Dynamic adversarial jamming-based reinforcement learning for designing constellations

To resist various types of jamming in wireless channels,appropriate constellation modulation is used in wireless communication to ensure a low bit error rate.Due to the complexity and variability of the channel environment,a simple preset constellation is difficult to adapt to all scenarios,so the online constellation optimization method based on Reinforcement Learning(RL)shows its potential.However,the existing RL technology is difficult to ensure the optimal convergence efficiency.Therefore,in this paper,Dynamic Adversarial Interference(DAJ)waveforms are introduced and the DAJ-RL method is proposed by referring to adversarial training in Deep Learning(DL).The algorithm can converge to the optimal state quickly by self-adaptive power and probability direction of dynamic strong adversary of DAJ.In this paper,a rigorous theoretical proof of the symbol error rate is given and it is shown that the method approaches the mathematical limit.Also,numerical and hardware experiments show that the constellations generated by DAJ-RL have the best error rate at all noise levels.In the end,the proposed DAJ-RL method effectively improves the RL-based anti-jamming modulation for cognitive electronic warfare.

Unsupervised Interpolation Recovery Method for Spectrum Anomaly Detection and Localization

With the growing efficiency of the use of unlicensed spectrum,the challenge of ensuring spectrum security has become increasingly daunting.Spectrum managers aim to accurately and efficiently detect and recognize anomaly behaviors in the spectrum.In this study,we propose a novel framework for spectrum anomaly detection and localization by spectrum interpolation recovery.Spectrum interpolation recovery refers to the recovery of the rest of the spectrum distribution based on a part of the spectrum distribution,which is achieved through a masked autoencoder(MAE)model with a core of multi-head self-attention(MHSA)mechanism.The spectrum interpolation recovery method restores the region where the masked abnormal signals are present,yielding anomaly-free results,with the difference between the restored and the masked representing the anomaly signals.The proposed method has been demonstrated to effectively reduce model-induced over-recovery of anomalous signals and dilute large-scale generation errors caused by anomalies,thereby improving the detection and localization performance of anomaly signals,and improving the area under the receiver operating characteristic curve(AUC)and the area under the precision-recall curve(AUPRC)by 0.0382(3.68%)and 0.1992(68.90%),respectively.On a designed dataset containing 3 variables of interference-to-signal ratio(ISR),signal-to-noise ratio(SNR),and anomaly type,the total recall of anomaly detection and localization at a 5%false alarm rate reached 0.8799 and 0.5536,respectively.Furthermore,a comparative study among different methods demonstrates the effectiveness and rationality of the proposed method.

Decentralized Distributed Deep Learning with Low-Bandwidth Consumption for Smart Constellations

For the space-based remote sensing system,onboard intelligent processing based on deep learning has become an inevitable trend.To adapt to the dynamic changes of the observation scenes,there is an urgent need to perform distributed deep learning onboard to fully utilize the plentiful real-time sensing data of multiple satellites from a smart constellation.However,the network bandwidth of the smart constellation is very limited.Therefore,it is of great significance to carry out distributed training research in a low-bandwidth environment.This paper proposes a Randomized Decentralized Parallel Stochastic Gradient Descent(RD-PSGD)method for distributed training in a low-bandwidth network.To reduce the communication cost,each node in RD-PSGD just randomly transfers part of the information of the local intelligent model to its neighborhood.We further speed up the algorithm by optimizing the programming of random index generation and parameter extraction.For the first time,we theoretically analyze the convergence property of the proposed RD-PSGD and validate the advantage of this method by simulation experiments on various distributed training tasks for image classification on different benchmark datasets and deep learning network architectures.The results show that RD-PSGD can effectively save the time and bandwidth cost of distributed training and reduce the complexity of parameter selection compared with the TopK-based method.The method proposed in this paper provides a new perspective for the study of onboard intelligent processing,especially for online learning on a smart satellite constellation.

Unsupervised Spectrum Anomaly Detection Method for Unauthorized Bands

With the rapid development of wireless communication,spectrum plays increasingly important role in both military and civilian fields.Spectrum anomaly detection aims at detecting emerging anomaly signals and spectrum usage behavior in the environment,which is indispensable to secure safety and improve spectrum efficiency.However,spectrum anomaly detection faces many difficulties,especially for unauthorized frequency bands.In unauthorized bands,the composition of spectrum is complex and the anomaly usage patterns are unknown in prior.In this paper,a Variational Autoencoder-(VAE-)based method is proposed for spectrum anomaly detection in unauthorized bands.First of all,we theoretically prove that the anomalies in unauthorized bands will introduce Background Noise Enhancement(BNE)effect and Anomaly Signal Disappearance(ASD)effects after VAE reconstruction.Then,we introduce a novel anomaly metric termed as percentile(PER)score,which focuses on capturing the distribution variation of reconstruction error caused by ASD and BNE.In order to verify the effectiveness of our method,we developed an ISM Anomaly Detection(IAD)dataset.The proposed PER score achieves superior performance against different type of anomalies.For QPSK type anomaly,our method increases the recall rate from 80%to 93%while keeping a false alarm rate of 5%.The proposed method is beneficial to broadband spectrum sensing and massive spectrum data processing.The code will be released at git@github.com:QXSLAB/vae_ism_ano.git.