JP4804815B2 - Random pulse generation source, method for generating random number and / or probability using the same, program, and semiconductor device - Google Patents

Description

  The present invention relates to a random pulse generation source provided with an α particle emitter that spontaneously generates a complete random pulse, and such a random pulse generation source is integrally mounted inside to generate a completely random signal. The present invention relates to a semiconductor device such as an IC that can generate an authentication signal, a random number, and a probability, and also relates to a method and program for generating a random number and / or probability from a random pulse.

  Examples of IC tags with improved security include those disclosed in Japanese Patent Publication Nos. 2003-337928 and 2003-16396. The former stores electric power by connecting to the terminal device, creates a one-time password by the CPU inside the IC with this electric power, holds this in the memory, and holds the password when the stored electric power is discharged In the latter, the authentication data is read twice or more from the tag, and a pseudo-random value is used as the authentication data.

Patent Publication 2003-337928 Patent Publication 2003-16396

  However, these IC tags do not include a completely random signal generator or a function for generating such a signal in an IC used for authentication. In this type of authentication IC, a storage circuit is incorporated, and authentication data is received from a signal transmitted from the outside by wireless communication or infrared communication, or created in an IC chip and written in the storage circuit. The data stored in the IC chip on the receiving side can be read from the outside, and the confidentiality is weak and there is a security problem. In addition, when using random number data for IC distinction, it is often a pseudo-random number created by a program on the transmission side, and confidentiality on the transmission side is indispensable, and regularity can be inferred from the random number data There were vulnerabilities, and the number of combinations was also limited, and there was concern that multiple identical identification numbers would be mixed. In addition, when such an IC tag is used for tracking, problems such as the accumulation of personal information on the tracking side occur, which has been an impeding factor for popularization.

  Further, in an authentication device such as an electronic lock, a conventional IC used for authenticating a partner does not have an authentication signal generation function built into the chip itself. For anti-clone (anti-counterfeiting), send an authentication signal from outside by wireless or infrared communication, store this authentication data in the memory element in the chip, and when it matches the stored data of the sender, the other party Was recognized as. Therefore, if the stored data is copied exactly, the sender cannot make a true / false decision. Therefore, as in the case of the IC tag described above, there is a security problem.

  Some ICs have built-in random number generators that use thermal noise or noise. However, thermal noise and noise are affected by environmental changes such as heat and electromagnetic waves, so pseudorandom numbers are handled instead of completely random signals. There was a limit to reliability. In addition, in the method using noise, the random number created on the basis of it is phenomenologically random, but the performance of the random number itself is not theoretically supported except based on various evaluation methods.

The present invention has been developed to solve the above-described conventional problems.
The random pulse generator (generation source) (hereinafter referred to as RPG) according to the present invention emits alpha particles, beta rays or alpha particles that are emitted along with nuclear decay, and an emitter that emits beta rays. a detector that detects α particles and the like, and the emitter is used as a solution, and the solution is dropped onto a detection surface of the detector.

  The RPG according to the present invention also includes an α particle, a beta ray or an α particle emitted along with the decay of the nucleus, an emitter that emits the beta ray, and a detector that detects the emitted α particle and the like. A member formed by depositing the solution or processing into a disk shape by a roll method or the like is configured to face the detector through a predetermined distance.

  Further, a semiconductor device including an IC according to the present invention is provided with a random pulse generation source that spontaneously generates a random pulse, and measures a time interval of a random pulse generated from the RPG or measures a voltage value, By converting to a digital value, a random number is generated and / or a probability is generated.

For example, when the present invention is applied to an authentication device, a random pulse generator (RPG) having an alpha particle emission source that is emitted infinitely by natural decay is incorporated in the IC body, and a completely random signal is obtained from the RPG. This signal is used as an authentication signal. That is, the IC chip itself always generates an authentication signal and rewrites it with new authentication data every time authentication is performed, so that copying does not make sense. At this time, almost infinite combinations of authentication signals are possible without using a program, management of authentication data on the IC chip manufacturing side is unnecessary, and complete security is established on the user side having the IC chip. be able to.
Further, as will be described in detail later, in the present invention, a complete random number and probability can be easily generated from an RPG signal in an IC chip. Make it easy to use.

