Key requirements for the design of robust chaotic PRNG

Abstract

The increasing number of e-transactions requires more secure and innovative schemes for secure information storage and transmission. Since the encryption should be unique for each transaction, there is a big necessity of new generators of very huge numbers of encryption keys, and chaotic random number generators seem to be perfectly suitable for this application. In this paper a new robust, gigaperiodic and simple in implementation chaotic generator is proposed. The generator construction is based on the principle of ring-coupling. The proposed chaotic generator successfully passed statistical and analytical tests: NIST, largest Lyapunov exponent, autocorrelation, cross-correlation, uniform distribution. For the best precision of uniform distribution, approximate density function has been applied, the distribution errors are analyzed by written software. The resulting chaotic system promises the designed robust implementation to cryptosystems.