Mathematical Modeling of Signal Detection in Non-gaussian Correlated Noise

The development of signal detection systems requires complete information about the type of random processes distributions in communication channels with noise. One of the advanced approaches that allows describe random processes is the use of moment and cumulant description of random variables. This approach makes it possible to significantly simplify the synthesis of signal detection systems in noise with a different type of distribution function. The authors of paper proposed the synthesis of the new cumulant models and methods for signal detection in additive correlated non-Gaussian noise. A stochastic polynomial of finite degree was used to synthesize a decision function, the optimal coefficients of which are found according to the adapted new moment quality criterion decision making. The nonlinear processing of signals in noise and taking into account the parameters of correlated non-Gaussian noise in the form of one-dimensional (1D) and two-dimensional (2D) moments can increase the signal processing efficiency compared to traditional Gaussian random process models.