  A method / program for generating random numbers and / or probabilities according to the present invention comprises: providing an RPG; measuring a time interval of a random pulse generated from the RPG; or measuring a voltage value and converting it to a digital value; Generating random numbers of an exponential distribution having a predetermined bit length, uniform random numbers, and / or generating probabilities from random pulses converted into digital values.

  In the above method, the step of generating a random number of an exponential distribution having a predetermined bit length, a uniform random number, and / or generating a probability from a random pulse is generated from an exponential distribution indicating a probability that the same pulse interval occurs for the random pulse. Find the required number of probabilities and create a uniform random number based on this probability.

  The step of generating a random number of an exponential distribution having a predetermined bit length, a uniform random number, and / or generating a probability from a random pulse obtains the probability from the time interval between t1 and t2 using the exponential distribution. When t2 time is assumed to be infinite time and t1 time is exceeded, it is recognized as a pulse having a predetermined probability.

  The step of generating a random number of an exponential distribution having a predetermined bit length, a uniform random number, and / or generating a probability from a random pulse varies a basic period of measurement when generating the random number from the exponential distribution. Thus, even if the average emission rate of the pulse fluctuates, the same calculation process is performed to realize a stable appearance probability density distribution of exponential distribution random numbers and uniform distribution random numbers.

The step of generating a random number of an exponential distribution having a predetermined bit length, a uniform random number, and / or generating a probability from a random pulse varies a basic period of measurement when generating the random number from the exponential distribution. As a result, the distribution of the exponential distribution random numbers is stabilized, and the generation of uniform distribution random numbers and the generation of probabilities are processed simultaneously.

The step of generating an exponential distribution random number, a uniform random number of a predetermined bit length from random pulses, and / or generating a probability is performed by changing a basic period of measurement when generating a random number from an exponential distribution. It automatically corrects fluctuations in the operating clock, such as the oscillation frequency of hardware such as a microcomputer that counts time.

  The present invention incorporates an original signal source that is not affected by environmental conditions and cannot be controlled by humans into the IC body, and uses a random signal generated by the IC itself as an authentication signal, a random number, and a source signal for probability generation. To do. Thereby, it is possible to construct an authentication system in which human operation is impossible and safety is established.

  In addition, there is no need to store data on the supply side of the authentication system, and the cost can be greatly reduced. The uniform random number generated in the present invention can be used as a probability generator that can handle a complete random number and cannot be illegal.

  In the specific description, an RPG-RFID (Radio Frequency Identification) chip to which a function of a wireless tag is added will be described as an example. However, the present invention is not limited to this, and can of course be applied to other authentication signals, semiconductor devices such as ICs for generating random numbers and probabilities.

  In addition, the present invention can be applied to an IC card, IC tag, PC-connectable device, device-embedded module, etc. incorporating a random pulse generation source (RPG) described below. A semiconductor device configured to generate a random number and / or generate a probability by measuring a time interval of a random pulse generated from the RPG or measuring a voltage value and converting it into a digital value. The present invention can also be applied to embedded IC cards, IC tags, PC-connectable devices, device-embedded modules, and the like.

The IC tag according to the present invention has the following structure. An RPG (random pulse generator) that generates completely random pulses is an improved version of the pulse generator using α particles described in Japanese Patent No. 2926539, which was acquired by the present inventor. As the alpha particle emitter, ²⁴¹ Am, ²⁴⁴ Cm, ²¹⁰ Pb- ²¹⁰ Po, ²¹⁰ Po, etc. which spontaneously collapses are used. More specifically, use all nuclides of the uranium series, all nuclides of the thorium series, all nuclides such as ²¹⁰ Pb- ²¹⁰ Po that form a radiation equilibrium between these nuclides, ²⁴⁴ Cm or ²⁴¹ Am Can do. Note that an RPG that uses a beta ray or a gamma ray to generate a random pulse may be used as long as the IC tag is used in a site where there is a shielded installation space at the use destination of the IC tag.

