High Resolution Time-Frequency Representations

The objective of the proposed Phase I work is to demonstrate the feasibility of using the Choi-Williams distribution to estimate the time ¿ frequency distribution of transients of interest to the IUSS. The specific objectives are to develop a metric for IUSS images and to show that the Choi-Williams distribution of these images may be used for classification. The Choi-Williams distribution is a modification of the Wigner distribution which allows infinite resolution in time and frequency simultaneously but does not exhibit the unwanted artifacts of the Wigner distribution. Amron has demonstrated its value for transients recorded at sonobuoys and has hardware and software to calculate it now. The metric will measure of the information in the Choi-Williams distribution display and the processing load required to form the display. The tasks will be to collect data; to develop the metric, to calculate the Choi-Williams distribution and the metric for the data; to write a final report; and, optionally, to move the code to the ADS hardware and operating system. In Phase II we will implement the Choi-Williams in hardware and software suitable for an IUSS site.