Optimisation algorithm for selecting quadratic time-frequency distributions: performance results and calibration

A methodology based on a resolution performance measure P for time-frequency distributions (TFDs) was defined to optimise TFDs and select the one which results into best concentration of signal components around their instantaneous frequency laws, and best suppression of the interference terms in the time-frequency plane. The algorithm which implements this methodology is described by Sucic and Boashash (see Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing-ICASSP'01, Salt Lake City, Utah, USA, May 2001). In this paper we investigate the performance and robustness of the algorithm for multicomponent signals of linear and non-linear FM laws. We also study its applicability to signals with more-than-two components at different distances from each other, as well as the two-component signals with non-equal amplitudes of the components' mainlobes.