Search

Automatic Modulation Classification (AMC) is a critical issue in electromagnetic spatial perception. Currently traditional recognition techniques are difficult to adapt to complex signal situations. Most existing modulation classification algorithms ignore the complementarity between different features and the importance of feature fusion. Based on this, we proposed a method for image feature fusion for AMC that fully uses the complementarity between different image features. The original signal is converted into an image by the Gramian Angular Field (GAF) method, and the GAF image is used as the input to the network, meanwhile the received signal is converted from the Inphase-Quadrature (I-Q) domain to the r- $\theta $ domain using the Accumulated Polar Feature conversion technique, and the original signal is feature coded from the r- $\theta $ domain and then converted into an image. The fused features of the two images are used as input to the neural network for model training to achieve automatic modulation classification of multiple types of signals. In the evaluation phase, the differences in the recognition effectiveness of the proposed method by different neural networks are discussed. Experiments show that the best performance is achieved using the Swin-Transformer network model, with a more than 90% recognition rate for the modulation method at signal-to-noise ratios greater than 4dB.

Automatic modulation classification (AMC) aims to automatically identify the modulation type of a detected signal in an intelligent wireless receiver, such as software-defined radio (SDR). Recently, deep learning-based methods such as convolutional neural networks have been applied to AMC, showing high-accuracy performance. However, the earlier studies do not consider various signal degradations that can possibly occur during the transmission and reception of wireless signals. Particularly, the signal reception can be often unstable, and the signal can be partially received due to the dynamic spectrum sensing or signal sensing in the intelligent wireless systems. The corrupted signal with missing samples considerably degrades the accuracy of modulation classification of the deep learning-based models, because it is very different from the training datasets. To address this issue, the preprocessing process of restoring the corrupted signal, called signal inpainting, is essential. Although it is significant for the modulation classification, no studies have been performed to investigate the effect of signal inpainting on AMC. To that end, this study proposes a generative adversarial network(GAN)-based signal inpainting method that fills in the missing samples in a wireless signal. The proposed inpainting method can restore the time-domain signal with up to 50% missing samples while maintaining the global structure of each modulation type. The correct recovery of the global structure enables the extraction of distinctive features that play a key role in the modulation classification. To investigate this effect of signal inpainting on AMC, we perform intensive experiments on the RadioML dataset that has been widely used in the AMC studies. We compare the accuracy performance of the two state-of-the-art AMC models with and without the proposed signal inpainting, respectively. Through the analysis of the results, we show that the proposed GAN-based inpainting method significantly improves the accuracy of AMC.

This study proposes robust convolutional neural network (CNN)-based automatic modulation classification (AMC) techniques. Traditional AMCs may be classified into two types: those that rely on ML (maximum likelihood-based AMCs) and those that rely on features. Numerous studies have been conducted on feature-based automatic modulation classification techniques. The current feature-based AMCs lack generalization capability and frequently target a small group of modulation techniques. The current paper develops three different CNN-based AMCs, each with a different classification layer (CL). The adopted classification layers are mean absolute error-based CL, a sum of squared errors-based CL, and crossentropy-based CL. The developed techniques can classify the received signals without feature extraction, where they can learn the features from the transmitted signals automatically during the offline training process, thus eliminating the necessity for feature extraction. A comparison study was done for the proposed CNN-based AMCs with three optimization algorithms at two signal-to-noise ratios. The proposed AMCs attain a true classification accuracy of up to 100% depending on the optimizer and loss function-base CL.

Modulation classification has been widely studied in recent years. However, few studies focus on the performance degradation in multipath fading channels, whose impact is non-negligible. In this paper, a convolutional neural network (CNN) employing cyclostationary (CS) feature, which maintain the essential characteristics in fading channels, is proposed for robust modulation classification. Our method can be implemented in two approaches, referred as CASE1 and CASE2. In CASE1, a single-structured CNN is designed for learning hybrid CS features to perform classification. And in CASE2, we present a CNN model based on a hierarchical structure to perform two-stage classification. Specifically, the coarse classification is performed by learning the second-order CS features with the first-level CNN. Next, another CNN can be selectively activated to learn from high-order CS features for fine classification within the subclass. In this way, our method uses CS features to provide favorable guidance for the learning process of CNN, thus improving the classification performance in fading channels. The experimental results demonstrate the advantages of the proposed method in terms of classification accuracy and computational complexity.

