Detection of narrow-band low-level signals: simulated performance of high-resolution eigenstructure methods versus conventional beamforming
In this report the detection of narrow-band low-level signals by passive sonar is addressed,specifically the detection of a low-level point source in a noise field modelled as a large number ofhigh-level source lines embedded in white additive noise. This model may give, with a conventional processor, a measured noise field where the low-level source can be masked by high-level sources along similar directions. In this context, high-resolution eigenstructure detection algorithms are shown to achieve a significantly higher probability of detection than the conventional beamformer. The performance of four detection algorithms is compared using synthetic data - a companion report will present real data results.
Jesus, Sergio M.