Dr Christos Chousidis

Internet Protocol (IP) suite and Ethernet physical layer are the current trend in home and industrial systems communication protocols. Unfortunately, in the field of power line automation protocols that are in wide use in home automation applications, the lack of IP convergence leads to difficulties, because of the variety and high costs of systems and solutions that need to be integrated. This paper presents a simple and reliable convergence mechanism of open power line Home Automation Protocols to IP (HAPoIP), which is implemented and tested for X10 protocol, on a low cost platform. The novelty and the achievement of the proposed system is the integration of heterogeneous automation networks through IP. In addition, the system was built in terms of software and hardware based on a new protocol.

Voice-to-MIDI real-time conversion is a challenging task that presents a series of obstacles and complications. The main issue is the tracking of the pitch. The frequency tracking of human voice can be inaccurate and computationally expensive due to spectral complexity of voice sounds. Moreover, with microphone-based systems, the presence of environmental noise and neighbouring sounds further affect the accuracy of the frequency tracking. Another issue with the conversion of voice into MIDI, is the presence of non-singing phonemes. As every sound picked up by the microphone would go through the conversion system, any voice or sounded phonemes produced by the user will result in a MIDI output. This research addresses such issues by applying a novel experimental method which employs electroglottography, known to the medical community as EGG, as a source for the pitch tracking operation. Electroglottography improves both the accuracy of the tracking and the ease of processing as it delivers a direct evaluation of the vocal folds operation whilst bypassing any contamination from other sound sources. Furthermore, to address the issue of non-singing phonemes, the proposed method employs the use of neural networks for a real-time classification of the vocal act produced by the user.

A prototype platform was designed, implemented and tested to act as the foundation of a novel intelligent home energy management system which can manage the overall consumption of a house while taking into account user preferences and dynamic tariffs. In order to reduce the implementation cost and to increase the platform stability, the system was designed to employ independent stand-alone devices based on microcontrollers.

The growing need for multimedia applications within wireless Local Area Networks (LAN) demands reliable and efficient broadcasting and multicasting transmission of throughput sensitive data, like audio and video. IEEE 802.11 standard which is the primary technology in wireless LANs was not initially designed to handle heavy broadcasting traffic. However, this raises a series of reliability problems mainly related to the lack of an effective feedback mechanism for multicasting and broadcasting transmission. This inherited drawback does not allow the standard to take full advantage of the bandwidth offered by its latest amendments. The main aim of this work is to offer an alternative congestion control mechanism especially for broadcasting. For this, the expanding use of the CTS-to-Self protection mechanism is proposed. The Medium Access Control (MAC) algorithm is appropriately modified and tested under various data traffic conditions. The simulation shows that the use of this amended MAC method in conjunction with the suitable data packet size can significantly improve throughput, in multimedia type data broadcasting over wireless ad-hoc networks.

Voice-to-MIDI real-time conversion is a challenging problem that comes with a series of obstacles and complications. The main issue is the tracking of the human voice pitch. Extracting the voice fundamental frequency can be inaccurate and highly computationally exacting due to the spectral complexity of voice signals. In addition, on account of microphone usage, the presence of environmental noise can further affect voice processing. An analysis of the current research and status of the market shows a plethora of voice-to-MIDI implementations revolving around the processing of audio signals deriving from microphones. This paper addresses the above-mentioned issues by implementing a novel experimental method where electroglottography is employed instead of microphones as a source for pitch-tracking. In the proposed system, the signal is processed and converted through an embedded hardware device. The use of electroglottography improves both the accuracy of pitch evaluation and the ease of voice information processing; firstly, it provides a direct measurement of the vocal folds' activity and, secondly, it bypasses the interferences caused by external sound sources. This allows the extraction of a simpler and cleaner signal that yields a more effective evaluation of the fundamental frequency during phonation. The proposed method delivers a faster and less computationally demanding conversion thus in turn, allowing for an efficacious real-time voice-to-MIDI conversion.

Gene expression programming (GEP) is a data driven evolutionary technique that well suits for correlation mining. Parallel GEPs are proposed to speed up the evolution process using a cluster of computers or a computer with multiple CPU cores. However, the generation structure of chromosomes and the size of input data are two issues that tend to be neglected when speeding up GEP in evolution. To fill the research gap, this paper proposes three guiding principles to elaborate the computation nature of GEP in evolution based on an analysis of GEP schema theory. As a result, a novel data engineered GEP is developed which follows closely the generation structure of chromosomes in parallelization and considers the input data size in segmentation. Experimental results on two data sets with complementary features show that the data engineered GEP speeds up the evolution process significantly without loss of accuracy in data correlation mining. Based on the experimental tests, a computation model of the data engineered GEP is further developed to demonstrate its high scalability in dealing with potential big data using a large number of CPU cores.

Sound recording and processing techniques can be used in designing diagnostic solutions for a variety of medical conditions related to the respiratory system. In this spectrum, cough monitoring for chronic or seasonal conditions is a significant medical practice. In this paper, a precise cough identification and monitoring system is presented. The system is utilising a convolutional neural network as a feature extraction algorithm and classification system. Including several functions of loading the audio data into the system and converting it into a set of spectrograms, as well as the pre-segmentation stage function, the model retains its relatively low-complexity, which allows accelerating the learning process, also enhanced using dropout. Due to limited audio data available, the dataset dimension was established at 600 samples, split into two equal-numbered groups - 300 samples of "cough" samples, and 300 of "non-cough" samples. The validation accuracy (thus the percentage of samples labelled correctly by the system during the validation process) yielded over 84%, suggesting that this can be a successful cough detection method for future medical applications and devices, such as potential respiratory system condition diagnostic tool.

