Journal of Applied Sciences

• Articles • Previous Articles     Next Articles

Modulation Classification Using Weighted Constellation Matching

ZHANG Lu-ping; WANG Jian-xin
  

  1. School of Electronic Engineering and Optoelectronic Technology,Nanjing University of Science and Technology,Nanjing 210094,China
  • Received:2008-09-24 Revised:2008-10-30 Online:2009-01-25 Published:2009-01-25
  • Contact: ZHANG Lu-ping

Abstract: Limitation of existing modulation classification based on constellation is studied, and an improved technique proposed. Eleven kinds of signals are divided into three groups with two higher-order moments. Fuzzy C-means clustering is applied to recover the unknown constellations. The recovered constellations are then matched with standard constellations and classified using an improved cost function. Simulation results show that the correct
identification rate is high even if there are more constellation candidates, with the average recognition rate being improved by more than 25% at SNR of 10 dB.

Key words: higher-order moments, constellation, clustering, modulation classification

CLC Number: