Research on Influencing Factors of Digital Signal Modulation Recognition

Abstract

In the real environment, modulation recognition has low classification recognition rate under low SNR and is affected by many factors such as symbol rate, frequency offset and adjacent channel crosstalk. Based on the combination of high-order cumulants and instantaneous features, this paper firstly analyzes the performance of modulation signal recognition in Gaussian environment. Then through the experimental verification, symbol rate, frequency offset, adjacent channel crosstalk has an impact on the accuracy of modulation recognition. The experimental results show that the ratio of symbol rate and sampling rate has a significant impact on the recognition results, while frequency offset and adjacent channel crosstalk have little impact on the recognition rate.

Keywords

• pattern recognition
• digital modulation
• higher order statistics
• multiple signal classification
• machine learning