Instantaneous Signal Detection From The Background Of Gaussian Interference

For passive sonar system operating in the background of strong interference, the most important task is to fulfill signal detection with high reliability. The optimum detector for passive detection is energy detector. Yet energy detector usually deteriorates due to fluctuation of background interference and a decrease in SNR.Bispectrum detection features a good capability in depressing Gaussian interferences, so it outperforms energy detector in the case of instantaneous signal detection. By incorporating bispectrum detection and energy detection method, a coalition detection system can be developed, which can obtain more information by making use of the second and third order statistics of underwater acoustic signals, and can adapt to a broader extent to complicated underwater acoustic environments, thus greatly improving underwater acoustic signal detection performances.In this paper stationary Gaussian random processes as well as bispectrums of instantaneous explosive wave signals are analyzed, and the differences between bispectrums are efficiently used for instantaneous signal detection. As compared with energy detector, bispectrum detector can depress Gaussian noise much better, which is proved via test simulation. In order to further increase instantaneous signal detection performances, we have constructed a coalition detection system upon bispectrum detection and energy detection. Test simulation results show that the incorporation of bispectrum detection and energy detection can bring out a much better capability in noise reduction and signal detection.