University Defence Research Centre (UDRC) in Signal Processing at Imperial College

In many applications, events are presented or displayed visually to an operator who is then responsible for detecting the presence and identity of threat targets using this visual information. This is not always effective for the detection of transient events. Such events are more likely to be detected by an auditory display and operators typically rely on listening to make a decision. This is mostly due to the human auditory system excelling in the detection of transient sounds in the presence of noise and the advantage of combined auditory and visual processing. Notwithstanding this superiority, there is still no effective way to automate this integration of auditory and visual information as part of the system display. Routine experience is that sonar post event analysis detects and classifies targets that were not reported operationally. The aim of this task is to identify candidate signal processing techniques for automated transient detection that exploits combined auditory and visual processing. The emphasis is on ways to integrate the auditory and visual information that characterise transient events. The task will compare early and late integration methods for auditory-visual processing.

Project Supervisor

Mr. Adrian Brown

Adrian Brown BSc MSc is a Principal Scientist in Dstl. His initial research was in ocean measurement and modelling with a particular interest in oceanographic internal waves. This work involved several sea trials including a month in the Atlantic aboard a weather ship. He then spent some time working in Operational Analysis (aka Operational Research) with a focus on engagement simulation modelling of submarine detection. Subsequently he was responsible for the MoD's reference document on Sonar Modelling, leading a team working on sonar performance modelling. More recently he has acted as scientific advisor to a major submarine sonar procurement, which included planning of and participation in submarine sonar trials.

Coming soon

Project Supervisor

Dr. Jonathan Perry

Jon Perry was born in Bristol, UK, in 1962. He studied Physics at Imperial College (BSc 1983, PhD 1988). He joined Space Department in what was then known as the Royal Aircraft Establishment as a Senior Research Fellow in 1986 and has continued to work in the scientific branch of the Civil Service ever since. Jon is now a Principal Scientist in Sensors and Countermeasures Department of Dstl (the Defence Science and Technology Laboratory – RAE’s successor organisation). His research interest is the imaging by aircraft and spaceborne reconnaissance radar of a range of ocean surface features, such as surface expressions of bathymetry and of internal waves, surface waves and wave wakes.

We propose to develop a novel dynamic clustering algorithm which exploits the notions of evolving clustering function and level lines to address the problem of tracking extended objects in complex enviroments where available information are sparse and intermittent. Such situations involve key challenges in the areas of:

• target recognition, classification and localization;
• multi-path mitigation (in that the method is expected to be robust to drop-out and corruption);
• non-stationary processing.

Preliminary investigation has shown that the proposed approach is promising and has set its mathematical foundations. This research aims at further developing the mathematical basis of the approach, to assess its effectiveness on specific case studies and address implementation issues, and to explore its applicability in the context of the key challenges listed above.

Project Supervisor

Prof. Alessandro Astolfi

Alessandro Astolfi was born in Rome, Italy, in 1967. He graduated in electrical engineering from the University of Rome in 1991. In 1992 he joined ETH-Zurich where he obtained a M.Sc. in Information Theory in 1995 and the Ph.D. degree with Medal of Honour in 1995 with a thesis on discontinuous stabilization of nonholonomic systems. In 1996 he was awarded a Ph.D. from the University of Rome “La Sapienza” for his work on nonlinear robust control. Since 1996 he is with the Electrical and Electronic Engineering Department of Imperial College, London (UK), where he is currently Professor in Non-linear Control Theory. From 1998 to 2003 he was also an Associate Professor at the Dept. of Electronics and Information of the Politecnico of Milano. Since 2005 he is also Professor at Dipartimento di Informatica, Sistemi e Produzione, University of Rome Tor Vergata. He has been visiting lecturer in "Nonlinear Control" in several universities, including ETH-Zurich (1995-1996); Terza University of Rome (1996); Rice University, Houston (1999); Kepler University, Linz (2000); SUPELEC, Paris (2001). His research interests are focused on mathematical control theory and control applications, with special emphasis for the problems of discontinuous stabilization, robust stabilization, robust control and adaptive control. He is author of more than 70 journal papers, of 20 book chapters and of over 160 papers in refereed conference proceedings. He is author (with D. Karagiannis and R. Ortega) of the monograph “Nonlinear and Adaptive Control with Applications” (Springer Verlag). He is Associate Editor of Systems and Control Letters, Automatica, IEEE Trans. Automatic Control, the International Journal of Control, the European Journal of Control, the Journal of the Franklin Institute, and the International Journal of Adaptive Control and Signal Processing. He has also served in the IPC of various international conferences.

