A Front End for Discriminative Learning in Automatic Modulation Classification

Müller, Francisco; Cardoso, Claudomir; Klautau, Aldebaro

doi:10.1109/lcomm.2011.022411.101637

Cited by 35 publications

(20 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Muller et al [11] employed a combination of discriminative learning and support vector machines (SVM) for modulation classification. Mendis et al [9] utilized deep belief networks (DBN) for modulation classification, although DBN has produced very impressive results but they are known to be very difficult to train and scale.…”

Section: Related Work a Modulation Classification Using ML Schemesmentioning

confidence: 99%

Adversarial Machine Learning Attack on Modulation Classification

Usama

Asim

Qadir

et al. 2019

2019 UK/ China Emerging Technologies (UCET)

View full text Add to dashboard Cite

Modulation classification is an important component of cognitive self-driving networks. Recently many ML-based modulation classification methods have been proposed. We have evaluated the robustness of 9 ML-based modulation classifiers against the powerful Carlini & Wagner (C-W) attack and showed that the current ML-based modulation classifiers do not provide any deterrence against adversarial ML examples. To the best of our knowledge, we are the first to report the results of the application of the C-W attack for creating adversarial examples against various ML models for modulation classification.

show abstract

Section: Related Work a Modulation Classification Using ML Schemesmentioning

confidence: 99%

Adversarial Machine Learning Attack on Modulation Classification

Usama

Asim

Qadir

et al. 2019

2019 UK/ China Emerging Technologies (UCET)

View full text Add to dashboard Cite

show abstract

“…In the sequel, the proposed HISTO-SVM is compared with two other AMC systems with respect to computational cost and accuracy: DLRT [6] and CSS [5] combined with linear SVMs. The two baselines were chosen because they were designed for real-time operation and had a lower implementation cost than the systems proposed in [2], [13], [16]- [19].…”

Section: Validation and Comparison With Baselinesmentioning

confidence: 99%

“…Now, their accuracies are compared. Following previous works in AMC (e. g., [5], [16]), the tests adopted modulations: BPSK, 4-PAM, 16-QAM and 8-PSK (hence, binary SVMs), and assumed the AWGN channel. Results with more realistic channels and considering impairments such as frequency offset ( ) were also obtained, but are omitted here because led to similar conclusions and space is limited.…”

Section: Validation and Comparison With Baselinesmentioning

confidence: 99%

“…The systems described in this work belong to the latter class, in which the system is composed by two stages: a front end that converts the input signal into features and a classifier [5]. Several front ends for feature-based AMC have been studied, such as the ones based on cyclostationarity, statistical methods, combination of higher order moments and spectral characteristics and wavelet transform [4].…”

Section: Introductionmentioning

confidence: 99%

“…This classifier is combined with two distinct front ends to compose AMC systems. The first system uses the concatenated sorted symbols (CSS) front end proposed by some of the authors in [5] and plays the role of one of the baselines, together with the discrete likelihood ratio test (DLRT) classifier [6]. A second contribution of this work is the proposal and evaluation of a novel front end called HISTO that, together with linear SVMs, leads to an AMC system that does not require multiplications.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

An Efficient FPGA IP Core for Automatic Modulation Classification

Cardoso

Castro

Klautau

2013

IEEE Embedded Syst. Lett.

Self Cite

View full text Add to dashboard Cite

This letter presents a new algorithm for automatic modulation classification (AMC) and its implementation and validation as an intellectual property (IP) core. AMC aims at accurately identifying the modulation scheme of a given communication system in a short period of time. The proposed IP core consists of a multiclass classifier composed by linear support vector machines and a new parameter extraction (front end) based on histograms. Based on its VHDL implementation and validation using FPGAs, the performance of the proposed system is compared with respect to accuracy and computational complexity to recently proposed AMC algorithms. The new AMC system does not require multipliers and achieves equivalent accuracy with respect to a baseline, while reducing by 50% and 90% the use of logic elements and memory, respectively.Index Terms-Field-programmable gate array (FPGA), intellectual property (IP) core, modulation classification.

show abstract

A survey of traditional and advanced automatic modulation classification techniques, challenges, and some novel trends

Abdel‐Moneim

El‐Shafai

Abdel-Salam

et al. 2021

Int J Communication

View full text Add to dashboard Cite

Automatic modulation classification (AMC) is an important stage in intelligent wireless communication receivers. It is a necessary process after signal detection, and before demodulation. It plays a vital role in various applications. Blind modulation classification is a very difficult task without information about the transmitted signal and the receiver parameters like carrier frequency, signal power, timing information, phase offset, existence of frequency-selective multipath fading, and time-varying channels in real-world applications. The AMC methods are divided into traditional and advanced methods. Traditional methods include likelihood-based (LB) and feature-based (FB) methods. The advanced methods include deep learning (DL) methods. In addition, the AMC methods are used to classify different modulation schemes such as ASK, PSK, FSK, PAM, and QAM with different orders and different signal-to-noise ratios (SNRs). This paper focuses on summarizing the AMC methoods, comparing between them, presenting the commercial software packages for AMC, and finally considering the new challenges in the implementation of AMC. K E Y W O R D S automatic modulation classification (AMC), deep learning (DL), feature-based (FB) methods, likelihood-based (LB) methods | INTRODUCTIONAutomatic modulation classification (AMC) is important in wireless communication systems used in military and civilian applications to enhance the efficiency of the spectrum utilization, redue the overhead, and resolve the shortage problems. Unfortunately, the restricted spectrum resources barely satisfy the ever-increasing demand for 5G 1,2 and Internet of Things (IoT) networks. 3 The AMC can be used for better management of the available spectrum. A simple block diagram of a communication system based on AMC is presented in Figure 1. 4 The AMC architecture contains two steps: signal preprocessing and a proper algorithm for classification. The preprocessing tasks involve reduction of noise, carrier frequency estimation, symbol period estimation, equalization, and signal power evaluation. On the other hand, the AMC methods comprise traditional methods including decisiontheoretic methods and feature-based methods 4,5 along with advanced methods 6 as shown in Figure 2.

show abstract

A Front End for Discriminative Learning in Automatic Modulation Classification

Cited by 35 publications

References 15 publications

Adversarial Machine Learning Attack on Modulation Classification

Adversarial Machine Learning Attack on Modulation Classification

An Efficient FPGA IP Core for Automatic Modulation Classification

A survey of traditional and advanced automatic modulation classification techniques, challenges, and some novel trends

Contact Info

Product

Resources

About