  Since alpha particles, beta rays, and gamma rays are not affected by the environment such as temperature, pressure, humidity, and electromagnetic waves, they cannot be manually manipulated. This characteristic IC cannot be realized by other methods and is an important factor in ensuring safety. If the IC tag is used in a site where complete safety is not required, an RPG using thermal noise, semiconductor jitter, or the like as a source of random pulses could be used.

  The RPG-RFID chip of the IC tag depends on the RPG and the electronic circuit for enabling the random pulse sent from the RPG to be used as the authentication signal, the circuit for storing the authentication signal, the storage element, and the communication form A circuit for transmitting the authentication signal to the other party, an antenna, a communication element, and the like are incorporated. In the case of an RPG-RFID chip, the power is supplied with RF energy from an antenna. However, if there is an installation space, a terminal may be provided for supply.

  In the authentication procedure of the RPG-RFID chip, first, a random signal (generation time and pulse time interval) from the built-in RPG is written in an internal storage element as an authentication signal of the RPG-RFID chip itself. Further, the stored signal is taken out from the outside using a communication means with the RPG-RFID chip or connected by direct contact and confirmed, and it is confirmed that there is no combination of the same authentication signals in the same lot.

When the RPG-RFID chip to be shipped is used for specific article management and the distribution range is relatively limited in the management of movement information from the time of shipment, the authentication signal for each RPG-RFID chip is sent at the time of shipment. Read through communication, save and use.
This authentication procedure and authentication data storage are performed for each communication, and the authentication data is replaced with a random signal generated by the RPG-RFID chip itself.

  Hereinafter, embodiments of an RPG-RFID chip with a random pulse generator (RPG) according to the present invention will be described in detail with reference to the drawings.

First, a random pulse generator (RPG) that spontaneously generates random pulses will be described.
The RPG is composed of an α-particle emitter, a PN-type semiconductor, a PNP-type semiconductor, a PIN photodiode, a phototransistor, a photodiode, or the like at the central portion of the RPG-RFID chip as shown in FIG. The elements of this pulse generator are selected according to the degree of safety required for the RPG-RFID chip. In FIG. 1, reference numeral 10 denotes an attachment portion of the α particle radiator and the detector, and 100 denotes a base (silicon base).

  The structure around the α particle detector of the RPG-RFID chip to which the α particle emitter is attached is shown in FIG. The release of α particles (He atoms) emitted from the α particle emitter is not affected at all by the environment such as temperature, pressure, electromagnetic waves, etc., and no power source is required and spontaneously released semipermanently according to the half-life. In other words, this system is characterized in that the source signal used for authentication can be a signal source that cannot be manually manipulated. Therefore, it becomes a signal generation source that cannot be changed from the outside. If the mounting part of the IC does not require complete safety, a diode or the like for generating noise in the pulse generator may be used. In FIG. 2, 11 is an α-particle emitter that is deposited or dropped, and 12 is a detector.

  As for the method of incorporating the α particle radiator into the IC, a method of using a product in which the α particle radiator is deposited or manufactured by the roll method and processed into a disk shape, and a solution of the α particle radiator are dropped. (Including injection by an ink jet method).

  The method of incorporating a disk-shaped α particle emitter is mounted directly above the α particle detection diode. As an attachment method, there are a method in which a spacer is attached so that the distance between the detection diode and the α-particle emitter is a predetermined distance, and a method in which the spacer is attached to the sealing plate. The attached structure is shown in FIG. In this figure, 11 is an alpha particle radiator, 12 is a detector, 13 is a sealing material, and 14 is a spacer for setting the distance between the radiator and the detector.

  When a detector is constructed by dropping a solution of an α-particle emitter, the following is performed. (See FIGS. 4A to 4C)

  First, the volume including the circumference of the dropping detector is calculated, and the volume that does not flow out to the outside is obtained. The amount of the solution to be dropped is the same volume as this volume. The concentration of the α-particle emitter is obtained by calculating the number of atoms of the α-particle emitter included in the dropping amount so as to be a predetermined number of pulses, and adjusting the concentration before dropping so as to be the concentration.