This proposal aims to investigate the processing techniques which may be applied to acoustic micro-Doppler signature (μ-DS) data. Specifically, methods to extract, classify and track, the μ-DS of  individual targets in the presence of background clutter and non-target backscatter signals will be developed. The proposed programme will build on existing radio frequency (RF) μ-DS research conducted at University College London. This will be carried out in combination with some novel acoustic signal processing approaches and will be supported by a data acquisition programme and signature characterization activities.

The proposed work is based upon the collection or simulation of suitable acoustic data followed by micro-Doppler signal extraction. The experimental equipment should allow a wide range of transmitted waveforms to be investigated and optimised for the target signatures required. The extracted data will then be classified using both methods developed at UCL for RF micro-Doppler data and methods particularly suitable for acoustic data such as those used in speech recognition. Signal processing methods will be developed to characterise target data and will include correction for the acoustic regime propagation conditions. Tracking filter techniques can then be applied to the data and some initial investigation of track classification methods made. This work will be closely supported by our Industrial collaborators, all of whom have significant expertise in acoustic technology. This input will help to ensure that commercially and practically viable ideas are developed.

Project Supervisor

Dr. Karl Woodbridge

Dr. Karl Woodbridge joined University College London in 1990 after 11 years working for Philips Electronics latterly as a project manager in the semiconductor electronics area. He is currently a Reader in Electronic Engineering in the Sensor Systems and Circuits group. He has research interests in the semiconductor and RF sensor areas. His RF research interests are mainly centred on radar sensors and networks. Current RF research activities at UCL include multistatic and netted radar systems, radar target detection, tracking and classification, terrain and marine radar sensing and pervasive passive detection using wireless transmissions. His research activities have been carried out in research investigator and technical consultancy roles for a wide range of customers in the civil and defence areas. He is a Chartered Engineer, a Fellow of the IET, a Fellow of the Institute of Physics and a visiting Professor in the Radar and Remote Sensing Group at the University of Cape Town. He has published or presented over 170 journal and conference papers in the above areas.

The increasing interest in bistatic and multistatic radar systems is a result of the potential they offer in sectors such as  remote sensing, navigation, automatic target recognition and related defence and commercial applications. Advantages of multistatic approaches over conventional monostatic systems include (i) the ability to operate in a covert mode (whereby the receiver may be passive with a relatively close stand off distance to the operational region compared to the transmitter), and (ii) increased survivability employing independent receiver manoeuvring with a reduced receiver cost that incorporate inexpensive passive receive only systems.  It is also possible to use multistatic systems to sample 3-dimensional space by receiving returns from appropriate 3-dimensional flightpaths. The diversity realised when the return from one transmitter is observed by several receivers offers enhanced performance. In a multistatic environment an enhanced radar cross section may be observed allowing targets that were not detectable using monostatic systems to be identified. At the other extreme strong targets that might mask other features in monostatic systems (the polyhedral effect) may be significantly reduced using multistatic radars.  Signal processing is fundamental to effective radar systems. In monostatic SAR imaging, algorithms such as the RDA, FDA, RMA, CSA have been developed and it is natural to aim to extend these to the multistatic domain. This involves revisiting the assumptions and approximations that are made in the monostatic high-resolution radar domain and which may be no longer valid in the multistatic spotlight domain. In multistatic radars used for extracting microdoppler signatures from moving targets techniques such as the Short Time Fourier Transform (STFT) and wavelets have been utilised. It is expected that alternative high-resolution methods will enable greater informative and more robust microdoppler signatures to be extracted. An important aim of the research proposal is to develop new signal processing techniques and demonstrate how they may be used to significantly improve resolution and focusing capability while reducing noise and/or clutter rejection for bistatic spotlight radar imaging and multistatic microdoppler radar signature extraction.. 
The following  key objectives are identified:
 
(i)    To derive new FrFT based algorithms for bistatic spotlight radar imaging methods (HiBistat), including PFA , RMA and CSA for (a) a stationary transmitter and constant velocity receiver, (b) a constant velocity transmitter and constant velocity receiver  and  (c) a stationary transmitter and an accelerating receiver

(ii)    To develop new short-time fractional Fourier transform (STFrFT) frequency signal representation and EMD based high resolution methods for multistatic radar imaging in order to extract robust microdoppler target signature (HiMicro).

