Passive Sonar Localization of a Moving Contact with Interference
Cross correlation of received acoustic signals is a common method for estimating the location of an underwater contact in passive sonar systems. Movement by the contact during the observation period and the presence of an interfering coherent sound source degrade the cross correlation and impair the accuracy of the location estimate. It is necessary to compensate for contact motion in order to avoid degradation of the correlation. One approach is to compress or expand the time axis of one of the received signals prior to correlating, with a scaling factor that compensates for that motion. This approach is computationally expensive. The Select-Correlate-Sum method, proposed in this work, uses an alternate approach to compensating for contact motion that is twice as efficient. This approach uses a scaling factor to select appropriate short-duration extraction pairs from signals received from two sensors. These uncompensated extraction pairs are correlated and then averaged to obtain an adequately long total correlation time. A scaling factor that properly compensates for contact motion also decorrelates a coherent interfering signal, causing it to affect the correlation like noise. In this way the compensation simultaneously avoids the correlation degradation caused by contact motion and the interference. Expressions are derived for the time delay and for the scaling factor which maximize the magnitude of the correlation peak. These expressions are verified using computer simulations. The performance of the Select-Correlate-Sum method is compared to the time-axis scaling method. The Select-Correlate-Sum method yields a somewhat higher variance in location estimate for a given observation period. However, to achieve a given level of location estimate variance, the Select-Correlate-Sum method requires about half the number of computations.
Master of Science (M.Sc.)
Electrical and Computer Engineering