In this paper, the design of a fully-digital chaos-based random bit generator (RBG) is reported. The proposed generator exploits a chaotic system whose map is implemented in the time domain where the state variables of the system are represented by the phase of digital ring oscillators. This results in an extremely robust and efficient entropy source which can be implemented as a digital standard-cell thus overcoming the main drawbacks of chaotic RBGs. An implementation in a 40nm CMOS technology shows a final throughput after post-processing of 12.5Mbit/s at 50MHz with a worst case current consumption below 40μA. The entropy rate of the source can be determined a priori and, in our implementation, it results to be >1.43 bits over 4 bits generated by the source in one clock cycle. After a 16 times compression in a 32-bit linear feedback shift register (LFSR), the final data has full-entropy. A method for a direct evaluation of the entropy after post-processing is provided which can cancel the pseudo-randomness introduced by the LFSR.
CITATION STYLE
Bucci, M., & Luzzi, R. (2016). A fully-digital chaos-based random bit generator. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9100, pp. 396–414). Springer Verlag. https://doi.org/10.1007/978-3-662-49301-4_25
Mendeley helps you to discover research relevant for your work.