Deep learning-based automatic modulation recognition networks are susceptible to adversarial attacks, posing significant performance vulnerabilities. In response, we introduce a defense framework enriched by tailored autoencoder (AE) techniques. Our design features a detection AE that harnesses reconstruction errors and convolutional neural networks to discern deep features, employing thresholds from reconstruction error and Kullback–Leibler divergence to identify adversarial samples and their origin mechanisms. Additionally, a restoration AE with a multi-layered structure effectively restores adversarial samples generated via optimization methods, ensuring accurate classification. Tested rigorously on the RML2016.10a dataset, our framework proves robust against adversarial threats, presenting a versatile defense solution compatible with various deep learning models.

With the development and application of deep learning in signal detection tasks, the vulnerability of neural networks to adversarial attacks has also become a security threat to signal detection networks. This paper defines a signal adversarial examples generation model for signal detection network from the perspective of adding perturbations to the signal. The model uses the inequality relationship of L2-norm between time domain and time-frequency domain to constrain the energy of signal perturbations. Building upon this model, we propose a method for generating signal adversarial examples utilizing gradient-based attacks and Short-Time Fourier Transform. The experimental results show that under the constraint of signal perturbation energy ratio less than 3%, our adversarial attack resulted in a 28.1% reduction in the mean Average Precision (mAP), a 24.7% reduction in recall, and a 30.4% reduction in precision of the signal detection network. Compared to random noise perturbation of equivalent intensity, our adversarial attack demonstrates a significant attack effect.