In this talk, we describe a chaotic electronic circuit designed to realize a physical random number generator that is easily integrated. The small footprint of the circuit enables massive parallel realization to achieve high-speed, true-random bit sequences. The analog circuit can be fully characterized, and conjugacy to a symbolic shift proves the presence of chaos. The symbolic representation also provides a rigorous means to extract the maximum entropy from the chaotic device. Analysis of the circuit dynamics reveals critical tunings that yield special Markov properties, which are essential for removing correlations in the random sequences. Practically important is the presence of a sensitive circuit statistic that enables efficient feedback control to the Markov tuning. Numerical simulation and breadboard experimental results demonstrate the effectiveness of the proposed physical random number generator device.

This content is only available as a PDF.