In this paper, we present a comprehensive survey and detailed comparison of techniques that have been applied to the problem of identifying the type of modulation contained within received wireless signals. Known as automatic modulation classification (AMC), the problem has been studied for many decades. AMC plays a significant role in both military and civilian scenarios and is the main step in smart receivers. With the development of software-defined radios and automatic communication systems, IoT technology and the spread of 5G technology, there has been exponential growth in the number of spectrum-using equipment making the issue of scarce spectrum resources more prominent. Although AMC techniques can be optimized from the classifier’s point of view, signal pre-processing also plays a critical role. Relevant data representation approaches include time-frequency analysis, cyclostationary transforms, and hybrid techniques. We provide a taxonomy of common approaches based on order and dimensionality along with an overall analysis of signal pre-processing algorithms for AMC. Furthermore, we reproduce the major existing schemes under uniform conditions, allowing an objective comparison among different methodologies. Finally, we create an open-source reproducible Python library to simulate these techniques, ensuring the usefulness for future research.

Many techniques have been developed for wireless signal recognition in many fifth generation (5G) enabled derivatives. Many harsh constraints, such as the large amount of model parameters and complex signal characteristics, drives intelligent recognition method in real-world settings. In this paper, we propose a generalizable, practical method for raw IQ signal recognition. Specifically, deep complex-valued convolutional neural network models, including a Complex-valued Visual Geometry Group (VGG) (CvVGG) model and a Complex-valued ResNet (CvRN) model, are proposed for handling raw signal IQ data. We examine the merit of complex-valued neural networks (CvNN) and validate their performance with experiments using two public datasets. With an SNR of 10dB, the proposed algorithm achieves a recognition accuracy of 96% on the RadioML2018.10a dataset. When performing drone recognition, CvNN can achieve a recognition accuracy of 99%. Our experimental results verify that deep complex-valued neural network models can achieve considerably improved accuracy with lower computation complexity and fewer model parameters than their real-valued counterparts.

Signal classification is a universal problem in adversarial wireless scenarios, especially when an eavesdropping radio receiver attempts to glean information about a target transmitter’s patterns, attributes, and contents over a wireless channel. In recent years, research surrounding the idea of Machine Learning (ML)-based signal classification has focused on modulation classification, with the downstream objective of demodulation. However, while the computer vision data domain has made significant progress in ensuring robust classification of images despite crafted perturbations, this success has not been translated to secure modulation classification. In this work, we perform the first-ever physical test of an eavesdropping ML-based modulation classifier radio, which we trained offline using a ensemble of i.i.d. models. Each model is trained with a weighted mixture of data perturbed by iterative, “least likely” white box attacks and non-attacked data. We then tested the ensemble online using coaxial-connected Software Defined Radios (SDRs). We conducted a case study comparing our results to the state-of-the-art computer vision approaches to investigate the presence of “label leaking”, model capacity sensitivity, understand the viability of parallel and sequential variations on perturbation training, and assess the effectiveness of iterative attack training. Our results show that perturbations can result in guessing-level classification performance from eavesdroppers, and that varying levels of robustness can be achieved against all presented attacks. These findings confirm that any receiver presents a new attack vector by utilizing ML techniques for classification tasks, and can be vulnerable to evasion attacks at little-to-no cost to transmitters. Consequently, we argue for the use of our training scheme in all ML-based classifying radios where security is a concern.

Deep learning is increasingly being used for many tasks in wireless communications, such as modulation classification. However, it has been shown to be vulnerable to adversarial attacks, which introduce specially crafted imperceptible perturbations, inducing models to make mistakes. This letter proposes an input-agnostic adversarial attack technique that is based on generative adversarial networks (GANs) and multi-task loss. Our results show that our technique reduces the accuracy of a modulation classifier more than a jamming attack and other adversarial attack techniques. Furthermore, it generates adversarial samples at least 335 times faster than the other techniques evaluated, which raises serious concerns about using deep learning-based modulation classifiers. [ABSTRACT FROM AUTHOR]

This paper presents channel-aware adversarial attacks against deep learning-based wireless signal classifiers. There is a transmitter that transmits signals with different modulation types. A deep neural network is used at each receiver to classify its over-the-air received signals to modulation types. In the meantime, an adversary transmits an adversarial perturbation (subject to a power budget) to fool receivers into making errors in classifying signals that are received as superpositions of transmitted signals and adversarial perturbations. First, these evasion attacks are shown to fail when channels are not considered in designing adversarial perturbations. Then, realistic attacks are presented by considering channel effects from the adversary to each receiver. After showing that a channel-aware attack is selective (i.e., it affects only the receiver whose channel is considered in the perturbation design), a broadcast adversarial attack is presented by crafting a common adversarial perturbation to simultaneously fool classifiers at different receivers. The major vulnerability of modulation classifiers to over-the-air adversarial attacks is shown by accounting for different levels of information available about the channel, the transmitter input, and the classifier model. Finally, a certified defense based on randomized smoothing that augments training data with noise is introduced to make the modulation classifier robust to adversarial perturbations. [ABSTRACT FROM AUTHOR]

