NON-LINEAR KEYSTREAM GENERATION FOR ENCRYPTION

摘要

Sensitive information that sends over public channels can be safeguarded by encrypting it. Encryption is the mutation of information into a representation unreadable by anyone without the decryption. In this paper a new encryption system based on the nonlinear property of the Hopfield neural network is proposed. The system is implemented by cascading the Clipped Hopfield Neural Networks in a parallel architecture. Experience results show that the output keystreams have a near optimal linear complexity, which is suitable for cryptographic systems with a high level of security.