Specification of channel codes types using the PSD and histogram attributes of the received signal

摘要

Blind detection of communication parameters are one of the main areas of research in urban and military communications. Among these studies many contributions have been devoted to extracting the parameters of block and convolutional codes, assuming the type of code is known. To achieve these purposes the type of the applied code should be specified before. In this article a method according to the histogram comparison of each code type is presented. This comparison is performed using the K nearest neighborhood (KNN) using the cosine, chi-squared, Euclidean distance, Earth Mover's Distance (EMD) and L1 metrics. The simulation results clearly express that how this algorithm successfully can classify these two code types with the probability more than 90 percent. It should be mentioned that this algorithm truly works in high error probability channels.