With the proliferation of frequency-using devices and the advent of the era of big data, spectrum management and control are faced with challenges of effectiveness and accuracy. Modulation classification technology is the foundation and key part of spectrum management and control. Therefore, the effectiveness of modulation classification technology in big data scenario is very important. In this paper, we consider not only the validity of the classification model under the background of big data, but also the dynamics of noise in the complex electromagnetic environment. So we construct a big dataset containing different signals under different MSNR, and use the big data to drive the deep learning model, and finally get the classification result. The proposed method can realize modulation classification only by training one model, which avoids the redundancy of model training in previous algorithms. The simulation results demonstrate the effectiveness and reliability of the proposed method.

Automatic modulation categorization (AMC) is used in many applications such as cognitive radio, adaptive communication, electronic reconnaissance, and non-cooperative communications. Predicting the modulation class of an unknown radio signal without having any prior information of the signal parameters is challenging. This paper proposes a novel multiscale deep-learning-based approach for the automatic modulation classification using radio signals. The approach considered the fixed boundary range-based Empirical wavelet transform (FBREWT) based multiscale analysis technique to decompose the radio signal into sub-band signals or modes. The sub-band signals computed from the radio signal combined with the deep convolutional neural network (CNN) are used to classify modulation types. The approach is tested using the radio signals of different signal-to-noise ratio (SNR) values and four different channel types such as additive white Gaussian noise (AWGN) combination of Rayleigh fading and AWGN channels, the combination of Rician flat fading and AWGN channels, and combination of Nakagami-m fading and AWGN channels for AMC. The results show that the proposed FBREWT based deep-learning approach achieves an overall classification accuracy of 97% for AMC using the radio signals with 10dB SNR for the AWGN channel. Moreover, the proposed approach has obtained the accuracy’s as 94.56%, 95%, and 97.33% using radio signals with 10 dB SNR values for Rayleigh fading, Rician flat fading, and Nakagami-m fading channels combined through AWGN link. The comparison of the proposed multiscale deep learning-based approach with existing methods is shown for AMC. [ABSTRACT FROM AUTHOR]

This letter presents our initial results for real-time orthogonal frequency division multiplexing (OFDM) signal modulation classification based on deep learning and software-defined radio. We generate a modulation classification dataset under a dynamic fading channel, including 6 different OFDM modulation signals, and propose a novel neural network with triple-skip residual stack (TRS) as the basic unit. Each TRS has multiple residual units with gradually increasing convolutional layers. Finally, a near real-time classification system is designed based on the proposed network and GNU Radio. The processing delay incurred by the detection and modulation classification is about 4 ms. It is worth mentioning that the classification accuracy can reach 64% at −10 dB, which is about 7% higher than ResNet and VGG. [ABSTRACT FROM AUTHOR]

In wireless communication, modulation classification is an important part of the non-cooperative communication, and it is difficult to classify the various modulation schemes using conventional methods. The deep learning network has been used to handle the problem and acquire good results. In the deep convolutional neural network (CNN), the data length in the input is fixed. However, the signal length varies in communication, and it causes that the network cannot take advantage of the input signal data to improve the classification accuracy. In this paper, a novel deep network method using a multi-stream structure is proposed. The multi-stream network form increases the network width, and enriches the types of signal features extracted. The superposition convolutional unit in each stream can further improve the classification performance, while the shallower network form is easier to train for avoiding the over-fitting problem. Further, we show that the proposed method can learn more features of the signal data, and it is also shown to be superior to common deep networks.

Deep learning has recently been applied to automatically classify the modulation categories of received radio signals without manual experience. However, training deep learning models requires massive volume of data. An insufficient training data will cause serious overfitting problem and degrade the classification accuracy. To cope with small dataset, data augmentation has been widely used in image processing to expand the dataset and improve the robustness of deep learning models. However, in wireless communication areas, the effect of different data augmentation methods on radio modulation classification has not been studied yet. In this paper, we evaluate different data augmentation methods via a state-of-the-art deep learning-based modulation classifier. Based on the characteristics of modulated signals, three augmentation methods are considered, i.e., rotation, flip, and Gaussian noise, which can be applied in both training phase and inference phase of the deep learning-based classifier. Numerical results show that all three augmentation methods can improve the classification accuracy. Among which, the rotation augmentation method outperforms the flip method, both of which achieve higher classification accuracy than the Gaussian noise method. Given only 12.5% of training dataset, a joint rotation and flip augmentation policy can achieve even higher classification accuracy than the baseline with initial 100% training dataset without augmentation. Furthermore, with data augmentation, radio modulation categories can be successfully classified using shorter radio samples, leading to a simplified deep learning model and a shorter classification response time.

In classification of signal modulation types in MIMO systems, it is difficult to achieve both high accuracy and high computational efficiency at the same time. State-of-the-art likelihood based methods incur massive increase in computational complexity when the number of transmitting antennas and the order of modulation increase. To make modulation classification feasible in time critical systems, we propose a low complexity algorithm with an ensemble of distribution tests. Three goodness-of-fit and a novel variance based distribution tests are employed to examine the mismatch between unknown signal and different modulation hypotheses. The results from all tests are combined by a multilayer perceptron classifier for improved robustness under a variety of channel conditions including AWGN channel and slow fading channels. The resulting solution achieves performance close to the maximum likelihood classifier at high SNR. Yet, it requires much lower computational complexity in all cases.

In modulation classification domain, handcrafted feature based method can fit well from a few labeled samples, while deep learning based method require a large amount of samples to achieve the superior classification performance. In order to improve the modulation classification accuracy under the constraint of limited labeled samples, this paper proposes a few-shot modulation classification method based on feature dimension reduction and pseudo-label training (FDRPLT), which combines handcrafted feature based method with deep learning based method. First, an optimal low-dimensional feature subset is created by the combination of the handcrafted features and autoencoder-extracted features post-processed by a feature selection algorithm. Then, a fully connected network (FCN), trained on a small number of labeled signals, is designed for the automated annotating, where unlabeled samples can be annotated and used for the later convolution neural network (CNN) training. The simulation results show that the classification accuracy of eight kinds of modulation types can reach to 98.3% when the SNR is 20dB.

In this paper, we study the automatic modulation classification in a non-orthogonal multiple access system. To mitigate the effect of interference, a likelihood-based algorithm and a fourth-order cumulant-based algorithm are proposed. Different from the maximum likelihood classifier for a single signal without interference, a likelihood function of the far and near users’ signals is derived. Then, a marginal probability for the far user is obtained by using the Bayesian formula. Hence, the modulation type can be determined by maximizing the marginal probability. The high computational complexity of the likelihood-based algorithm renders it impractical; accordingly, it serves as a theoretical performance bound. On the other hand, we construct a feature vector through the estimated fourth-order cumulants of the received signal including the superposed signal and noise. For each modulation pair, using the mean and covariance matrix of the estimated feature vector, its probability density function can be obtained. Then, the key is to calculate the mean and covariance matrix of the estimated feature vector. To solve this problem, the moments of the superposed signal are derived. Therefore, modulation classification can be performed by maximizing the probability density function. Extensive simulations verify that the two proposed algorithms perform well under a wide range of signal-to-noise ratios and observation lengths. [ABSTRACT FROM AUTHOR]

We consider a communication scenario, in which an intruder tries to determine the modulation scheme of the intercepted signal. Our aim is to minimize the accuracy of the intruder, while guaranteeing that the intended receiver can still recover the underlying message with the highest reliability. This is achieved by perturbing channel input symbols at the encoder, similarly to adversarial attacks against classifiers in machine learning. In image classification, the perturbation is limited to be imperceptible to a human observer, while in our case the perturbation is constrained so that the message can still be reliably decoded by the legitimate receiver, which is oblivious to the perturbation. Simulation results demonstrate the viability of our approach to make wireless communication secure against state-of-the-art intruders (using deep learning or decision trees) with minimal sacrifice in the communication performance. On the other hand, we also demonstrate that using diverse training data and curriculum learning can significantly boost the accuracy of the intruder. [ABSTRACT FROM AUTHOR]

An automatic modulation classification has a very broad application in wireless communications. Recently, deep learning has been used to solve this problem and achieved superior performance. In most cases, the input size is fixed in convolutional neural network (CNN)-based modulation classification. However, the duration of the actual radio signal burst is variable. When the signal length is greater than the CNN input length, how to make full use of the complete signal burst to improve the classification accuracy is a problem needs to be considered. In this paper, three fusion methods are proposed to solve this problem, such as voting-based fusion, confidence-based fusion, and feature-based fusion. The simulation experiments are done to analyze the performance of these methods. The results show that the three fusion methods perform better than the non-fusion method. The performance of the two fusion methods based on confidence and feature is very close, which is better than that of the voting-based fusion.

Digital signal modulation recognition is meaningful for military application and civilian application. In the non-cooperation communication scenario, digital signal modulation recognition will help people identify communication target and have better management over them. In order to the classification accuracy, deep learning is widely used to complete this task. However, current papers have not considered the deployment of deep learning in compute capability and storage limited edge equipment. In this paper, we utilize neural network pruning techniques to reduce the convolution parameters and floating point operations per second (FLOPs), which will pave a wide way to deploy signal classification convolution neural network (CNN) in edge equipment. We set the Average Percentage of Zeros (APoZ) criterion for convolution layers. Compared to original CNN, the experiment result shows that light CNN convolution layer could use only 1.5%~5% parameter and 33%~35% time without losing significant accuracy.

We consider the problem of efficient blind parameter estimation in terahertz (THz)-band ultra-massive multiple-input multiple-output (UM-MIMO) systems. UM-MIMO antenna arrays are crucial to overcoming the distance problem in THz communications. Following recent advancements in THz transceiver design, such arrays are intrinsically compact and reconfigurable. We propose dynamically tuning the modulation types, operating frequencies, and indices of transmitting antenna elements for efficient resource utilization in THz UM-MIMO systems. Furthermore, we propose THz-specific signal processing techniques at the receiver side for blindly detecting the transmission parameters. In particular, we propose a low-complexity antenna index, frequency index, modulation mode, and modulation type detectors, and we study the complexity and performance tradeoffs of the proposed schemes. [ABSTRACT FROM AUTHOR]

The modulation classification problem can be effectively handled by modern machine learning techniques, however, the available data for training is usually limited due to costly measurements in practice. Concerning the problem of insufficient training data, we propose a conditional variational auto-encoder (CVAE)-enhanced learning model for modulation classification. Different from the conventional data augmentation approaches where the data is independently generated, we introduce a feedback unit from the classifier to the CVAE generative network, to ensure that the generated data helps improving the classification accuracy. A two-stage generating-training algorithm is proposed. Via both simulation and practical implementation on a universal software radio peripheral (USRP) platform, it is shown that the proposed method effectively improves the classification accuracy in realistic propagation environment. [ABSTRACT FROM AUTHOR]

Recent advancements in radio frequency machine learning (RFML) have demonstrated the use of raw in-phase and quadrature (IQ) samples for multiple spectrum sensing tasks. Yet, deep learning techniques have been shown, in other applications, to be vulnerable to adversarial machine learning (ML) techniques, which seek to craft small perturbations that are added to the input to cause a misclassification. The current work differentiates the threats that adversarial ML poses to RFML systems based on where the attack is executed from: direct access to classifier input, synchronously transmitted over the air (OTA), or asynchronously transmitted from a separate device. Additionally, the current work develops a methodology for evaluating adversarial success in the context of wireless communications, where the primary metric of interest is bit error rate and not human perception, as is the case in image recognition. The methodology is demonstrated using the well known Fast Gradient Sign Method to evaluate the vulnerabilities of raw IQ based Automatic Modulation Classification and concludes RFML is vulnerable to adversarial examples, even in OTA attacks. However, RFML domain specific receiver effects, which would be encountered in an OTA attack, can present significant impairments to adversarial evasion. [ABSTRACT FROM AUTHOR]

Deep learning (DL) is a newly addressed area of research in the field of modulation classification. In this letter, a constellation density matrix (CDM) based modulation classification algorithm is proposed to identify different orders of ASK, PSK, and QAM. CDM is formed through local density distribution of the signal’s constellation generated using LabVIEW for a wide range of SNR. Two DL models, ResNet-50 and Inception ResNet V2 are trained through color images formed by filtering the CDM. Classification accuracy achieved demonstrates better performance compared to many existing classifiers in the literature. [ABSTRACT FROM AUTHOR]

