Study Of The Recognition Of Digital Modulation Signals Based On Modified Self-organizing Feature Map Neural Network

The modulation recognition of communication signals plays a vital role in the signals detection, spectrum monitoring and modern electronic warfare. It requires to accurately judge out the modulation method of received signals based on less prior information. With the development of communication technology, intelligent recognition of digital modulation signals is becoming a new research direction in this field. Artificial neural network have the ability of self-organization learning to input pattern, then formed certain self-awareness. Self-organizing feature map neural network is applied to the modulation recognition in this paper, fully embodied the system of intelligent design.This paper mainly for the following three jobs:1. Constructed five characteristic parameters based on instantaneous characteristic to identify seven common digital signals(2ASK,4ASK,2FSK,4FSK,BPSK,QPSK and 16QAM). The simulation results show the modulated signals instantaneous characteristics graph and characteristic parameters distribution graph.2. Studied the two main factors of K-means clustering algorithm effected clustering results in data clustering. In view of the disadvantages, global K-means clustering algorithm is introduced and improved using K center collocation. The experiment results show the differences in data clustering under the three algorithms.3. The clustering center of K-means algorithm is as the neurons initialized weights vector, and the network's topological structure is changed correspondingly. The simulation results show that the improved algorithm has certain advantages.