JP2009217513A - Random number generator and random number generation method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To generate a true random number value by a method other than measuring time intervals based on a phenomenon independently occurring on the basis of a physical phenomenon. <P>SOLUTION: The random number generator includes: a circuit for generating a random signal from a noise source; a circuit for generating an output waveform where voltage simply increases as time goes by in a prescribed cycle; a voltage value acquisition circuit for determining that the random signal has been generated and acquiring a voltage value when the last random signal has been generated from the output waveform; and a circuit for converting the voltage value into a digital value to output it as a random number value. Thus, it is possible to determine a random number value on the basis of the voltage value in the voltage waveform where the voltage simply increases as time goes by in the prescribed cycle and the random number value can be generated without using a measuring means by a counter. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a technique using a random pulse signal generated from a noise source or the like, and more particularly to a random number generation device and a random number generation method for generating a natural random number.

  In recent years, random numbers have come to be frequently used for various devices including game machines, or for processing in experiment planning and encryption key generation. Random numbers are often generated by software processing using a computer. However, the generation of random numbers by software is limited in terms of true random numbers with strict irregularities. On the other hand, for example, random numbers can be generated based on random pulses obtained by using white noise generated from thermal noise sources such as resistors, diodes, or conductors. In this case, it can be regarded as a genuine random number. It has been known.

  When a random pulse from a thermal noise source is continued on the time axis and a pulse exceeding a predetermined threshold (hereinafter referred to as “sampling pulse”) is detected, an independent event occurs. If so, the occurrence interval between these independent events is known to follow an exponential distribution. That is, the value represented by the time interval between sampling pulses can be regarded as a natural random number according to an exponential distribution.

  By the way, the time interval between the sampling pulses, that is, the time interval (t) from the previous sampling pulse generation time to the next generated sampling pulse is measured using the time measuring means. This time measuring means generally includes a clock signal generating means for generating a clock signal having a frequency higher than that of the sampling pulse, and a counting means (counter) for counting the number of clocks of the clock signal. The time interval (t) between sampling pulses is measured based on the counted number of clocks. In FIG. 5, the time interval between sampling pulses is measured by a counter, and the count values are set to a random value 1, a random value 2, a random value 3,.

  However, the time interval measured by the counter as described above is represented by an integer value from 0 to 255 in the case of an 8-bit counter, for example, and the time (t) that is originally expressed as a continuous amount is represented by a discrete value of the counter. This means that it is expressed approximately with various values. Of course, if the number of bits of the counter is increased to increase the range of values that can be taken as discrete values, the accuracy of quantization is improved, but it is still digital data.

  Therefore, an object of the present invention is to generate a genuine random number value from an event that occurs independently based on a physical phenomenon by a method other than measurement of a time interval.

In order to achieve the above object, a random number generator according to the present invention includes a noise source, a circuit that generates a random signal from the noise source, and a circuit that generates an output waveform whose voltage monotonously increases with time at a predetermined period. A voltage value acquisition circuit that detects the occurrence of the random signal and acquires the voltage value at the time when the random signal was generated last time from the output waveform; and converts the voltage value into a digital value. And a circuit for outputting as a random value.
Also, the output waveform by the random number generator of the present invention is a sawtooth wave or a logarithmic curve, and the circuit for generating the output waveform is the output waveform in response to the voltage value acquisition circuit detecting a random signal. Is generated again from the initial state.

In order to achieve the above object, a random number generation method according to the present invention includes a process for detecting a random signal generated from a noise source, and a waveform generation process for generating an output waveform whose voltage monotonously increases with time in a predetermined cycle. A voltage value acquisition process for acquiring a voltage value at the time when the random signal is generated from the output waveform, and a process for A / D converting the voltage value and outputting it as a random value.
Further, the output waveform according to the random number generation method of the present invention is a sawtooth wave or a logarithmic curve, and further, the voltage value acquisition process receives a reset signal as a command to repeat generation of the output waveform from the initial state when a random signal is detected. It outputs to the said waveform generation process, It is characterized by the above-mentioned.

  According to the present invention, since the random number value is determined based on the voltage value in the voltage waveform in which the voltage monotonously increases with time within a predetermined period from the event that occurs independently based on the physical phenomenon, the measuring means using the counter It becomes unnecessary. For this reason, it is possible to generate a random value without being affected by the counter resolution, and it is possible to generate a highly accurate random value.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram for explaining a case where a voltage value at the time of generating each sampling pulse is measured using a sawtooth wave from a sawtooth wave generation circuit for generating a random number. Sampling pulses are indicated by a, b, c, d. Of course, only random pulses generated from thermal noise sources, etc., are filtered out of a large number of random pulses, for example, those with absolute magnitudes ignoring positive and negative signs greater than a predetermined threshold. What remains is the sampling pulse shown in the figure. The present invention does not employ a conventional method of measuring the number of clocks between each sampling pulse. Specifically, as shown in FIG. 1, a sampling pulse (for example, a) is set to the start time of the sawtooth wave period (at this time, the voltage value is set to 0), and the next sampling is performed. The voltage value (V _b ) at the time of detecting the pulse (b) is measured. The voltage value in the sawtooth wave generation circuit can be converted into a random number value by performing A / D conversion or the like and digitizing the voltage value. In the case of FIG. 1, the voltage value (V _b ) can be obtained as a random value at the timing when the sampling pulse (b) occurs. Incidentally, the horizontal axis represents time of the sawtooth wave shown in FIG. 1, the vertical axis indicates the voltage value.

Further, in order to measure the voltage value at the timing when the next sampling pulse is generated, the voltage value is returned to 0 when the voltage value (V _b ) of the sampling pulse ( _b ) is measured, and then the sampling pulse c Is detected, the voltage value (V _c ) at that time is measured. Then, this voltage value (V _c ) can be obtained as a random value. If this process is repeated each time a sampling pulse is detected, it is possible to measure the voltage value when the sampling pulse is generated without using a counter.

  The sawtooth wave generation circuit resets the voltage increase of the sawtooth wave at a constant cycle and repeats the voltage increase with time until it reaches a predetermined maximum voltage value again to form a continuous sawtooth wave. For example, as shown in FIG. 1, the time interval between the sampling pulses ef may be longer than the period of the sawtooth wave. In this case, the predetermined sawtooth wave generation circuit may perform reset processing after reaching a predetermined maximum voltage value and detect the next sampling pulse (f) in the second and subsequent cycles. The period of the sawtooth wave may be set to an appropriate value according to the sampling pulse generation rate from the type of the thermal noise generation source, etc., and preferably, for example, the average number of occurrences of the sampling pulse per unit time Set a larger value. It is also possible to set an appropriate period in accordance with the resolution of the A / D converter described later, and as a result, it is desirable to reliably detect the sampling pulse.

  FIG. 2 shows an embodiment of a random number generator 100 that obtains random values from voltage values based on the use of the sawtooth wave. The random number generation device 100 includes a random pulse generation circuit 1, a voltage value acquisition circuit 2, a sawtooth wave generation circuit 3, an A / D converter 4, a random value storage circuit 5, and an input / output circuit 6. The random number generation device 100 is an example for explaining the contents of the present invention, and is not limited thereto, and includes changes in a range that does not change or depart from the technical idea of the present invention.

  The random pulse generation circuit 1 is a circuit that generates random pulses from a thermal sound source such as thermal noise. Although not explicitly shown in FIG. 2, the random pulse generation circuit 1 has a filtering function for generating a sampling pulse by extracting a random pulse having a predetermined size or more from the random pulse, or generating a random pulse. A sampling pulse generation circuit is provided separately from the circuit 1. The sampling pulse obtained from the random pulse generation circuit 1 (or the sampling pulse generation circuit) is input to the voltage value acquisition circuit 2.

