A robust signal detection method based on Monte Carlo optimization in uncertain ocean environment

Existing detection methods will have mismatch problem when apply to the real uncertain ocean and this will lead to the performance degradation. Based on the Monte Carlo optimization this paper proposed a robust signal detector using the Bayesian theory and generalized likelihood ratio test (GLRT). The detector uses a priori knowledge of the environment in the GLRT to avoid the nonlinear optimization process. This guarantees the detection performance and reduces the computational complexity at the same time. Matched-ocean detector, mean-ocean detector and energy detector were also present as a comparison. Results from simulations and SWellEx-96 experiment data show that the proposed detector has leading performance over mean-ocean detector and energy detector. The computational consumption is also reduced.