Classification of Digital Modulation Types in Multipath Environments
Abstract
As the expansion of digital communication applications still continues, the need for automated classification of digital modulation types increases. This study attempts to give a partial solution to this problem by proposing a classification scheme which identifies nine of the most popular digital modulation types; namely 2-FSK, 4-FSK, 8-FSK, 2-PSK, 4-PSK, 8-PSK, 16-QAM, 64-QAM and 256-QAM. Higher-order statistics parameters are selected as class features, and a hierarchical neural network-based classifier set-up proposed for the identification of all modulation types considered except those within the M-QAM family. Specific M-QAM types identification is obtained via equalization-based schemes. This study considers the effects due to real-world multipath propagation channels and additive white Gaussian noise. Results show a consistent overall classification performance of at least 68% for severe multipath propagation models and for SNR levels as low as 11dB.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 2001
- Accession Number
- ADA390810
Entities
People
- George Hatzichristos
Organizations
- Naval Postgraduate School