Apparatus and method for stabilizing output sequence of real random number generator

• US 9,841,949 B2
• Filed: 02/05/2015
• Issued: 12/12/2017
• Est. Priority Date: 02/12/2014
• Status: Active Grant

• Current Assignee: Electronics and Telecommunications Research Institute
• Original Assignee: Electronics and Telecommunications Research Institute
• Inventors: Park, Sun-Chun, Ryu, Hyung-Yul, Cho, Young-Jun
• Primary Examiner(s): Sandifer, Matthew
• Application Number:
  US14/614,544
• Publication Number:
  US 20150227344A1
• Time in Patent Office: 1,041 Days
• CPC Class Codes: G06F 7/58 Random or pseudo-random number generators
• CPC Class Codes: G06F 7/588 Random number generators, i.e. based on natural stochastic processes
• CPC Class Codes: G06N 7/02 using fuzzy logic computing arrangements based on biological models G06N3/00; computing arrangements using knowledge-based models G06N5/00

Abstract

An apparatus and method that stabilizes an output sequence of a real random number generator by adjusting the input voltage of the random number generator or the comparative voltage of a comparator. The apparatus for stabilizing an output sequence of a real random number generator includes a verification unit for performing a monobit test on random number values sampled by and output from a comparison unit, and calculating an impact value. An adjustment unit adjusts a voltage of any one of a random number generation unit and the comparison unit using resulting values of the monobit test and the impact value output from the verification unit.

Claims

1. A real random number generator comprising:

a processor; and

one or more units comprising executable code configured and executed by the processor using algorithm associated with a non-transitory storage device,

wherein the one or more units comprise;

a verification unit for performing a monobit test on random number values sampled by and output from a comparison unit, and calculating an impact value; and

an adjustment unit for adjusting a voltage of any one of a random number generation unit and the comparison unit using resulting values of the monobit test and the impact value output from the verification unit,

wherein the adjustment unit comprises;

a fuzzy function unit for fuzzifying the resulting values of the monobit test from the verification unit using a preset fuzzy function;

a fuzzy engine unit for performing fuzzy inference by applying fuzzy values from the fuzzy function unit to a fuzzy rule; and

a defuzzy function unit for converting the fuzzy values indicated by a fuzzy set by the fuzzy engine unit into numerical values corresponding to the fuzzy values.

2. The real random number generator of claim 1, wherein the verification unit performs the monobit test using the following monobit test equation:

X=(n₀−

n₁)²/n

where X denotes calculation results of the monobit test, n₀ denotes a number of 0'"'"'s, n₁ denotes a number of 1'"'"'s, and n denotes a total number of bits.

3. The real random number generator of claim 2, wherein the impact value becomes a negative (−

) value when (n₀−

n₁) is greater than 0 in the monobit test equation, whereas the impact value becomes a positive (+) value when (n₀−

n₁) is less, than 0.

4. The real random number generator of claim 1, wherein the preset fuzzy function is composed of values between probabilities of 0.95 and 0.99 of a chi-square distribution, and respective distribution values are 3.84 and 6.64.

5. The real random number generator of claim 1, wherein the adjustment unit outputs values, obtained by multiplying the values output from the defuzzy function unit by the impact value output from the verification unit, as adjustment values.

6. A method for stabilizing an output sequence of a real random number generator including a verification unit, a comparison unit, a random number generation unit and an adjustment unit, the method comprising:

performing, by the verification unit, a monobit test on random number values sampled by and output from the comparison unit, and calculating an impact value; and

adjusting, by the adjustment unit, a voltage of any one of the random number generation unit and the comparison unit using resulting values of the monobit test and the impact value, thereby stabilizing the output sequence of the real random number generator,

wherein the verification unit, comparison unit, and adjustment unit each comprise executable code configured and executed by a processor using algorithm associated with a non-transitory storage device, and

wherein adjusting the voltage comprises;

fuzzifying the resulting values of the monobit test using a preset fuzzy function;

performing fuzzy inference by applying fuzzy values obtained from fuzzifying to a fuzzy rule; and

converting the fuzzy values indicated by a fuzzy set by the fuzzy inference into numerical values corresponding to the fuzzy values.

7. The method of claim 6, wherein performing the monobit test and calculating the impact value comprises performing the monobit test using the following monobit test equation:

X=(n₀−

n₁)²/n

where X denotes calculation results of the monobit test, n₀ denotes a number of 0'"'"'s, n₁ denotes a number of 1'"'"'s, and n denotes a total number of bits.

8. The method of claim 7, wherein the impact value becomes a negative (−

) value when (n₀−

n₁) is greater than 0 in the monobit test equation, whereas the impact value becomes a positive (+) value when (n₀−

n₁) is less than 0.

9. The method of claim 6, wherein the preset fuzzy function is composed of values between probabilities of 0.95 and 0.99 of a chi-square distribution, and respective distribution values are 3.84 and 6.64.

10. The method of claim 6, wherein adjusting the value further comprises outputting values, obtained by multiplying the numerical values by the impact value, as adjustment values.