(iii)    To develop a Fractional Fourier Transform computational engine that will run on state-of-the-art high performance VLIW DSP/ARM architectures suitable for embedded system designs 

(iv)    To investigate the comparative performance of the new high-resolution methods using realistic simulated models and real data sets

Project Supervisor

Prof. John Soraghan

John J. Soraghan received the BEng and the MEngSc. degrees in 1978 and 1982, respectively, both from University College Dublin, Dublin, Ireland, and the PhD degree in electronic engineering from the University of Southampton, UK, in 1989. His PhD work on Synthetic Aperture Radar Processing was carried out in collaboration with the Royal Aircraft Establishment, Farnborough, U.K. From 1979 to 1980 he joined Westinghouse Electric Corporation, USA as an electronic engineer working on their TPS43 and TPS64 radar systems. In 1986, he joined the Department of Electronic and Electrical Engineering, University of Strathclyde, Glasgow, UK, as a Lecturer in the Signal Processing Division. He became a Senior Lecturer in 1999, a Reader in 2001, and a Professor in 2003. From 1989 to 1991, he was Manager of the Scottish Transputer Centre, and from 1991 to 1995, he was Manager of the DTI Centre for Parallel Signal Processing. Since 1996, he has been Manager of the Texas Instruments’ DSP Elite Centre in the University. He was Head of the Institute for Communications and Signal Processing from 2005-2007. He currently holds the Texas Instruments Chair in Signal Processing in the Centre of Excellence in Signal and Image Processing (CeSIP) University of Strathclyde. His main research interests include advanced linear and non-linear signal processing theory and algorithms with applications to telecommunications; biomedical; multimedia systems; remote sensing and defence. He has been organiser and Technical Chair for the biannual European DSP in Education and Research Symposium (EDERS) since 2004. He has supervised thirty PhD students to graduation, holds three patents, and has published over 230 papers.

Robots, whether operated or autonomous, have widespread applications in, e.g. manufacturing, transport, earth and space exploration, health care and weaponry. The ability to sense and interact with its environment plays an important role for a robot to mimic certain intelligent behaviour as humans. Audition is one of such indispensible senses that are used by humans and animals to recognise their environment in their daily lives. It is therefore highly desirable for a robot to have hearing abilities, to some extent, as a human does. Unfortunately, despite not being limited to only two sensors, robots are still far from approaching the hearing capabilities that are inherent to the human auditory system. A great deal of research in robot audition has been done in the audio domain. It is known however that rather than using only auditory organs, humans are able to infer the meaning of spoken sentences by reading the movement of mouth and facial muscles. In other words, human speech is inherently bimodal: audio and visual in both production and perception. This project attempts to use both the audio and visual modalities for the problem of source separation of target speech in the presence of multiple competing speech interferences and sound sources in room environments for a robotic system. The ultimate goal of the project is to provide progress towards machine perception of auditory scenes within an un-controlled natural environment based on the combination of visual information for the enhancement of audio based blind source separation algorithms.

Project Supervisor

Dr. Wenwu Wang

Wenwu Wang is currently a Lecturer at Centre for Vision Speech and Signal Processing, University of Surrey, where he joined since May 2007. Prior to this, he was a Postdoctoral Research Associate at King's College London (from May 2002 to December 2003) and Cardiff University (from January 2004 to April 2005). He also worked in UK industry, first as a DSP Engineer at Tao Group Ltd (now Antix Labs Ltd) (from May 2005 to August 2006), then as an R&D engineer at Creative Labs (from September 2006 to April 2007). During spring 2008, he has been a visiting scholar at the Perception and Neurodynamics Lab and the Center for Cognitive Science, The Ohio State University. He obtained the PhD degree in April 2002 from Harbin Engineering University, China. His research interests include blind signal processing, audio-visual signal processing, machine learning and perception, and machine audition (listening). He is a member of the IEEE, and belongs to the IEEE Signal Processing, and Circuits and Systems Societies. He has served or currently serves as a reviewer, program committee member, or editor for a number of international journals and conferences.

A popular approach for automatic minehunting with sidescan sonar is to focus on shadow regions of objects. This is thought by many as more dependable than the highlight region. In good conditions, and given prior knowledge, it can be used to accurately classify the object into broad shape classes. However, under changing conditions, and in the presence of clutter such as sand ripples, this simplified approach tends to lose its effectiveness. The highlight region of objects has also been considered for classification but so far results have proven to be too dependent on the specific sonar conditions and the approach is therefore also currently considered unreliable. The difficulty is to extract features that are invariant, or at least tolerant, to shift, scale, orientation, background, and multiple views. Due to the highly textured appearance of modern sidescan sonar imagery, recent efforts have explored the potential of using texture for classification. This has lead to some proposals for features based on fractal measures.