This paper introduces a novel automatic modulation classification (AMC) method using the graph-based fractional lower-order cyclic-spectrum analysis in the $\alpha$ -stable noise environment. The noise in the practical communication scenario usually exhibits impulse characteristics in the statistical sense, which could be modeled as the $\alpha$ -stable distribution. This would make the second- or higher-order statistics of the received signal vanish, and thus the performances of the conventional AMC algorithms designed for Gaussian noise significantly deteriorate when directly employed in the $\alpha$ -stable noise. In our proposed framework, the fractional lower-order cyclic spectrum (FLOCS) analysis is first invoked to acquire the polyspectra of the signal corrupted by the $\alpha$ -stable noise. Then, the graph-based AMC mechanism is systematically established upon the graph representation of the FLOCS to identify the modulation type according to the discrepancies between the graph features derived from the training and test data. The performance of our proposed new algorithm is theoretically analyzed, and the correct classification probability $P_{\text{cc}}$ over the modulation candidate set is formulated analytically. The remarkable scalability and efficiency of our proposed approach for the modulation candidate set variation are also theoretically investigated in detail. Monte Carlo simulation results demonstrate the effectiveness and superiority of the proposed AMC scheme. [ABSTRACT FROM AUTHOR]

As the process of identifying the modulation format of the received signal, automatic modulation classification (AMC) has various applications in spectrum monitoring and signal interception. In this paper, we propose a dictionary learning-based AMC framework, where a dictionary is trained using signals with known modulation formats and the modulation format of the target signal is determined by its sparse representation on the dictionary. We also design a dictionary learning algorithm called block coordinate descent dictionary learning (BCDL). Furthermore, we prove the convergence of BCDL and quantify its convergence speed in a closed form. Simulation results show that our proposed AMC scheme offers superior performance than the existing methods with low complexity.

In the context of Automatic Modulation Classification (AMC), some recent works have utilized multiple features to train the neural network. With an ultimate aim to develop a systematic approach to select the most diverse and unique features, we propose and demonstrate a novel method to select the most diverse $m\atopwithdelims ()2$ features from a larger feature set. Bhattacharyya distance metric for the dissimilarity between two probability distributions is utilized to select the features with the highest distance for all modulation pairs within a test pool. The proposed approach is analyzed for three different neural networks based classifiers, amidst AWGN and frequency-selective fading channels. A substantial reduction in computational complexity is achieved with an acceptable compromise on the classification performance. [ABSTRACT FROM AUTHOR]

A novel method based on constellation structure is proposed to identify PSK and QAM modulation of different orders, in the slow and flat fading channel. The proposed method does not require training for threshold optimization and considers carrier frequency, symbol rate, and phase offset unknown. The symbol rate is estimated using the spectrum of the instantaneous phase of the complex baseband signal. Carrier frequency offset (CFO) is estimated and corrected from the downconverted signal and downsampled to the estimated symbol rate for extraction of constellation points. The phase offset is determined based on the symmetrical structure of constellation. The features extracted using k-medoids are used for classification of the final modulation scheme. Results show that the proposed algorithm outperforms some existing classifiers and offers lower computational complexity compared to algorithms based on subtractive clustering. [ABSTRACT FROM AUTHOR]

In this paper, modulation classification and symbol decoding problems are jointly considered and optimal strategies are proposed under various settings. In the considered framework, there exist a number of candidate modulation formats and the aim is to decode a sequence of received signals with an unknown modulation scheme. To that aim, two different formulations are proposed. In the first formulation, the prior probabilities of the modulation schemes are assumed to be known and a formulation is proposed under the Bayesian framework. This formulation takes a constrained approach in which the objective function is related to symbol decoding performance whereas the constraint is related to modulation classification performance. The second formulation, on the other hand, addresses the case in which the prior probabilities of the modulation schemes are unknown, and provides a method under the minimax framework. In this case, a constrained approach is employed as well; however, the introduced performance metrics differ from those in the first formulation due to the absence of the prior probabilities of the modulation schemes. Finally, the performance of the proposed methods is illustrated through simulations. It is demonstrated that the proposed techniques improve the introduced symbol detection performance metrics via relaxing the constraint(s) on the modulation classification performance compared with the conventional techniques in a variety of system configurations. [ABSTRACT FROM AUTHOR]

An innovative automatic modulation classification via graph-based constellation analysis for ${M}$ -ary QAM signals is presented. In our framework, a unified mesh model for the constellation diagrams of the ${M}$ -QAM signals within the modulation candidate set is first constructed and exploited to transform the received ${M}$ -QAM signal into graph domain. The concise graph representation of the received ${M}$ -QAM signal is established from its constellation according to the positions of the recovered symbols in the mesh model. Then, the modulation feature vector is built from the eigenvector(s) corresponding to the maximum eigenvalue of its adjacency matrix. The modulation type can be identified by measuring the angle between the feature vectors resulting from the training data and the test data. Monte Carlo simulation results and theoretical analysis demonstrate that the proposed method with lower computational complexity can provide superior performance to the existing subtractive clustering technique, and is robust to the residual phase and timing offsets. [ABSTRACT FROM AUTHOR]

