A robust signal detector based on mode weighting

Conventional signal detectors widely used in underwater acoustics usually assume that the environmental parameters are deterministic, which is generally not the case in an actual ocean due to its random variations both in space and time. Compared to the signal fluctuations inducted by the varying environment, the normal mode depth functions which can be extracted from the receiving data demonstrate much higher stability, especially those at lower orders. According to this, a robust signal detector is proposed in this paper. Firstly, the optimal matched mode processing (MMP) detector was derived based on the Neyman-Pearson (NP) criterion. Then the weighted mode amplitude vector was given from the perturbed ocean environment using principal component analysis (PCA) method. Finally, the weighted mode amplitude vectors are matched with the one deduced from the received signal to get a robust MMP detector. Computer simulations show that the detector proposed here can detect the target in an uncertain ocean more efficiently and only the first several principle components are needed in the robust MMP detector.