3 ms·

> their output distribution is not uniform or gaussian https://www.worldscientific.com/worldscibooks/10.1142/7200#t=aboutBook https://www.worldscientific.com/w

> their output distribution is not uniform or gaussian https://www.worldscientific.com/worldscibooks/10.1142/7200#t=aboutBook https://www.worldscientific.com/worldscibooks/10.1142/7200#t... > "More in details, Ando and Graziani demonstrated that the Chua's circuit can be used to generate either a gaussian or a uniform white noise. To do this, they investigate the statistical and spectral characteristics of the signals generated by a Chua's circuit with respect to different values of the parameters α and β. They then applied the x^2 method to determine which set of parameters leads to a signal having the statistical and spectral characteristics more similar to those of either a Gaussian or uniform noise. By using this method, a signal with a Gaussian-like distribution is obtained with a confidence of 95% for this set of parameters: ..." They've got charts of the probability density functions for each that seem reasonable to me.

cpldcpu 1y ago
Nice, I was not aware of that! Quite interesting. Thank you for the source. It seems to be a corner case. As I learned to know them, chaotic circuits have unpredictable cyclic behavior. Chua's circuit typically follows an oscillatory behavior with a double attractor. If true randomness is the goal, it is much easier to use other sources of randomness like avalanche transistors, jitter of ring-oscillators in the analog domain or LFSRs if you are in the digital domain.