Accurate and efficient implementation of the time–frequency matched filter

Abstract

The discrete time-frequency matched filter should replicate the continuous time-frequency matched filter. But the methods differ. To avoid aliasing the discrete method transforms the real-valued signal to the complex-valued analytic signal. The theory for the time-frequency matched filter does not consider the discrete case using the analytic signal. We find that the performance of the matched filter degrades when using the analytic, rather than real-valued, signal. This performance degradation is dependent on the signal to noise ratio and the signal type. In addition, we present a simple algorithm to efficiently compute the time-frequency matched filter. The algorithm with the real-valued signal, comparative to using the analytic signal, requires one-quarter of the computational load. Hence the real-valued signal-and not the * Boualem Boashash is also with The University of Queensland.

Two helicopter classification methods with a high pulse repetition frequency radar

“…9 and 10, we can see that the T-lag has little effect on the classification results whereas the frequency-lag affects the classification results significantly. The reason is (15) and…”

Accelerometer-based fetal movement detection

“…Based on the time-frequency representation of acceleration data, frequency bands suitable for analysis can be determined. Furthermore, information can be extracted from such representations to build time-frequency templates and atoms that can be used in applications like time-frequency matched filter [18] and time-frequency matching pursuit [19]. In this preliminary characterization of acceleration data, we have found that artifacts can mimic fetal movements (see Fig.…”

A passive DSP approach to fetal movement detection for monitoring fetal health

“…Techniques such as TF matching pursuit [17] and TF matched filter [18] may use TF templates and atoms that can be generated based on these features. For instance, instantaneous frequency (IF) and group delay (GD) [19] and morphology of energy distribution [20] can be used as TF features in a detection application.…”