Evolutionary Optimization of Residual Neural Network Architectures for Modulation Classification

Perenda, Erma; Rajendran, Sreeraj; Bovet, Gérôme; Pollin, Sofie; Zheleva, Mariya

doi:10.1109/tccn.2021.3137519

Cited by 10 publications

(19 citation statements)

References 45 publications

(66 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…the literature [22][23][24][25] to compare the modulation recognition accuracy with the Intra-InterNet network.…”

Section: Layer Name Input Size Kernel Size Step Size Number Of Kernelsmentioning

confidence: 99%

“…The CLDNN [24] network feeds both the raw samples and the output of the convolutional layers into the LSTM to provide longer temporal relationships for feature extraction. The 1D CNN [25] network has a low computational cost and is suitable for processing time-series signals, but the recognition accuracy is not high.…”

Section: Layer Name Input Size Kernel Size Step Size Number Of Kernelsmentioning

confidence: 99%

“…However, the classification accuracy of low SNR signals is limited under a complex noise floor. Literatures [22][23][24][25] all use CNN or LSTM or the combination of the two network models to identify the modulation mode of wireless signals, but their recognition accuracy under complex noise is not high. In order to achieve efficient and high-accuracy identification of wireless signal modulation methods in complex noise environments, this paper proposes a modulation signal classification method based on the CNN-LSTM network based on signal periodicity characteristics (Intra-InterNet) for timeseries signals.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

An improved CNN‐LSTM network for modulation identification relying on periodic features of signal

Zhou,

Li,

Wang

2023

IET Communications

View full text Add to dashboard Cite

Aiming at the problem that modulated signals' classification and recognition accuracy is not high in the complex and changeable electromagnetic environment, this paper improves the CNN‐LSTM network and proposes a modulation recognition method of the CNN‐LSTM network (Intra‐InterNet) based on the periodic characteristics of the signal. The method uses the feature extraction ability of the convolutional neural network to extract the correlation features within the cycle and between adjacent cycles of different kinds of modulated signals. At the same time, the relevant features of the signals of non‐adjacent cycles are extracted by using the characteristics of the LSTM. This feature extraction method effectively enhances the feature extraction capability of the network by utilizing the periodicity of the modulated signal. Through the identification and verification of various modulation signals, the experimental results show that: under the signal‐to‐noise ratio of −6 dB ∼ 6 dB, the average identification accuracy rate reaches 87.29%, which effectively improves the accuracy of signal modulation mode identification.

show abstract

“…the literature [22][23][24][25] to compare the modulation recognition accuracy with the Intra-InterNet network.…”

Section: Layer Name Input Size Kernel Size Step Size Number Of Kernelsmentioning

confidence: 99%

Section: Layer Name Input Size Kernel Size Step Size Number Of Kernelsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

An improved CNN‐LSTM network for modulation identification relying on periodic features of signal

Zhou,

Li,

Wang

2023

IET Communications

View full text Add to dashboard Cite

show abstract

“…Efficient model designs [ 10 , 11 ] focus on acceleration over compression through the use of optimized convolutional operations or network architectures. Recently, as a means of improving accuracy, deepening on CNN models has become a popular trend, as demonstrated by ResNet [ 6 ], VGGNet [ 7 ] and Xception [ 12 ].…”

Section: Introductionmentioning

confidence: 99%

Joint design and compression of convolutional neural networks as a Bi-level optimization problem

Louati

Bechikh

Louati

et al. 2022

Neural Comput & Applic

View full text Add to dashboard Cite

Over the last decade, deep neural networks have shown great success in the fields of machine learning and computer vision. Currently, the CNN (convolutional neural network) is one of the most successful networks, having been applied in a wide variety of application domains, including pattern recognition, medical diagnosis and signal processing. Despite CNNs’ impressive performance, their architectural design remains a significant challenge for researchers and practitioners. The problem of selecting hyperparameters is extremely important for these networks. The reason for this is that the search space grows exponentially in size as the number of layers increases. In fact, all existing classical and evolutionary pruning methods take as input an already pre-trained or designed architecture. None of them take pruning into account during the design process. However, to evaluate the quality and possible compactness of any generated architecture, filter pruning should be applied before the communication with the data set to compute the classification error. For instance, a medium-quality architecture in terms of classification could become a very light and accurate architecture after pruning, and vice versa. Many cases are possible, and the number of possibilities is huge. This motivated us to frame the whole process as a bi-level optimization problem where: (1) architecture generation is done at the upper level (with minimum NB and NNB) while (2) its filter pruning optimization is done at the lower level. Motivated by evolutionary algorithms’ (EAs) success in bi-level optimization, we use the newly suggested co-evolutionary migration-based algorithm (CEMBA) as a search engine in this research to address our bi-level architectural optimization problem. The performance of our suggested technique, called Bi-CNN-D-C (Bi-level convolution neural network design and compression), is evaluated using the widely used benchmark data sets for image classification, called CIFAR-10, CIFAR-100 and ImageNet. Our proposed approach is validated by means of a set of comparative experiments with respect to relevant state-of-the-art architectures.

show abstract

“…Automatic modulation awareness (AMA) was initially inspired by its implementation in military scenarios [ 1 , 2 , 3 ]. Electronic warfare and threat analysis are two examples of military uses that require the recognition of signal modulations in order to identify adversary transmitting units, to prepare jamming signals and to recover the intercepted signal.…”

Section: Introductionmentioning

confidence: 99%

Modulation Awareness Method for Dual-Hop Cooperative Transmissions over Frequency-Selective Channels

Marey

Mostafa

2022

Sensors

View full text Add to dashboard Cite

Modulation awareness and cooperative transmissions have individually received a significant amount of research in the scholarly literature. However, a limited number of works are principally concerned with the combination of the two topics, and they are restricted to frequency-flat wireless channels. In this study, we propose a new modulation awareness method applicable to dual-hop amplify-and-forward cooperative broadcasts. The suggested method is built on the creation of theoretical representations of cross-correlation functions of the received signals. We conceptually prove that a family of modulation types generates spikes for certain cross-correlation functions, while others do not. We create a numerous layer hypothesis evaluation for the purpose of making judgments centered on this attribute. The suggested method has a number of benefits, such as the ability to operate on both frequency-flat and frequency-selective channels, as well as the absence of the necessity of channel awareness or noise power. Computer simulations analyze the performance of the proposed method, which delivers adequate awareness performance in a variety of operational scenarios.

show abstract

Evolutionary Optimization of Residual Neural Network Architectures for Modulation Classification

Cited by 10 publications

References 45 publications

An improved CNN‐LSTM network for modulation identification relying on periodic features of signal

An improved CNN‐LSTM network for modulation identification relying on periodic features of signal

Joint design and compression of convolutional neural networks as a Bi-level optimization problem

Modulation Awareness Method for Dual-Hop Cooperative Transmissions over Frequency-Selective Channels

Contact Info

Product

Resources

About