Research on Automatic Modulation Classification System Based on Higher Order Cumulants

• Main Authors: Wei-Cheng Huang, 黃威誠
• Other Authors: Li-Chung Chang
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/3q8b6t

碩士 === 國立臺灣科技大學 === 電機工程系 === 106 === For many applications in the field of communication, such as spectrum surveillance and electronic warfare, which requires to detect the modulation type of a given communication signal. With the increasing variety of digital modulation and complexity, the automatic modulation classification system has become more and more important. The function of the automatic modulation identification system is to detect the received unknown signal and identify its modulation type, and then demodulate the signal to obtain the content transmitted in the signal. This technology is quite important in the field of military intelligence. There are two common methods for automatic modulation identification systems. First one is Maximum Likelihood (ML), and second one is based on Feature-Based (FB). Higher Order Cumulants (HOCs) is one of the methods using FB. The advantage of HOCs is even the signal pass the additivity white Gaussian noise (AWGN) channel, the unknown modulation signal type can still be identified by HOCs. In this paper, we compare 7 different identification systems using HOCs. Under the condition that 12 kinds of modulation type and multi-carrier signals can be successfully identified, the best probability correct classification system and the lowest complexity classification system are proposed. Then we compare the classification system we proposed to the original classification system.