Unsupervised Spectrum Anomaly Detection With Distillation and Memory Enhanced Autoencoders

Spectrum is the fundamental medium for transmitting information services, including communication, navigation, and detection. Spectrum anomalies can lead to substantial economic losses and even endanger life safety. Anomaly detection constitutes a critical component of spectrum risk management. Through spectrum anomaly detection, anomalous spectrum usage behaviors, such as malicious user activities, can be identified. Given the significant limitations of current spectrum anomaly detection algorithms in terms of accuracy and localization capabilities, this paper proposes an approach for detecting spectral anomalies that utilizes knowledge distillation and memory-enhanced autoencoders. First, the pre-trained network with robust feature extraction capabilities is distilled into the teacher network. Subsequently, both an autoencoder and a memory-enhanced autoencoder with an identical structure are trained to predict the teacher network's normalized outputs on a spectrum devoid of anomalies. Finally, in the case of an anomalous spectrum, difference exist between the normalized outputs of the teacher network and the outputs of different student networks, as well as among the outputs of different student networks, which facilitates the process of anomaly detection. The outcomes of experiments reveal that the proposed algorithm is more effective on both synthetic spectral datasets and real IQ signals, demonstrating its proficiency in accurately detecting and locating anomalies.