  The sawtooth wave generating circuit 3 is a circuit that generates a sawtooth wave as shown in FIG. 1 at a predetermined cycle. Since the time and the voltage value correspond in a proportional relationship within the period, the elapsed time can be identified by measuring the voltage value. The voltage value acquisition circuit 2 acquires the voltage value of the sawtooth wave generated by the sawtooth wave generation circuit 3 when receiving the sampling pulse input from the random pulse generation circuit 1 (or the sampling pulse generation circuit). The voltage value is output to the A / D converter 4. In addition, the voltage value acquisition circuit 2 outputs a reset signal to the sawtooth wave generation circuit 3 at the time of acquiring the voltage value, and in response to this, the sawtooth wave generation circuit 3 generates the sawtooth wave in a cycle. Return to the initial state. That is, even before the predetermined maximum voltage value as the sawtooth wave is reached, the voltage value is reset to 0, and the process returns to the process of generating the sawtooth wave at a predetermined cycle again.

  The A / D converter 4 converts the analog voltage value acquired by the voltage value acquisition circuit 2 into a digital value, determines a random value, and outputs it to the random value storage circuit 5. The random value storage circuit 5 is a circuit that stores each random value output from the A / D converter 4. The input / output circuit 6 is a circuit that reads a random value stored in the random value storage circuit 5 in response to a request from the outside and outputs the random value to the outside.

  According to the random number generation device 100 using the sawtooth wave in the present embodiment, the measured voltage value can be directly used as a random number value. Therefore, the counter means for measuring the time interval (t) between the sampling pulses is provided. It can be lost. Moreover, since the data output from the voltage value acquisition circuit 2 is an analog value, it has high affinity with an exponential distribution in which the time interval (t) between sampling pulses is expressed as a continuous function, and outputs an exponential distribution with little distortion. It becomes possible to do. As a result, when performing a test of true randomness based on the result of the exponential distribution, the test process and the test accuracy can be made equivalent as compared with the conventional one while simplifying the circuit.

  Next, a second embodiment of the present invention will be described. The second embodiment uses a logarithmic curve instead of the sawtooth wave described above. FIG. 3 is a diagram illustrating an example of a random number generation device 200 that acquires a voltage value using a logarithmic curve. The logarithmic waveform generation circuit 7 is used instead of the sawtooth wave generation circuit 3 in FIG. 1, and the other circuits are the same as those in FIG. FIG. 4 is a diagram showing a relationship with the sampling pulse when the logarithmic curve is generated by the logarithmic waveform generation circuit 7. As in the case of FIG. 1, each time interval (t) when sampling pulses are generated at a, b, c, d... Is identified by the voltage value in the logarithmic curve.

For example, the voltage at the time when the next sampling pulse (b) is detected by matching a certain sampling pulse (for example, a) with the start point of the period of the logarithmic curve (at this time, the voltage value is set to 0). The value (V ′ _b ) is measured, and this voltage value is acquired as a random value. The voltage value in the logarithmic waveform generation circuit 7 increases with time according to the curvature of a predetermined logarithmic curve. Similarly to the case of the sawtooth wave shown in FIG. 1, the voltage value (V ′ _b ) of the sampling pulse (b) is measured in order to measure the voltage value when the next sampling pulse (c) is generated. Then, the voltage value is returned to 0, the voltage value (V ′ _c ) when the sampling pulse (c) is generated is measured, and this voltage value is acquired as a random value. If this process is repeated each time a sampling pulse is detected, a random value can be obtained without using a counter.

  Instead of directly measuring the time interval between sampling pulses, the basic concept of measuring the voltage value when the next sampling pulse arrives and identifying this voltage value as a random value is the logarithmic curve even for sawtooth waves. But it is the same. Therefore, the effect of the present embodiment is basically the same as the case where the sawtooth wave is used in that the counter means can be eliminated, but the following points are the characteristics particularly when the logarithmic curve is used. Can be mentioned.

