Neural Crypto-Coding as DES: Turbo over Land Mobile Satellite (LMS) channel

Abstract

Modern wireless communications have increased requirements for reliable communication with significant security mechanisms. In this paper cryptography and error correction coding is combined in a single step known as "Crypto-Coding". The focus is on the application of neural networks to Crypto Coding. This approach provides data security for reliable transmission as encryption and error correction takes place in a single step. The computational costs and increase efficiency, mainly in practical system performance is evaluated over Land Mobile Satellite (LMS) channel. We are combining Data Encryption Standard (DES) and Turbo coding in a single step to significantly reduce realizations. A number of experimental studies show applicability of the technique to a selected set of areas in cryptography and error correction coding. Simulation results show that the BER performance of the neural Crypto Coding is close to the Crypto-Coding, for data length equal to 1024 bits, and achieves a BER of about 4 x 10. 5 at an SNR of 10 dB. The proposed algorithms are well suited for implementation on a VLSI-platform.