ture random number generator based chaotic design help me

[jodenma]

turerandom

now i need design one ture random number generator based chaotic ,but in theory simulation the result cann't pass sp800-22,maybe i need modify my theory equation.help me , can you give me some advice,thanks.

 

[jodenma]

turerandom

by the way,i want to see some article about random number generator based thaotic theory ,can you help me .

 

[piao]

Hi, jodenma,

I am intersted in your work, is it pure logic circuit to implement such a random number generator?

 

[IanP]

Re: ture random number generator based chaotic design help m

Maybe it is not exactly what you are after but I would use "white noise generator" and just digitalize its output.
If you use 8-bit fast A/D converter you have 256-number random generator, if you use 16-bit A/D you have 65536-number random generator...

 

[piao]

that's a good idea, but how to sense white noise? using just the ADC, if so, the white noise signal should be amplified .

Another question is how to prove your radom number is true?

Thansk in advance!

 

[moorhuhn]

Re: ture random number generator based chaotic design help m

I would suggest to use several ring oscillators, where the inverters are done by the weak transistors. The thermal noise will produce a jitter. Assuming you have 4 of them where the typical middle frequency of each oscillator is the prime number, you get a pretty good seed for your random generator.

 

[seinfield]

Re: ture random number generator based chaotic design help m

there are some cellular automatas that generates pseudo-random numbers.

 

[flatulent]

Re: ture random number generator based chaotic design help m

The US NSA says that galactic noise is the best random source. I have seen some systems where the random variation in disk access time was used as the seed. In the thermal noise - ADC approach above I suggest that only the low order bits be used to insure genuine randomness.

To test the results you can do a FFT and see if the spectrum is flat. Another approach is to do an autocorrelation calculation and see if one or more shifts off all give zero value.

 

[jodenma]

now more popular ture random generator is based on white noise . the test standard all most is NIST fips140-2 or sp800-2