Past Projects

Modes of Action of Compression Therapy

Professor Bob Williams
Professor Keith Harding (External Professor, University Hospital of Wales)
Dr Michael Clark
Mr Theofanis Bethaves

This study has been consolidated over a number of years. The initial work in this area attracted a £140K grant under the HEFCW Quality Research interactive. The grant was shared with the University of Wales College of Medicine’s Wound Healing Research Unit; this collaboration continues. The original study involved the development of techniques for measuring pressure distribution under bandages, with a view to determining bandaging system efficacy. The outcomes were very successful. The team has been successful in obtaining funding from Smith and Nephew to enable a six-year extension of the programme. The funding supports a senior research fellow, travel and equipment. Currently the work involves multi-sensor ambulatory pressure measurements, investigations into linkages between compression and limb temperature and the assessment of blood flow through bandages. The work is recognised internationally.

Implementation of a Pipeline for a VHDL-microcontroller

Professor Rainer Bermbach (Fachhochschule Braunschweig/Wolfenbüttel)
Dr Ralf Patz
Mr Martin Kupfer (Fachhochschule Braunschweig/Wolfenbüttel)

This work concerns the practical issues surrounding the increase of processing power of microcontroller cores when using a multi-stage pipeline. The project is based around a PIC-compatible VHDL-based microcontroller core which has been developed in a feasibility study at the University of Applied Sciences Braunschweig/Wolfenbuettel. The original PIC from Arizona Microchip contains a 2-stage pipeline and requires four cycles to execute one instruction. The current VHDL-based PIC core achieves 25 MIPS ( 100MHz) in a Xilinx Spartan 3 FPGA, while the original PIC can only achieve 5 – 10 MIPS (20 – 40 MHz). The speed increase is predominantly limited by the FPGA resources. The focus in this project is on the investigation into the increase of processing power of the VHDL-PIC architecture by implementation of a multi-stage pipeline. The limitation due to hazards and appropriate solutions will be investigated. A limit of four to five stages is proposed to minimize the implementation complexity. The study will only focus on the hardware side of multi-stage pipeline implementations.

Integrated Instrumentation Approaches to Wound Assessment

Professor Bob Williams
Professor Keith Harding (External Professor, University Hospital of Wales)
Dr Michael Clark
Mr Theofanis Bethaves

This study commenced in 2000-2001 and complements the bandage efficacy work. It concerns the development and application of instrumentation systems to investigate the role of pressure and microclimate in the formation and treatment of pressure wounds. The work is presently in a review phase.

Novel Wound Assessment Strategies

Professor Bob Williams
Professor M A Wahab
Mr Garry Jackson

This commercial research study is conducted through the Centre for Electronic Product Engineering. The project is for two years in the first instance and is supported by £90k funding from Smith and Nephew.

Image Analysis in the Assessment of Lesions

Dr Andreas Hoppe
Professor Bob Williams
Dr David Wertheim (Kingston University)
Dr WG Jiang (Department of Surgery, University Hospital of Wales)
Professor Keith Harding (External Professor, University Hospital of Wales)

The assessment of cell migration might provide a better understanding in wound healing and cancer metastasis. This study examines the movement of single cancer cells, from microscope images, within a cell cluster. A novel interactive computer system has been developed which enables the description of a cell through a natural cubic spline interpolation. Image processing functions have been developed to facilitate the tracking process of cells in consecutive frames. The work has resulted in a successful PhD submission.

Automatic Recognition of Coarse – Grained Non – Regular Structures using Non – Parametric Processing and Soft Computing Methods

Mr Alexander Stolpmann
Professor Bob Williams
Dr J. Angele (FH Wolfenbttel)

The project examines new statistical and soft-computing methods for the recognition and classification of coarse-grained non-regular objects. An expert system is under development with the objective of enabling automatic texture discrimination, image segmentation and classification. The system uses a fuzzy logic and neural network configuration so providing a versatile tool for image identification. This project is in collaboration with the Computer Science Department at FH Wolfenbuttel in Germany.

