We propose a framework for joint entropy coding and encryption using Chaotic maps. We begin by observing that the message symbols can be treated as the symbolic sequence of a discrete dynamical system. For an appropriate choice of the dynamical system, we could back-iterate and encode the message as the initial condition of the dynamical system. We show that such an encoding achieves Shannon's entropy and hence optimal for compression. It turns out that the appropriate discrete dynamical system to achieve optimality is the piecewise-linear Generalized Luroth Series (GLS) and further that such an entropy coding technique is exactly equivalent to the popular Arithmetic Coding algorithm. GLS is a generalization of Arithmetic Coding with different modes of operation.
GLS preserves the Lebesgue measure and is ergodic. We show that these properties of GLS enable a framework for joint compression and encryption and thus give a justification of the recent work of Grangetto et al. and Wen et al. Both these methods have the obvious disadvantage of the key length being equal to the message length for strong security. We derive measure preserving piece-wise non-linear GLS (nGLS) and their skewed cousins, which exhibit Robust Chaos. We propose a joint entropy coding and encryption framework using skewed-nGLS and demonstrate Shannon's desired sensitivity to the key parameter. Potentially, our method could improve the security and key efficiency over Grangetto's method while still maintaining the total compression ratio. This is a new area of research with promising applications in communications.
N. Nagaraj, Vaidya, P. G., and Bhat, K. G., “Joint Entropy Coding and Encryption Using Robust Chaos”, arXiv preprint nlin/0608051, 2006.