Radar signal classification method using HMM and neural networks

摘要

A method for classifying radar signals using HMM and neural network according to the present invention is provided. The method includes a pre-processing process of normalizing, extracting and quantizing feature point RF data and pulse repetition interval (PRI) data; An HMM learning process for learning HMMs (hidden Markov models) for each N1 radio frequency (RF) and N2 pulse repetition interval (PRI) change pattern; And a process of constructing a fully connected multilayer neural network that calculates probability for M characteristics using a fully connected multilayer neural network having a feature value according to the HMM, the minimum and maximum values of the RF and the PRI, and continuously changing For a new type of radar signal, a radar signal classification method using an HMM and a neural network can be provided.