  A required amount is sucked into a dropping jig such as the microsyringe 20, and a dropping amount of α-particle emitter solution 11 ′ determined in advance is pushed out and dropped onto the surface of the detector 12 (see FIG. 4A). After dripping, it seals with the epoxy sealing material 15 (refer FIG. 4B), or it seals so that it may seal with the disk 16 for sealing, etc. (refer FIG. 4C). A similar procedure is used when the ink jet method is used to drop the solution by jetting the solution.

The distance between the α-particle radiator and the circuit portion (distance d in FIGS. 1 and 2) is set larger than the range of the α-particle determined by the following formula in order to prevent malfunction due to the α-particles. The range of α particles in air is generally as shown in FIG. The horizontal axis is the distance between the alpha particle emitter and the detector, and the vertical axis is the count value (arbitrary unit). The calculation when ²¹⁰ Pb- ²¹⁰ Po is used for the α particle emitter is shown below.

The relationship between the energy E (MeV) of the α particles and the range R (cm) in the air has been well investigated and is expressed by the Geiger equation.
E = 2.12R ^2/3 (4 <E <7)
R = 0.323E ^3/2 (3 <R <7)
Therefore, the range of Poα particles in the air is

Ａ：固体物質の原子量
ρ：密度
Ｓｉ中におけるα粒子の飛程は、

よって、回路と分離する為にＳｉの遮蔽材（遮蔽壁）を用いる場合は２９μｍ以上とすれば良いことになる。 If the range in air is known, the range Rs (cm) of the α particles in the solid substance can be obtained by the following Bragg-Kreman equation.

A: Atomic weight of solid substance ρ: Range of α particles in density Si

Therefore, in the case where a Si shielding material (shielding wall) is used for separation from the circuit, it may be 29 μm or more.

²¹⁰ Pb- ²¹⁰ Po emits extremely weak beta rays. Unlike alpha particles, beta rays exponentially decrease the number of electrons reached with distance as shown in FIG. The beta ray is an electron itself and does not affect the internal IC circuit at all, but the structure does not leak to the outside so as not to cause unnecessary misunderstanding. The necessary thickness of the shielding material is obtained by the following formula. In FIG. 8, the horizontal axis indicates the thickness of Al, and the vertical axis indicates the number of electrons.
Since the maximum energy Em of the Pb-210 beta ray is 0.061 MeV, the maximum range Rm is calculated from the Willson equation.
Rm = 2400 (Em) ² = 2400 × (0.061) ² = 2400 × 0.003721 = 8.9304 mg / cm ²

よって、ＲＰＧ−ＲＦＩＤチップに必要な図１に示す封止材は33μｍ以上あれば良いことになる。 In addition, from the figure of the isotope handbook (revised 3rd edition, Maruzen, April 1985), at 60 KeV, the range in Al is 6 mg / cm ² . Since the density of Al is 2.7 g / cm ³ ,

Therefore, the sealing material shown in FIG. 1 required for the RPG-RFID chip may be 33 μm or more.

  By the way, the present invention can be easily incorporated in not only an RPG-RFID chip but also a general IC. FIGS. 5A and 5B are layout diagrams for incorporation in a DIP type IC. The position where the alpha particle emitter is incorporated does not need to be in the center as shown in the figure, and can be changed according to the design of the IC circuit. (The same applies to the MM chip) In the figure, reference numeral 30 denotes an attachment portion of the α emitter, 31 denotes a sealing cap or pitch, and 32 denotes a diode chip to which the α emitter is dropped.

FIG. 6 shows an entire block circuit of an example of the RPG-RFID chip according to the present invention. The RPG-RFID chip includes an RPG 40, a waveform shaping amplifier circuit 41, a logic circuit (control calculation) 42, a storage circuit 43, an RF circuit 44, and an antenna 45.
The waveform shaping amplifier circuit 41 is for making the peak waveform output from the RPG 40 a shaped rectangular wave that can be digitally processed. The logic circuit 42 is a processing circuit for handling a random pulse signal as a target signal. The storage circuit 43 is a circuit for storing data after signal processing. The RF circuit 44 is a circuit for performing communication via an antenna 45 with the outside.

