Convolutional Radio Modulation Recognition Networks

By Tim O'Shea, Johnathan Corgan, T. Clancy. International Conference on Engineering Applications of Neural Networks, 2016

1,385 citations recorded by the source corpus.

We study the adaptation of convolutional neural networks to the complex-valued temporal radio signal domain. We compare the efficacy of radio modulation classification using naively learned features against using expert feature based methods which are widely used today and e show significant performance improvements. We show that blind temporal learning on large and densely encoded time series using deep convolutional neural networks is viable and a strong candidate approach for this task especially at low signal to noise ratio.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Signal Processing

View Paper: https://www.semanticscholar.org/paper/3bf21df6125bb513ceb95e57de04cc9a7de70af6

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Signal Processing

Citations in the corpus

Deep Learning for RF Device Fingerprinting in Cognitive Communication Networks (2018, methodology): “A CNN operating on the time-domain complex baseband signal has demonstrated high performance at classifying the modulation scheme used by a transmitter [24].”
Machine Learning for Wireless Networks with Artificial Intelligence: A Tutorial on Neural Networks (2017, methodology): “tions. In particular, CNNs have been applied to various domains such as in image and video recognition, recommender systems, and natural language processing (and more recently wireless communications [142]). They constitute a class of deep, feed-forward ANNs that consist of one input layer, one outp”
Spectrum Data Poisoning with Adversarial Deep Learning (2018, background): “cognitive radio tasks such as spectrum sensing [17], spectrum access [18], [19] and modulation recognition [20].”
End-to-End Learning From Spectrum Data: A Deep Learning Approach for Wireless Signal Identification in Spectrum Monitoring Applications (2017, methodology): “ation. In total, 225,225 snapshots were collected. B. CNN network structure The convolutional neural network structure utilized for endto-end learning from spectrum data is derived from O’Shea at al. [9], i.e the CNN2 network, as it has shown to signiﬁcantly outperform traditional signal identiﬁcati”
Deep Power Control: Transmit Power Control Scheme Based on Convolutional Neural Network (2018, methodology): “, autoencoder and recurrent neural network (RNN), as tools for solving complex problems such as the classification of communication signals [5], [8] and impaired channel recovery [9].”
Wireless interference identification with convolutional neural networks (2017, methodology): “[9] compared the performance of different classifiers for classifying 11 modulations, 8 digital and 3 analog modulations.”
Deep Learning Convolutional Neural Networks for Radio Identification (2018, background): “[1] and [8] apply deep learning at the physical layer, specifically focusing on modulation recognition using convolutional neural networks.”
Deep architectures for modulation recognition (2017, methodology): “Guided by work that shows depth being more important than number of feature maps [2] we will establish a baseline convolutional network similar to that used in Radio Convolutional Modulation Networks [13]. Our ﬁrst step is to tune the number of ﬁlters and number of taps per ﬁlter and view those as u”
Over-the-Air Deep Learning Based Radio Signal Classification (2017, methodology): “We showed in [12] and [13] that these methods can be readily applied to simulated radio time series sample data in order to classify emitter types with excellent performance, obtaining equivalent accuracies several times more sensitive than existing best practice methods using feature based classifi”
Grant-Free Non-Orthogonal Multiple Access for IoT: A Survey (2019, methodology): “For the physical layer, learning algorithms have been applied to demodulation [173], channel codes [174], CS [175], coherent [176] and blind [177] detection, etc.”
Adversarial Attacks on Deep-Learning Based Radio Signal Classification (2018, methodology): “For example, [3] uses a convolutional neural network (CNN) for channel decoding, [4] studies DL-based wireless resource allocation, and [5], [6] use DL for the classical task of radio signal (modulation) classification.”
State-of-the-Art Deep Learning: Evolving Machine Intelligence Toward Tomorrow’s Intelligent Network Traffic Control Systems (2017, background): “Furthermore, O’Shea and Corgan [246] demonstrated the viability of using deep CNNs in CRNs.”
Resource Allocation for Multi-Channel Underlay Cognitive Radio Network Based on Deep Neural Network (2018, background): “deep learning, has gained in popularity in fields including image classification and speech recognition [5], [6].”
Automatic Modulation Classification: A Deep Learning Enabled Approach (2018, background): “recognition is proposed in [28], [29], and the simulations with a dataset called RML2016.”
Artificial Neural Networks-Based Machine Learning for Wireless Networks: A Tutorial (2017, methodology): “tions. In particular, CNNs have been applied to various domains such as in image and video recognition, recommender systems, and natural language processing (and more recently wireless communications [142]). They constitute a class of deep, feed-forward ANNs that consist of one input layer, one outp”
Deep Learning Based Transmit Power Control in Underlaid Device-to-Device Communication (2019, background): “Recently, deep learning technology has begun to be applied in wireless communication systems [10], [12], [14]–[16], e.”
Modulation Classification Based on Signal Constellation Diagrams and Deep Learning (2019, neutral): “In [7], [9], and [12], adaptation of DL to the complex temporal signal domain is studied.”