Deep Learning based Effective Watermarking Technique for IoT Systems Signal Authentication

Abstract

In order to identify cyber-attacks, this research suggests a special watermarking technique for dynamic IoT System signal validation. IoT Systems (IoTSs) can extract a group of randomly generated characteristics from their produced signal and then periodically watermark these attributes into the transmission owing to the proposed efficient watermarking technique. Using dynamic watermarking for IoT signal authentication, a potent deep learning technique is used to detect cyber-attacks. Based on an LSTM structure, the proposed learning system enables IoT devices to extract a set of random features from the signal they release, hence enabling dynamic watermarking of the signal.