The automatic modulation classification for surface and underwater sensors in the perception layer is crucial in the Internet of Underwater Things (IoUT), where Deep Learning (DL) is becoming an important tool to improve classification accuracy. This work focuses on the radio environment in the perception layer. The biggest challenge in popular DL-based methods is deploying the algorithm in edge devices with limited computing power. Network pruning has been found to be a critical and effective method for network lightweight and the improvement of resources, thus mitigating potential interference. While not studied in previous work, this paper fills the hole in algorithm deployment’s criterion selection and accuracy loss. Specifically, we develop a Convolutional Neural Network (CNN) based lightweight framework on distinguishing modulated signals from generated datasets (which is named DLocean) in different Signal-to-Noise Ratios (SNR). The performance of the lightweight framework is tested on the edge device. The experiments demonstrate that the proposed model compensates for accuracy and can successfully classify the modulation schemes with 93.4% accuracy at the SNR=5 dB. Our results also show that the proposed framework can improve performance without exceeding the original network complexity on edge device deployment.

In military and civilian scenarios, many signal recognition tasks are often accompanied by the growth of signal classes due to signal camouflage or increased communication devices. However, conventional approaches based on deep learning cannot effectively cope with the growth of classes. To address this problem, in this paper, we propose a complex-valued class-incremental learning (CIL) framework for signal recognition (C-SRCIL), which can extract the latent complex-valued features of modulation signals and complete the CIL for complex-valued signal recognition. In our C-SRCIL, We first decouple the feature representation from the classifier to combat catastrophic forgetting, and introduce complex-valued neural networks and an integrated loss function in the feature representation. Then a complex-valued CIL adapter is designed based on the nodalization idea for updating the decoupled classifier in the incremental stage. Experiments on datasets RADIOML2016.10A and SIGNAL2022 show that C-SRCIL brings about a 1.32% performance improvement with a 39.96% reduction in training time compared to the state-of-the-art. It suggests that C-SRCIL extends the classification boundaries of existing models while preserving previous knowledge.

Deep learning (DL) models are vulnerable to adversarial attacks, by adding a subtle perturbation which is imperceptible to the human eye, a convolutional neural network (CNN) can lead to erroneous results, which greatly reduces the reliability and security of the DL tasks. Considering the wide application of modulation recognition in the communication field and the rapid development of DL, by adding a well-designed adversarial perturbation to the input signal, this article explores the performance of attack methods on modulation recognition, measures the effectiveness of adversarial attacks on signals, and provides the empirical evaluation of the reliabilities of CNNs. The results indicate that the accuracy of the target model reduce significantly by adversarial attacks, when the perturbation factor is 0.001, the accuracy of the model could drop by about 50% on average. Among them, iterative methods show greater attack performances than that of one-step method. In addition, the consistency of the waveform before and after the perturbation is examined, to consider whether the added adversarial examples are small enough (i.e., hard to distinguish by human eyes). This article also aims at inspiring researchers to further promote the CNNs reliabilities against adversarial attacks.

With the emergence of the information age, mobile data has become more random, heterogeneous and massive. Thanks to its many advantages, deep learning is increasingly applied in communication fields such as modulation recognition. However, recent studies show that the deep neural networks (DNN) is vulnerable to adversarial examples, where subtle perturbations deliberately designed by an attacker can fool a classifier model into making mistakes. From the perspective of an attacker, this study adds elaborate adversarial examples to the modulation signal, and explores the threats and impacts of adversarial attacks on the DNN-based modulation recognition in different environments. The results show that, regardless of a white-box or a black-box model, the adversarial attack can reduce the accuracy of the target model. Among them, the performance of the iterative attack is superior to the one-step attack in most scenarios. In order to ensure the invisibility of the attack (the waveform being consistent before and after the perturbations), an appropriate perturbation level is found without losing the attack effect. Finally, it is attested that the signal confidence level is inversely proportional to the attack success rate, and several groups of signals with high robustness are obtained.

Network Intrusion Detection Systems (NIDS) have become a strong tool to alarm attacks in computer and communication systems. Machine learning, especially deep learning, has made huge success in fields of industry and academic. Network intrusion activity can be a time series event. In this paper, we adopt a time-related deep learning approach to detect network intrusions. A stacked sparse autoencoder (SSAE) is first built to extract the features with the greedy layer-wise strategy. And then, we propose a time- related intrusion detection system based on the variants of Recurrent Neural Network (RNN). We study the performance of proposed approach on the binary classification with a benchmark dataset UNSW- NB15. Based on the study of parameter time steps, it is proved that our time- related model is effective for intrusion detection. The experiment results show that the accuracy of the proposed approach reaches over 98% and the false alarm rate is as low as 1.8%. The performance of our model is superior to that of the standard RNN- based approach and approaches based on Deep Neural Network and shallow machine learning.

It has been recognized that fixed spectrum and channel allocation will lead to waste of spectrum resources when multiple agents communicate at the same time. Dynamic allocation of channels is proposed to maximize the utilization of spectrum resources. In the environment of multiple unmanned aerial vehicles (UAVs), it is necessary to ensure that each UAV can communicate successfully without interfering with other UAVs. Dynamic allocation of channels plays an important role in such systems. In this paper, we propose a dynamic channel allocation scheme based on deep reinforcement learning for multi-UAV systems. A slotted time system is used by all the UAVs. Di2642erent from the traditional method, the occupancy of each channel is scanned first in each time slot. Then a channel will be selected for data transmission, with feedback from the environment when the transmission is over. The proposed channel allocation scheme incorporates a long short-term memory (LSTM) into the deep reinforcement learning framework, to better learn from the past experience and better adapt to the the highly dynamic environment in a multi-UAV system. The experimental results show that compared with the traditional reinforcement learning method (Q- learning and Deep Q Network (DQN)), the proposed method achieves faster convergence and better performance with respect to average collision rate, average reward, and average successful communication rate.

In the last decade, wireless multimedia device is widely used in many fields, which leads to efficiency improvement, reliability, security, and economic benefits in our daily life. However, with the rapid development of new technologies, the wireless multimedia data transmission security is confronted with a series of new threats and challenges. In physical layer, Radio Frequency Fingerprinting (RFF) is a unique characteristic of IoT devices themselves, which can difficultly be tampered. The wireless multimedia device identification via Radio Frequency Fingerprinting (RFF) extracted from radio signals is a physical-layer method for data transmission security. Just as people’s unique fingerprinting, different Internet of Things (IoT) devices exhibit different RFF which can be used for identification and authentication. In this paper, a wireless multimedia device identification system based on Ensemble Learning is proposed. The key technologies such as signal detection, RFF extraction, and classification model are discussed. According to the theoretical modeling and experiment validation, the reliability and the differentiability of the RFFs are evaluated and the classification results are shown under the real wireless multimedia device environments.