Time- Frequency signal processing (TFSP) is the representation and processing of signals in a joint time-frequency domain. Most natural and man-made signals are non-stationary; that is, their frequency content varies with time. Classical signal processing, which represents signals in terms of time alone or frequency alone, is often unsuitable for processing non-stationary signals because it does not transparently represent the variation of frequency content with time. TFSP overcomes this limitation by representing the signal as a function of time and frequency, known as a time- frequency distribution (TFD).

Although the variety of possible TFD is literally infinite, many independently discovered TFDs are special cases of the quadratic class. All members of the class are linear transforms of each other. Each member is defined by kerner function, and many desirable properties of TFDs can be expressed as constraints on the kernel function. Early developments in radio communication made it necessary to extend the notion of frequency (previously considered time-invariant) to allow for a time-dependent instantaneous frequency (IF). Attempts to define IF sowed the seeds of time-frequency signal analysis (TFSA), and the need to measure IF motivated much subsequent research in TFSA, including the author's own work on polynomial TFDs.

Whereas past research in the field has tended to concentrate on analysis, more recent research has extended over the full range of processing, including filtering, detection and reconstruction of signals. Thus TFSA grew into TFSP.

This lecture:

Introduces concepts, algorithms and applications of TFSP in an rigorous but intuitive manner, making maximal use of classical signal-processing concepts familiar to students and practicing engineering;

Introduces the concept of instantaneous frequency (IF) and its estimation; Discusses the relationships between time-frequency representations and other representations in terms of time and lag (t,\tau), Doppler and lag (\nu,\tau), and Doppler and frequency(\nu,f);

Introduces a range of TFSP techniques and discusses the underlying assumptions affecting their applicability;

Illustrates the power of TFSP in solving real-life problem, including analysis, synthesis, filtering, estimation, detection and classification of signals, that would not be tractable using classical methods, in applications selected form telecommunications, radar, sonar, and seismic processing, source identification, medical diagnosis and machine fault detection;

Demonstrates software implementations of TFSP methods; Shows the relationships between TFSP and other emerging fields of signal processing, such as higher-order spectra and wavelets;

Offer exercises by which participants may test their knowledge and develop basic skills. Explore, using real-life signal, the various aspects of time-frequency signal analysis in a step-by-step tutorial approach. The tutorial is based on an easy to use Time-Frequency Signal Analysis (TFSA) package designed by the author.

Biography

chProfessor Boualem Boashash is a fellow of the IEEE "for contributions to time-frequency signal analysis and signal processing education". He is also a fellow of IE Aust and a Fellow of the IREE.

He obtained a Diplome d'ingenieur-Physique - Electronique from the ICPI University of Lyon, France, in 1978. He received an M.S. degree from the Institut National Polytechnique de Grenoble, Grenoble, France in 1979. He received his Doctorate (Docteur-Ingenieur) from the same university in May 1982.

In 1979 he joined Elf-Aquitaine Geophysical Research Centre, Pau, France, where he was a research engineer in the signal processing group. In May 1982, he joined the Institut National des Sciences Appliquees de Lyon, Lyon, France, where he was a Maitre-Assistant associe. In January 1984, he joined the Electrical Engineering Department of the University of Queensland, Australia, as a Lecturer, Senior Lecturer (1986) and Reader (1989). In 1990, he joined Bond University, Graduate School of Science and Technology, as Professor of Signal Processing. In 1991, he joined the Queensland University of Technology as the foundation Professor of Signal Processing and Director of the Signal Processing Research Centre.

Professor Boashash was the technical chairman of ICASSP 94, the premium conference in Signal Processing. He is currently the chairman of the International Symposium on Signal Processing and its Applications (ISSPA), organised regularly in Australia since 1987. He has written over two hundred technical publications, edited two books and has supervised over twenty PhD students and five Masters students. Professor Boashash's research interests include: time-frequency signal analysis, spectral estimation, signal detection and classification, and higher-order spectra. He is also interested in wider issues such as the effect of engineering developments on society.