Markov Evaluation of Information Signal
O. V. Opalikhina, Mikhail A. Zhelavskiy · 2025
The article describes an algorithm for stochastic evaluation of a broadband signal isolated against the background of additive noise with constant and nonuniform spectral density. As a criterion for optimal processing of information parameters, the Bayesian criterion of the average risk minimum was chosen. It is proposed to correct the free parameters of the neural chain using a Markov discrete channel with memory. The Markov model allows you to minimize the root mean square error by creating an inverse system. As a neural chain model, the rapidly increasing differentiable logistic function was chosen. Both linear and nonlinear mathematical operations are applicable to the logistic function. Adjusting the slope parameter of the logistic function allows minimizing errors when solving the problem of blind signal separation. Cryptographic protection is achieved by forming a broadband signal according to the law of a pseudo random floating-code M-sequence generated. To simulate Markov discrete channel with memory the Wolfram Mathematica computer environment is used. The program code is written in C++.