This paper initiates the efforts to design an intelligent/cognitive nano receiver operating in terahertz band. Specifically, we investigate two essential ingredients of an intelligent nano receiver—modulation mode detection (to differentiate between pulse-based modulation and carrier-based modulation) and modulation classification (to identify the exact modulation scheme in use). To implement modulation mode detection, we construct a binary hypothesis test in nano-receiver’s passband and provide closed-form expressions for the two error probabilities. As for modulation classification, we aim to represent the received signal of interest by a Gaussian mixture model (GMM). This necessitates the explicit estimation of the THz channel impulse response and its subsequent compensation (via deconvolution). We then learn the GMM parameters via expectation–maximization algorithm. We then do Gaussian approximation of each mixture density to compute symmetric Kullback–Leibler divergence in order to differentiate between various modulation schemes (i.e., ${M}$ -ary phase shift keying and ${M}$ -ary quadrature amplitude modulation). The simulation results on mode detection indicate that there exists a unique Pareto-optimal point (for both SNR and the decision threshold), where both error probabilities are minimized. The main takeaway message by the simulation results on modulation classification is that for a pre-specified probability of correct classification, higher SNR is required to correctly identify a higher order modulation scheme. On a broader note, this paper should trigger the interest of the community in the design of intelligent/cognitive nano receivers (capable of performing various intelligent tasks, e.g., modulation prediction, and so on). [ABSTRACT FROM AUTHOR]

By adopting the cumulative distribution function of the received signal’s amplitude as feature, a support vector machine-based algorithm is proposed to jointly classify the modulation format and estimate the optical signal-to-noise ratio (OSNR) in coherent optical communication systems. Three commonly-used quadrature-amplitude modulation (QAM) formats are considered. Numerical simulations have been carried out in the OSNR ranges from 5 to 30 dB, and results show that the proposed algorithm achieves a very good modulation classification (MC) performance, as well as high OSNR estimation accuracy with a maximum estimation error of 0.8 dB. Optical back-to-back experiments are also conducted in OSNR ranges of interest. A 99% average correct MC rate is observed, and mean OSNR estimation errors of 0.38, 0.68, and 0.62 dB are noticed for 4-QAM, 16-QAM, and 64-QAM, respectively. Furthermore, compared with the neural networks-based joint estimation algorithm, the proposed algorithm attains better performance with comparable complexity. [ABSTRACT FROM AUTHOR]

Blind and non-cooperative classification of the modulation employed in signals originating from unknown or partly known sources has widespread applications in civilian and military contexts. One of the most recent and interesting approaches to digital modulation which has been enabled by multiantenna transceivers is the spatial modulation, where the indices of the transmit antennas activated in a given symbol period are utilized to transmit information bits. Clearly, existing modulation classification methods, designed for the identification of conventional modulation types, are not capable of classifying the family of spatially modulated signals that make use of the space dimension. In this letter, for the first time in the literature, a modulation classification method is proposed for a modulation type belonging to the family of spatial modulations: the binary space shift keying modulation. [ABSTRACT FROM AUTHOR]

In this letter, we propose a modulation classification algorithm, which is based on the received signal’s amplitude for coherent optical receivers. The proposed algorithm classifies the modulation format from several possible candidates by differentiating the cumulative distribution function (cdf) curves of their normalized amplitudes. The candidate with the most similar cdf to the received signal is selected. The measure of similarity is the minimum average distance between these cdfs. Five commonly used quadrature amplitude modulation formats in digital coherent optical systems are employed. Optical back-to-back experiments and extended simulations are carried out to investigate the performance of the proposed algorithm. Results show that the proposed algorithm achieves accurate classification at optical signal-to-noise ratios of interest. Furthermore, it does not require carrier recovery. [ABSTRACT FROM PUBLISHER]

In this paper, we investigate the cooperative modulation classification problem under multipath scenarios with blind channel information. Multipath channels cause severe degradation on the modulation classification performance, which has not yet been thoroughly solved in the existing literature. To address this issue, a likelihood-based classifier using the expectation-maximization algorithm is proposed, which is capable of finding the maximum likelihood estimates of unknown parameters in a tractable way. Furthermore, to evaluate the upper bound performance of the proposed algorithm, the Cramér–Rao lower bounds of the joint estimates of unknown parameters are derived. Extensive simulations show that the classification performance of the proposed algorithm with good initialization scheme is close to the performance upper bound in the high signal-to-noise ratio region. The results also demonstrate that the proposed algorithm provides significant performance improvement in the multipath channels compared with conventional approaches. [ABSTRACT FROM PUBLISHER]