  As a method of digitizing and storing the signal from the RPG, as shown in FIG. 9, since the RPG random pulse has a random crest value and pulse interval, the pulse interval is measured or the crest value voltage is measured. It is also possible to perform digital conversion by sample-and-hold and use the measured value or converted value as a random value. A combination of voltage and number of clock pulses can also be used.

  Next, when a pulse generated by RPG is used as an authentication signal, the probability of occurrence of exactly the same pulse interval is obtained.

As an example of calculation, the calculation is made in the case where the average emission rate n = 10 pulses normally used.
The probability of the pulse interval is obtained by the following formula.

１０個のパルス間隔が平均のパルス間隔0.1secで出現したとすると、同じ時間間隔で１０個のパルスが次に出現する確率としては、次の表の計算結果となる。 When the average RPG release rate n = 10, it is assumed that an average number of 10 pulses per second is used as the authentication signal.

Assuming that 10 pulse intervals appear at an average pulse interval of 0.1 sec, the probability of the next occurrence of 10 pulses at the same time interval is the calculation result in the following table.

上記表の計算結果から同じパルス間隔のパルス列10個が出現する確率は、１/2.97E＋21となる。

From the calculation results in the above table, the probability that 10 pulse trains having the same pulse interval will appear is 1 / 2.97E + 21.

Therefore, in the method using this method, the probability of the same combination with at most 10 pulses can be realized as 1 / 2.97E + 21 or less. This probability is the probability of the same pulse generator, and each pulse generator has a different n value. Therefore, the probability that different pulse generators have the same pulse train is (1 / 3E + 21) ² or less.

  The number of pulses to be used can be selected according to the purpose of use of an IC that uses this method for authentication, and safety can be improved.

Next, a method for generating uniform random numbers will be described.
The probability of the pulse appearing in the time interval t1 to t2 is obtained by the above equation (1).
Transform this equation,
t ₂ = -1 / n * log (e ^-nt1 -p) ----------- (2)

Here, since it is t2> t1, the equation replaced by _{t k = t 2, t k} -1 = t 1, successively t ₁ from t ₀ = _0, t _2, by determining the ...., A threshold of time width (Δt) having a uniform appearance probability of 1 / p can be obtained.

t _{k + 1} ＝ −1 / n ＊ log (e ^-ntk −p) ----------- (3) where t ₀ = 0
In this equation (3), each t _k value can be uniquely determined by fixing n and p. When the pulse interval time measurement is performed with a 16-bit length (0-65535), considering the number of significant digits after conversion, the data length to be replaced by the above threshold value must be about 8 bits long, and the effect of errors will be affected at short time intervals It cannot be ignored. Accordingly, p is set to 256, and the 0-65535 (16-bit length) section is converted to the 0-255 (8-bit length) section by the threshold value. When 16 bits are required as the bit length, it can be generated by connecting two pieces of 8-bit data.

  In this method, when the n value increases, the appearance time interval becomes relatively small, and an error becomes a problem when the time interval measurement resolution is constant. Conversely, when the n value decreases, the time interval becomes relatively large, and when the time interval measurement resolution is constant, a rounding error (overflow) occurs in the upper numerical value.

Therefore, the fluctuation of the n value is dealt with by changing the basic period of time measurement. In other words, when the measured n value is small, the basic period of measurement is increased. On the other hand, when the n value is large, by decreasing the basic period of measurement, the distribution of numerical data as a measurement result is always constant. The arithmetic processing circuit can be simplified. In addition, since fluctuations in the basic period of time measurement depend only on the average number of releases per unit time, fluctuations in the reference operating frequency of the arithmetic circuit are also corrected within a range where the operation speed of the arithmetic circuit is in time from the viewpoint of processing capability . Therefore, the accuracy of the frequency of the oscillation circuit that generates the clock of the arithmetic circuit is not required. In particular,
t _{k + 1} = −1 / n * log (e ^-ntk −p) --------- (4) where t ₀ = 0
In the equation, when the fluctuation component of n is m and n = n × m,
t _{k + 1} = −1 / nm * log (e ^-nmtk −p) ---------- (5) where t ₀ = 0
It becomes. Here, by changing the basic period of measurement time according to m (tk = tl / m)
t _{l + 1} = −1 / n * log (e ^-ntl −p) --------- (6) where t ₀ = 0
Thus, the basic arithmetic expression can be applied as it is.