Statistical Feature Analysis with Computer-Aided Processing of X-ray Tracings

Mr Frank Höwing
Professor D Wermser (University of Mainz/Trier)
Professor Bob Williams

This project studies the articulatory movements of a variety of different human speakers from a continuous X-ray film. To date over 1500 digitised tracings of the vocal tract shape, tongue and lip positions and the midsagittal profile from the lips to the glottis have been produced. Currently all the feature information extracted has to be manually traced. The aim of the project is to provide a statistical model comprising about 20 auditory parameters that will automatically characterise all the essential aural features from the 2-D X-ray image. It has been found that such features can greatly assist the training of people who are either deaf or have speech impediments. The project has been developed together with the Linguistics Department and Medical Centre at the University of Mainz/Trier.

The Application of Optical Textural Analysis Techniques to the Inspection of Coated Steel Production

Dr Scott Woodham
Professor Bob Williams
Dr John Rees (Stafford University)

The quality of hot-dipped iron/zinc coated steel (Galvanneal) is ultimately dependent on a number of parameters which are controllable to varying extents (zinc bath chemistry, coating conditions, furnace annealing, etc.) The aim of the project is to develop an appropriate inspection algorithm and classifier system to recognise the optimal alloyed state from optical surface measurements. Techniques based around the Wavelet Transform, GLCM measurements and conditional morphological methods have been developed. The evaluation of the extracted features is based on statistical and soft computing techniques.

The work has resulted in a successful PhD submission.

The Analysis of Doppler Blood Flow Signals using DSP Methods and Pattern Recognition Techniques

Professor M A Wahab
Professor R J Williams
Mr Ralf Patz

This project concerns investigation into novel analyses of Doppler ultrasound blood flow signals.
The aim is to arrive at efficient techniques of processing and feature extraction of the forward and reverse blood flow signals to enable quantitative assessment of vascular diseases in the upper and lower limbs. The signal analysis work will concentrate on the application of multi-resolution time-frequency representations of the ultrasound signals. The widely used spectrogram, which is based on the short time Fourier Transform, does not provide sufficient features that allow accurate classification of the degree of the vascular disease or the patient’s response to treatment. Multi-resolution time-frequency analysis may provide more information about the blood turbulence and hence, indicate stenosis in a blood vessel. Statistical and neural network recognition techniques will be developed to provide the quantitative measures required by the medical community.

Instrumentation for Stress Measurements in Structural Anchors

Mr John Dunn
Professor Bob Williams
Mr Keith Minton (Cintec International)
Mr Clive Thomas

This project is conducted collaboratively with Cintec International in Newport, Gwent. The aim of the project is to develop instrumentation that will enable the installation of structural anchors to be better controlled. Additionally, the intention is to embed instrumentation in anchors that will allow ongoing stress measurements of the anchors and also enable the assessment of building structures under dynamic loading conditions. The instrumentation development concerns stress transduction and signal processing and communication ASIC development. The work is supported under the Teaching Company Programme.

Harmonic Control in AC Power Applications

Mr Costakis Nicolaou
Professor M A Wahab

This study concerns the development of new approaches to controlling harmonics in AC-AC power converters. The anticipated benefits are reduced acoustic noise and greater efficiency. The work has three strands; real-time monitoring of supply and load harmonics, novel waveform switching and synthesis techniques, and innovative methods of waveform injection for harmonics control. It is anticipated that the control strategies will also result in reduced EMI problems.

ISDN Test methodologies

Mr Patrick Sugrue
Professor M A Wahab

This work concerns the development of test methodologies for ISDN system development. There are three lines of work. Spectral analysis techniques to assess the data transmission characteristics of twisted pair transmission lines for; an investigation into passive D channel monitoring to diagnose faults on the ISDN Sbus; investigations into bit error rate analysis to measure data transmission quality on the ISDN S bus. Test system development includes the application of ASIC design for fast implementation of spectrum analysis and other signal processing strategies.