to other published schemes.

Two fundamental principles dominate the signal processing of the auditory system: selectivity and adaptation. The response of the auditory system is selective for various acoustic features and the representation of these acoustic features adapts over time. This thesis is concerned with the characterisation of selectivity and adaptation in the human auditory system. Initially, selectivity for modulation rate and adaptation to intensity are characterised in a central auditory model. Next, selectivity for temporalmodulation rate and selective adaptation to both intensity and temporal modulation rate are characterised in psychophysical data.

Powerline Communication (PLC) has long been established for low data rate applications by the electric supply companies. Since 1991, the European CENELEC standard EN 50065 has ruled the use of 3 - 148.5KHz frequency range for narrow band PLC applications. Sim- ilar standard has been established by the IEEE in the US, where a frequency range of 50 - 450KHz is available. The fast growth of Internet since the 1990s accelerated the demands for digital communication services. Furthermore, with the develop- ment of in-home networking, there is a need to establish high speed data links between multiple household devices. This makes PLC sys- tems march rapidly into the high frequency range above 1MHz. Exist- ing broadband PLC system in the 1.6 - 30MHz frequency range only provides data rates smaller than 200Mbps. With the growing demand of multimedia services such as High De nition (HD) video streaming, much faster transmission speed up to Gigabits per second is required and this can be achieved by increasing the operating frequencies. Ultra Wideband (UWB) transmission in free space provides extremely broad bandwidth for short-range, high data rate applications. If UWB signals could be transmitted over the powerline channels in the high frequency range above 30MHz, data rates up to gigabits per second could be achieved. In this thesis, the possibility of implementing ultra wideband trans- mission over the low voltage indoor powerline is investigated. The starting point is to understand the signal propagation characteristics over powerline cables, in the UWB frequency range. Experimental re- sults indicate that the signal degrades at an acceptable rate over the mains cable in a scaled down UWB frequency band (50MHz - 1GHz), which provides a potential operation band for UWB over PLC ap- plications. Key component for the PLC system, a broadband Radio Frequency (RF) coupler is designed and developed, to introduce UWB signals to the transmission channel. With the channel properties and coupling unit, extensive experimental investigations are carried out to analyse the powerline network environment, including channel loss, noise and radiated emission. Furthermore, theoretical channel capac- ity and link budget are derived from measured parameters. It is shown that the indoor powerline is a suitable media for data transmission in the high frequency range from 50 to 550MHz in the home environment. Finally, system level performance is analysed by modelling the Phys- ical Layer (PHY) data transmission. The Multiband-OFDM UWB proposal for IEEE 802.15.3a standard is used to predict the transmis- sion performance under di erent propagation paths and data rates. The research work conducted in this project has proven that UWB over PLC is highly feasible for future in-home applications. With the global promotion of smart grid applications, UWB over PLC will play an important role in providing high speed data transmission over the power networks.

The increasing popularity of wireless communications and the higher data requirements of new types of service lead to higher demands on wireless networks. Relay based cellular networks have been seen as an effective way to meet users’ increased data rate requirements while still retaining the benefits of a cellular structure. However, maximizing the probability of providing service and spectrum efficiency are still major challenges for network operators and engineers because of the heterogeneous traffic demands, hard-to-predict user movements and complex traffic models. In a mobile network, load balancing is recognised as an efficient way to increase the utilization of limited frequency spectrum at reasonable costs. Cooperative control based on geographic load balancing is employed to provide flexibility for relay based cellular networks and to respond to changes in the environment. According to the potential capability of existing antenna systems, adaptive radio frequency domain control in the physical layer is explored to provide coverage at the right place at the right time. This thesis proposes several effective and efficient approaches to improve spectrum efficiency using network wide optimization to coordinate the coverage offered by different network components according to the antenna models and relay station capability. The approaches include tilting of antenna sectors, changing the power of omni-directional antennas, and changing the assignment of relay stations to different base stations. Experiments show that the proposed approaches offer significant improvements and robustness in heterogeneous traffic scenarios and when the propagation environment changes. The issue of predicting the consequence of cooperative decisions regarding antenna configurations when applied in a realistic environment is described, and a coverage prediction model is proposed. The consequences of applying changes to the antenna configuration on handovers are analysed in detail. The performance evaluations are based on a system level simulator in the context of Mobile WiMAX technology, but the concepts apply more generally.

