Low Frequency Active Signal Detection Methodology and Simulation Employing Discrete Wavelet Transforms.

摘要

This document is the final report on the Low Frequency Active / Low Probability of Intercept Sonar Signals research work for DARPA. This report highlights results using discrete wavelet theory in the development of near-optimal detection schemes. We also detail the construction of the sonar signals, their characteristics, and their effectiveness. Since we wish to emphasize practical results over theoretical results in this report, we give only a summary of the theory. Our main premise is that wavelet theory, especially the development of the 'iota 2-closet' discrete wavelet bases, shows substantial performance improvement over standard bases (i.e. the delta basis and the Fourier basis) for signal detection and computational efficiency. Additionally, we have created a simulation system to simulate the effects of transmitting a sonar signal underwater - including Doppler effects and bottom bounce - to test the effectiveness of the detection method as well as to highlight the characteristics of the various sonar signals that we produce. User directions and details of the front-end are given in the User's Guide: Low Frequency Active Simulation System, Reference.