In this paper, the problem of chaotic secure communication is addressed from an iterative learning perspective, where the plaintext spans over a pre-specified finite-time interval. In the cryptographic scheme, the plaintext is modulated through chaotic parameter modulation, and then chaotically masked by a nonlinear mechanism. An efficient learning algorithm is exploited for decryption, and the sufficient convergence condition is derived by which the learning gain can be chosen. Furthermore, two fundamental techniques are given in order to enhance security. Case studies are carried out to demonstrate the effectiveness of the proposed method.
展开▼
