High Performance Time-Frequency Distributions for Practical Applications

Abstract

This chapter presents in three interrelated sections the key concepts and techniques needed to design and use high performance time-frequency distributions (TFDs) in real-world practical applications. Section 6.1 first presents, in a heuristic approach, the core concepts forming the field of time-frequency signal processing, incorporating recent developments, such as the design of high resolution quadratic TFDs for multicomponent signal analysis. Section 6.2 outlines methods of assessment of the performance of time-frequency techniques, in terms of the resolution performance of TFDs in separating closely spaced components in the time-frequency domain. A performance measure is defined using key attributes of TFDs, such as the components' mainlobes and sidelobes, and cross-terms. This method of assessment of TFDs performance has led to improvements in designing high resolution quadratic TFD for time-frequency analysis of multi component signals. Section 6.3 presents a methodology for selecting the optimal TFD for a given real-life signal under application-specific constraints. The methodology, based on the performance measure, allows for emphasis of signal features in specific regions of interest in the time-frequency domain.