This paper proposes a new information model of smart consumers where the appliances are not regarded independently such as in traditional scheduling algorithms or home energy management systems, but through a set of common attributes that make up the generic appliance model. Each appliance is generalized by its set of particular attributes. The novel vision is introduced and the attributes of the generic appliance model are explained. The functionality of the system is described with emphasis on the interactions between system components. Under the novel vision, home energy management systems could be developed into widely used practical implementations that would allow better management and user interaction within the home environment.

Background: Heart rate is an essential diagnostic parameter indicating a patient's condition. The assessment of heart rate is also a crucial parameter in the diagnostics of various sleep disorders, including sleep apnoea, as well as sleep/wake pattern analysis. It is usually measured using an electrocardiograph (ECG)-a device monitoring the electrical activity of the heart using several electrodes attached to a patient's upper body-or photoplethysmography (PPG). Methods: The following paper investigates an alternative method for heart rate detection and monitoring that operates on tracheal audio recordings. Datasets for this research were obtained from six participants along with ECG Holter (for validation), as well as from fifty participants undergoing a full night polysomnography testing, during which both heart rate measurements and audio recordings were acquired. Results: The presented method implements a digital filtering and peak detection algorithm applied to audio recordings obtained with a wireless sensor using a contact microphone attached in the suprasternal notch. The system was validated using ECG Holter data, achieving over 92% accuracy. Furthermore, the proposed algorithm was evaluated against whole-night polysomnography-derived HR using Bland-Altman's plots and Pearson's Correlation Coefficient, reaching the average of 0.82 (0.93 maximum) with 0 BPM error tolerance and 0.89 (0.97 maximum) at +/- 3 BPM. Conclusions: The results prove that the proposed system serves the purpose of a precise heart rate monitoring tool that can conveniently assess HR during sleep as a part of a home-based sleep disorder diagnostics process.

The use of network infrastructures to replace conventional professional audio systems is a rapidly increasing field which is expected to play an important role within the professional audio industry. Currently, the market is dominated by numerous proprietary protocols which do not allow interoperability and do not promote the evolution of this sector. Recent standardization actions are intending to resolve this issue excluding, however, the use of wireless networks. Existing wireless networking technologies are considered unsuitable for supporting real-time audio networks, not because of lack of bandwidth but due to their inefficient congestion control mechanisms in broadcasting. In this paper, we propose an amendment of the IEEE 802.11 MAC that improves the performance of the standard for real-time audio data delivery. The proposed amendment is offering a solution for the balancing of data flow density in wireless ad-hoc networks for a multi-broadcasting environment. It is based on two innovative ideas. First, it provides a protection mechanism for broadcasting and second, it replaces the classic congestion control mechanism, based in random backoff, with an alternative traffic adaptive algorithm, designed to minimize collisions. The proposed MAC is able to operate as an alternative mode allowing regular Wi-Fi networks to coexist and interoperate efficiently with audio networks, with the last ones being able to be deployed over existing wireless network infrastructures.

The acts of speaking and singing are different phenomena displaying distinct characteristics. The classification and distinction of these voice acts is vastly approached utilizing voice audio recordings and microphones. The use of audio recordings, however, can become challenging and computationally expensive due to the complexity of the voice signal. The research presented in this paper seeks to address this issue by implementing a deep learning classifier of speaking and singing voices based on bioimpedance measurement in replacement of audio recordings. In addition, the proposed research aims to develop a real-time voice act classification for the integration with voice-to-MIDI conversion. For such purposes, a system was designed, implemented, and tested using electroglottographic signals, Mel Frequency Cepstral Coefficients, and a deep neural network. The lack of datasets for the training of the model was tackled by creating a dedicated dataset 7200 bioimpedance measurement of both singing and speaking. The use of bioimpedance measurements allows to deliver high classification accuracy whilst keeping low computational needs for both preprocessing and classification. These characteristics, in turn, allows a fast deployment of the system for near-real-time applications. After the training, the system was broadly tested achieving a testing accuracy of 92% to 94%.

The research presented in this paper is aiming to address the safety issue that hearing-impaired people arefacing when it comes to identifying a siren sound. For that purpose, a siren identification system, usingdeep learning, was designed, built, and tested. The system consists of a convolutional neural network thatused image recognition techniques to identify the presence of a siren by converting the incoming soundinto spectrograms. The problem with the lack of datasets for the training of the network was addressed bygenerating the appropriate data using a variety of siren sounds mixed with relevant environmental noise. Ahardware interface was also developed to communicate the detection of a siren with the user, using visualmethods. After training the model, the system was extensively tested using realistic scenarios to assess itsperformance For the siren sounds that were used for training, the system achieved an accuracy of 98 per cent.For real-world siren sounds, recorded in the central streets of London, the system achieved an accuracy of 91per cent. When it comes to the operation of the system in noisy environments, the tests showed that the systemcan identify the presence of siren when this is at a sound level of up to -6 db below the background noise.These results prove that the proposed system can be used as a base for the design of a siren-identification application for hearing-impaired people