Based on an online and distributed implementation of the expectation-maximization (EM) algorithm, a hybrid likelihood-based modulation classifier is proposed for a sensor network subject to unknown nonidentical flat fading. The proposed algorithm compares favorably in terms of computational complexity with respect to other maximum likelihood classifiers that rely on the batch (offline) EM algorithm for parameter estimation. Upon the reception of a new sample, each sensor computes the a posteriori probability of the corresponding symbol based on the average consensus algorithm, which relies on local communications among nearby sensors and promotes scalability and power efficiency. Then, each sensor updates its statistics with the innovation obtained from the samples received during the last symbol interval and new estimates for local sensor parameters are computed. Simulation results demonstrate that the proposed online and distributed EM-based classifier can achieve performance close that of a clairvoyant classifier equipped with perfect channel state information. [ABSTRACT FROM PUBLISHER]

The problem of modulation classification for a multiple-antenna (multiple-input multiple-output (MIMO)) system employing orthogonal frequency-division multiplexing (OFDM) is investigated under the assumption of unknown frequency-selective fading channels and signal-to-noise ratio (SNR). The classification problem is formulated as a Bayesian inference task, and solutions are proposed based on Gibbs sampling and mean field variational inference. The proposed methods rely on a selection of the prior distributions that adopts a latent Dirichlet model for the modulation type and on the Bayesian network (BN) formalism. The Gibbs sampling method converges to the optimal Bayesian solution, and using numerical results, its accuracy is seen to improve for small sample sizes when switching to the mean field variational inference technique after a number of iterations. The speed of convergence is shown to improve via annealing and random restarts. While most of the literature on modulation classification assumes that the channels are flat fading, that the number of receive antennas is no less than that of transmit antennas, and that a large number of observed data symbols are available, the proposed methods perform well under more general conditions. Finally, the proposed Bayesian methods are demonstrated to improve over existing non-Bayesian approaches based on independent component analysis (ICA) and on prior Bayesian methods based on the “superconstellation” method. [ABSTRACT FROM PUBLISHER]

A novel automatic modulation classification method based on the graph presentation of the cyclic spectrum is proposed. In our proposed scheme, the periodicity and the symmetry of the cyclic spectrum will be exploited to establish a concise feature representation of multiple graphs. The modulated signal is first transformed from the cycle-frequency domain into the graph domain. Consequently, the concise graph-presentation, namely, a set of weighted directed rings, will be formulated as the robust features of the original signal. Those features can be easily expressed by the corresponding adjacency matrices. It can be verified that the adjacency matrices are sparse and the non-zero entries therein can be registered as the efficient feature parameters. Through the Hamming distance measure to enumerate the difference between the feature parameters resulting from the training data and the test data, one can perform the modulation classification. Monte Carlo simulation results demonstrate that our proposed method can achieve much better classification accuracy than the existing technique when the cyclic spectrum is used. [ABSTRACT FROM PUBLISHER]

The problem of efficient modulation classification (MC) in multiple-input multiple-output systems is considered. Per-layer likelihood-based MC is proposed by employing subspace decomposition to partially decouple the transmitted streams. When detecting the modulation type of the stream of interest, a dense constellation is assumed on all remaining streams. The proposed classifier outperforms existing MC schemes at a lower complexity cost, and can be efficiently implemented in the context of joint MC and subspace data detection. [ABSTRACT FROM PUBLISHER]

In this paper, we consider the problem of automatic modulation classification with multiple sensors in the presence of unknown time offset, phase offset and received signal amplitude. We develop a novel hybrid maximum likelihood (HML) classification scheme based on a generalized expectation maximization (GEM) algorithm. GEM is capable of finding ML estimates numerically that are extremely hard to obtain otherwise. Assuming a good initialization technique is available for GEM, we show that the classification performance (in terms of the probability of error) can be greatly improved with multiple sensors compared to that with a single sensor, especially when the signal-to-noise ratio (SNR) is low. We further demonstrate the superior performance of our approach when simulated annealing (SA) with uniform as well as nonuniform grids is employed for initialization of GEM in low SNR regions. The proposed GEM based approach employs only a small number of samples (in the order of hundreds) at a given sensor node to perform both time and phase synchronization, signal power estimation, followed by modulation classification. We provide simulation results to show the efficiency and effectiveness of the proposed algorithm. [ABSTRACT FROM PUBLISHER]