This thesis concerns mathematical models of wireless communication networks, in particular ad-hoc networks and 802:11 WLANs. In ad-hoc mode each of these devices may function as a sender, a relay or a receiver. Each device may only communicate with other devices within its transmission range. We use graph models for the relationship between any two devices: a node stands for a device, and an edge for a communication link, or sometimes an interference relationship. The number of edges incident on a node is the degree of this node. When considering geometric graphs, the coordinates of a node give the geographical position of a node. One of the important properties of a communication graph is its connectedness | whether all nodes can reach all other nodes. We use the term connectivity, the probability of graphs being connected given the number of nodes and the transmission range to measure the connectedness of a wireless network. Connectedness is an important prerequisite for all communication networks which communication between nodes. This is especially true for wireless ad-hoc networks, where communication relies on the contact among nodes and their neighbours. Another important property of an interference graph is its chromatic number | the minimum number of colours needed so that no adjacent nodes are assigned the same colour. Here adjacent nodes share an edge; adjacent edges share at least one node; and colours are used to identify di erent frequencies. This gives the minimum number of frequencies a network needs in order to attain zero interference. This problem can be solved as an optimization problem deterministically, but is algorithmically NP-hard. Hence, nding good asymptotic approximations for this value becomes important. Random geometric graphs describe an ensemble of graphs which share common features. In this thesis, node positions follow a Poisson point process or a binomial point process. We use probability theory to study the connectedness of random graphs and random geometric graphs, which is the fraction of connected graphs among many graph samples. This probability is closely related to the property of minimum node degree being at least unity. The chromatic number is closely related to the maximum degree as n ! 1; the chromatic number converges to maximum degree when graph is sparse. We test existing theorems and improve the existing ones when possible. These motivated me to study the degree of random (geometric) graph models. We study using deterministic methods some degree-related problems for Erda}os-R enyi random graphs G(n; p) and random geometric graphs G(n; r). I provide both theoretical analysis and accurate simulation results. The results lead to a study of dependence or non-dependence in the joint distribution of the degrees of neighbouring nodes. We study the probability of no node being isolated in G(n; p), that is, minimum node degree being at least unity. By making the assumption of non-dependence of node degree, we derive two asymptotics for this probability. The probability of no node being isolated is an approximation to the probability of the graph being connected. By making an analogy to G(n; p), we study this problem for G(n; r), which is a more realistic model for wireless networks. Experiment shows that this asymptotic result also works well for small graphs. We wish to nd the relationship between these basic features the above two important problems of wireless networks: the probability of a network being connected and the minimum number of channels a network needs in order to minimize interference. Inspired by the problem of maximum degree in random graphs, we study the problem of the maximum of a set of Poisson random variables and binomial random variables, which leads to two accurate formulae for the mode of the maximum for general random geometric graphs and for sparse random graphs. To our knowledge, these are the best results for sparse random geometric graphs in the literature so far. By approximating the node degrees as independent Poisson or binomial variables, we apply the result to the problem of maximum degree in general and sparse G(n; r), and derived much more accurate results than in the existing literature. Combining the limit theorem from Penrose and our work, we provide good approximations for the mode of the clique number and chromatic number in sparse G(n; r). Again these results are much more accurate than existing ones. This has implications for the interference minimization of WLANs. Finally, we apply our asymptotic result based on Poisson distribution for the chromatic number of random geometric graph to the interference minimization problem in IEEE 802:11b/g WLAN. Experiments based on the real planned position of the APs in WLANs show that our asymptotic results estimate the minimum number of channels needed accurately. This also means that sparse random geometric graphs are good models for interference minimization problem of WLANs. We discuss the interference minimization problem in single radio and multi-radio wireless networking scenarios. We study branchand- bound algorithms for these scenarios by selecting di erent constraint functions and objective functions.