  When the voltage value at the time of generating the sampling pulse is measured based on the sawtooth wave, the distribution of the voltage value is expressed as an exponential distribution, which means that the smaller the time interval (t), the higher the frequency of occurrence. means. In other words, a case in which the time interval between sampling pulses is such that the period of the sawtooth wave repeats many times and the next sampling pulse arrives is unlikely to occur. By the way, in the actual use of random numbers, whether or not the generated random numbers are authentic is verified, and only the random numbers that pass the verification are put to practical use. There is an evaluation of regularity. This is that there is no correlation between the i-th value and the j-th value, that is, all values appear uniformly and flatly in a predetermined time width. It can be paraphrased as sex assessment. For this reason, for example, an exclusive OR operation process is performed on a random value obtained by measuring a time interval between pulses generated from a thermal noise source, so that every time interval (t) is equal. It is verified whether it is converted into a flat straight line distribution that means various probabilities. In the case of time interval measurement using a sawtooth wave, an additional calculation process is required to determine whether this flat straight line is formed.

  On the other hand, when the voltage value is measured with a logarithmic curve and the time interval (t) between sampling pulses is calculated, the time interval (t) does not exhibit the characteristics of the exponential distribution as described above. That is, since the time interval (t) is output as a logarithmically compressed value, if the sampling pulse is a genuine random number, a flat straight line distribution is obtained. This eliminates an additional calculation process required for measuring the time interval by using the sawtooth wave, and has an effect that one item of the test as a true random number can be directly evaluated.

  In the first embodiment or the second embodiment, the maximum value of the voltage value generated by the sawtooth wave or the logarithmic curve can be set as the maximum value of the random number value. Easy to set.

  In the above-described embodiment, the example in which the voltage value obtained from the output waveform of the sawtooth wave or the logarithmic curve has been described. However, the present invention is not limited to this. It will be easily understood that the technical idea of the present invention is realized by an arbitrary output waveform in which voltage values are uniquely associated within a certain period.

It is a block block diagram of the apparatus which implement | achieves the random number generation test method of this invention. It is a flowchart which shows the test | inspection operation | movement procedure as one embodiment of the random number generation test method of this invention. It is a flowchart which shows the test | inspection operation | movement procedure as one embodiment of the random number generation test method of this invention, and is related with a timer process especially. It is an image figure which shows the random number generation test method of this invention. It is a figure when the time interval between sampling pulses is measured with the counter.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Random pulse generation circuit 2 Voltage value acquisition circuit 3 Sawtooth wave generation circuit 4 A / D converter 5 Random value storage circuit 6 Input / output circuit 7 Logarithmic curve generation circuit

Claims (6)

Noise sources,
A circuit for generating a random signal from the noise source;
A circuit for generating an output waveform in which the voltage monotonously increases with time in a predetermined cycle;
A voltage value acquisition circuit that detects the occurrence of the random signal and acquires the voltage value at the time when the random signal was generated last time from the output waveform;
A circuit that converts the voltage value into a digital value and outputs the random value;
Random number generator with   The random number generator according to claim 1, wherein the output waveform is a sawtooth wave or a logarithmic curve.   The random number generation device according to claim 1, wherein the circuit that generates the output waveform generates the output waveform from the initial state again in response to the voltage value acquisition circuit detecting a random signal. Processing to detect random signals generated from noise sources;
Waveform generation processing for generating an output waveform in which the voltage monotonously increases with time in a predetermined cycle;
A voltage value acquisition process for acquiring a voltage value at the time of occurrence of the random signal from the output waveform;
A process of A / D converting the voltage value and outputting it as a random value;
Random number generation method including   5. The random number generation method according to claim 4, wherein the output waveform is a sawtooth wave or a logarithmic curve.   6. The random number generation according to claim 4, wherein the voltage value acquisition process outputs a reset signal as a command to repeat generation of the output waveform from an initial state to the waveform generation process when a random signal is detected. Method.

Images