In this project, a novel combination of state-of-the-art feature extraction and classification methods will be brought to bear on the challenging problem of target detection and classification using sidescan sonar imagery. We propose an extension of current texture extraction and classification methods for sidescan sonar target detection by using dual-tree complex wavelet (DTCWT) based local multifractal descriptors and support vector machine (SVM) classifiers for anomaly detection. Our approach will be to extract well localised smoothness and textural descriptors, fractal signatures, and lacunarity using dual-tree complex wavelets. These features will be carefully fed into a support vector machine classifier. In this context, the performance of the DTCWT will be compared directly to other existing wavelet-based fractal extraction methods by performing classification, firstly on some standard texture datasets, and ultimately on sidescan sonar imagery. We will investigate whether lacunarity and other textural descriptors are complementary to fractal and multifractal dimension features.

Project Supervisor

Prof. Nick Kingsbury

Nick Kingsbury is Professor of Signal Processing at the University of Cambridge, Department of Engineering, and head of the Signal Processing and Communications Research Group. He has worked in the areas of digital communications, audio analysis and coding, and image processing. He has developed the dual-tree complex wavelet transform and is especially interested in the application of complex wavelets and related multiscale and multiresolution methods to the analysis of images and 3-D datasets. He has been involved in many projects for the MoD, from his early career days at Marconi Space and Defence Systems to recent projects in the Data and Information Fusion DTC and UDRC programmes.

Target detection in in sonar imagery has been widely researched in the past. Yet, the performances achieved are operationally unacceptable for all but the most benign seabed types. Because of the limited resolution of existing imaging sonars and the strong effect of speckle on target scattering, most algorithms have focused on the acoustic shadow casted by the target when illuminated by the sonar. These techniques fail when the seabed is complex (3D texture). Recent developments in sonar sensors have significantly improved the resolution (synthetic aperture sonars, SAS) and the introduction of interferometric systems enables the joint recovery of imagery and 3D. This project has the following key objectives: 

1- Develop robust models of 3D textures in sonar imagery. This will be based on the use of ground truthed data from interferometric side scan and simulation tools developed in Heriot-Watt University

2- Develop context adaptive detection and classification algorithms. Once clutter models are understood, parameters estimation can be tackled and the models integrated into the detection & classification algorithms.

3- Review generative model based classification using SAS data.

Project Supervisor

Prof. Yvan Petillot

I am a Professor at Heriot Watt University. My main areas of interest are image understanding, sensor fusion and underwater robotics. I am and active member of the Oceans Systems Laboratory and the Signal And Image Group. Finally I am a director ofSeeByte Ltd, a Spin-out of the Oceans Systems Laboratory commercialising some of the technologies developed in the Oceans Systems Laboratory. I hold an Engineering degree in Telecommunications with a specialisation in Image and signal processing from ENSTBr. I also have a M.Sc. in optics and signal processing and a Ph.D. in image processing.

Person tracking has become increasingly important for a series of military and civilian applications like security, surveillance, smart-environments, medicine and others. Typically to track a moving and/or speaking person in a cluttered environment we need to employ microphones or cameras that look for spatio-temporal changes in the data they collect from the monitored environment in order to detect and track people. Nevertheless, these standalone systems have limited success since environments are normally reverberant affecting the performance of the audio systems while the monitored space is often poorly lighted and crowded limiting the performance of the video systems. In this research proposal we envision a person tracking system that will combine new and robust versions of the stand-alone systems in order to provide accurate multi-modal location estimates in real-time. The data fusion mechanisms which will be investigated and employed, will resemble to a degree the way humans locate and track objects by using the cues provided by both our eyes and ears. This assumes development of new technologies that can detect the presence of speech,identify pre-specified objects of interest, attenuate the presence of competing speakers, enhance the lighting conditions, improve or restore the quality of image and speech etc. Employment of the system in realistic environments will ensure that the increased effectiveness of the multi-modality approach is demonstrated at the end of the project.

Project Supervisor

Dr. Tania Stathaki

Tania Stathaki was born in Athens, Hellas. In September 1991 she received the Masters degree in Electronics and Computer Engineering from the Department of Electrical and Computer Engineering of the National Technical University of Athens (NTUA) and the Advanced Diploma in Classical Piano Performance from the Orfeion Athens Conservatory of Music. She received the Ph.D. degree in Signal Processing from Imperial College in September 1994. She is currently a Reader in the Department of Electrical and Electronic Engineering of Imperial College. Previously, she was Lecturer in the Department of Information Systems and Computing of Brunel University in UK, Visiting Lecturer in the Electrical Engineering Department of Mahanakorn University in Thailand and Assistant Professor in the Department of Technology Education and Digital Systems of the University of Pireus in Greece. Her current research interests lie in the areas of signal and image processing and computer vision.