Nowadays, computational grids have evolved to a stage where they can comprise many volunteered resources owned by different individual users and/or institutions, such as desktop grids and volunteered computing grids. This brings benefits for large-scale computing, as more resources are available to exploit. On the other hand, the inherent characteristics of the volunteered resources bring some challenges for efficiently exploiting them. For example, jobs may not be able to be executed by some resources, as the computing resources can be heterogeneous. Furthermore, the resources can be volatile as the resource owners usually have the right to decide when and how to donate the idle Central Processing Unit (CPU) cycles of their computers. Therefore, in order to utilise volunteered resources efficiently, this research investigated solutions from different aspects. Firstly, this research proposes a new computational Grid architecture based on Java and Java application migration technologies to provide fundamental support for coping with these challenges. This proposed architecture supports heterogeneous resources, ensuring local activities are not affected by Grid jobs and enabling resources to carry out live and automatic Java application migration. Secondly, this research work proposes some job-scheduling and migration algorithms based on resource availability prediction and/or artificial intelligence techniques. To examine the proposed algorithms, this work includes a series of experiments in both synthetic and practical scenarios and compares the performance of the proposed algorithms with existing ones across a variety of scenarios. According to the critical assessment, each algorithm has its own distinct advantages and performs well when certain conditions are met. In addition, this research analyses the characteristics of resources in terms of the availability pattern of practical volunteer-based grids. The analysis shows that each environment has its own characteristics and each volunteered resource’s availability tends to possess weak correlations across different days and times-of-day.

Structured audio uses a high level representation of a signal to produce audio output. When it was rst introduced in 1998, creating a structured audio representation from an audio signal was beyond the state-of-the-art. Inspired by object coding and structured audio, we present a system to reproduce audio using Expressive MIDI, high-level parameters being used to represent pitch expression from an audio signal. This allows a low bit-rate MIDI sketch of the original audio to be produced. We examine optimisation techniques which may be suitable for inferring Expressive MIDI parameters from estimated pitch trajectories, considering the e ect of data codings on the di culty of optimisation. We look at some less common Gray codes and examine their e ect on algorithm performance on standard test problems. We build an expressive MIDI system, estimating parameters from audio and synthesising output from those parameters. When the parameter estimation succeeds, we nd that the system produces note pitch trajectories which match source audio to within 10 pitch cents. We consider the quality of the system in terms of both parameter estimation and the nal output, nding that improvements to core components { audio segmentation and pitch estimation, both active research elds { would produce a better system. We examine the current state-of-the-art in pitch estimation, and nd that some estimators produce high precision estimates but are prone to harmonic errors, whilst other estimators produce fewer harmonic errors but are less precise. Inspired by this, we produce a novel pitch estimator combining the output of existing estimators.

Mobile ad hoc networks are of growing interest because of their unique characteristics and advantages in many practical applications. QoS provision acts as a major challenge in the routing protocol design in the real-world mobile ad hoc networks, especially for the real-time services. OFDM is a new technology which has many advantages over the other modulation schemes. Because of its prominent features, many popular wireless standards have adopted it as physical layer modulation, such as IEEE 802.11 series, WiMAX, 3GPP LTE etc, and it is extended to multiuser environment known as OFDMA. So far none of the existing ad hoc routing protocols fully account for the OFDMA based mobile ad hoc networks. In this thesis, a QoS routing protocol is proposed for OFDMA based mobile ad hoc networks. A signal strength-based sub-channel allocation scheme is proposed in the routing protocol aiming to reduce the signalling overhead and cochannel interference. The performance of the proposed routing protocol is compared with other alternative proposals through simulations using OPNET simulator. Moreover, a partial time synchronization and a null subcarrier based frequency synchronization algorithms are also proposed for OFDMA based ad hoc network to further support and facilitate the proposed sub-channel allocation scheme and routing protocol.