Low-complexity cyclostationary-based modulation classifying algorithm

Pedro Rodríguez

doi:10.1016/j.aeue.2017.02.008

Return

Low-complexity cyclostationary-based modulation classifying algorithm

2017-04-01

PRE

DOI：10.1016/j.aeue.2017.02.008

Original

Save

Abstract

En 中文

In this paper a low-complexity cyclostationary-based modulation classifier is presented, which is capable of distinguishing between OFDM, GFSK and QPSK modulations. The classifier computes and analyses the cyclic autocorrelation of the received signals in an implementation-efficient manner. Instead of computing a high number of values of the cyclic autocorrelation like other implementations, which leads to a non-implementable solution, it computes 3 values, allowing a real-time hardware implementation of the algorithms at a limited cost. The performance is evaluated through simulations in MATLAB, under white Gaussian noise and receiver impairments such as frequency offset, I/Q imbalance and DC offset. In order to assess the actual performance and complexity of the classifying algorithm, an FPGA-based implementation is presented in this document. Performance results with real signals are provided, which validate the ones obtained through simulations. (C) 2017 Elsevier GmbH. All rights reserved.

Keywords:

Cognitive radio

Cyclostationary detector

FPGA implementation

Modulation classifier

Spectrum sensing