Signal multiple encodings by using autoencoder deep learning

Anaz, Ammar Sameer
Al-Ridha, Moatasem Yaseen
Al-Nima, Raid Rafi Omar

Bulletin of Electrical Engineering and Informatics
Vol 12, No 1: February 2023

Encryption is a substantial phase in information security. It permits only approved persons to get private information. This study suggests a signal multi-encryptions system (SMES) technique for coding and decoding signals created by a deep autoencoder network (DAN). The DAN of four layers is employed for a coding package of signals multiple times before decoding or restructuring the original signals again. The suggested SMES offers a high level of security as it can produce and exploit multiple encryptions for signals. Many statistical calculations are applied to measure the reliability of the system. The outcomes are promising where noteworthy encryptions-decryptions are obtained.