  If the operating frequency used in the arithmetic circuit is about 25 MHz, the basic cycle clock is 40 nS. If the n value is 10, the average pulse interval is 100 mS, and the measurement basic period is 100 μS. The frequency dividing number for generating the period is 100 μS / 40 nS = 2500 counts. Therefore, the minimum value (resolution) of the fluctuation value m is 0.4 / 100 = 0.004 (0.4%), and there is no problem with the resolution of the fluctuation of the measurement reference time with respect to the fluctuation of the n value.

In this method, the measurement time required to obtain 256 values increases as n decreases. For example, the 255th threshold value is 1.386 seconds when n = 4, 1.848 seconds when n = 3, and n = 2. Sometimes it becomes 2.773 seconds and you have to count up to this time. The time width longer than that can be rounded to the maximum value (65535) of the time measurement data. (The appearance frequency of the maximum data 255 does not change even when rounding is performed.) Therefore, the max value of the measurement varies depending on the n value.
Next, a method for creating a probability is described.

The probability of using natural decay can be calculated by equation (1). (1)
p = e ^-nt1 -e ^-nt2 --------------- (7)
Therefore, it is only necessary to determine t ₁ and t ₂ that satisfy the probability value to be generated and determine whether or not there is a pulse between them. Here, when the t ₂ and infinity, two items of e ^-NT2 is negligible 0, measured time width can determined per by comparing t _1. By setting t ₂ to infinity, a region where the frequency of occurrence of e ^-nt is small is used, so that the resolution can be widened. Therefore, the rounding error during probability calculation can be minimized.

Formula when t ₂ is infinite
p = e ^-nt1 ------------------------ (8)
Taking the logarithm of respect to both sides, and seek t ₁
t ₁ = -log (p) / n -------------------- (9)
Thus, the value of t ₁ can be obtained from the set probability value (p) and the number of pulses per second (n).

  Even in this case, the required measurement time varies due to the fluctuation of the n value, and particularly when the probability value is small, it is necessary to enable long-time measurement. Therefore, the problem of counter overflow and insufficient accuracy can be solved by changing the measurement basic period in the same manner as the calculation of the uniform random number. However, this method can handle very small probability values, but if the probability value is small, the maximum value of measurement may increase and the measurement time may increase. However, the problem of response time can be avoided by providing a mechanism for storing past measurement values in the arithmetic circuit unit.

The above equation (9), t ₁ = −log (p) / n, may be calculated when a probability value is set from the outside, and no calculation occurs during measurement. Therefore, even if the calculation time takes some time, the influence on the actual operation is very small. In order to flexibly deal with the set probability value and select various probability values by external input, the above logarithmic calculation can be calculated by a microcomputer without using a table for omitting the calculation. I can do it.

Next, the setting of the reference measurement time for the n value for changing the basic period will be described.
p = e ^−nt1 −e ^−nt2 --------------------- (7)
T ₀ is 0, and the time interval of the maximum threshold value (255) is p = ^255/256 = 1−e ^−nt255
Than,
t ₂₅₅ =-(1 / n) * log (1/256) = 5.545177 / n ------------ (10)
Can be obtained.

If we set the 255th threshold + 5% as the time when the 16-bit timer overflows, with some margin, the maximum time interval that can be counted by the 16-bit timer is
t _max = 5.545177 * 1.05 / n = 5.822436 / n --------------- (11)
From this, the cycle of the timer count reference clock is
t _clk = (5.822436 / 65536) /n=88.84333 μS / n.
If the basic operating frequency of the arithmetic circuit is 25 MHz, the system clock cycle is
t _sys = 40 nS
And the value of the prescaler for generating the reference clock t _{clk of the} timer is
t _pres = (88843.33 / 40.816) /n=2176.7/n.

As described above, equation (4) can be used to obtain a threshold value for uniform random number generation by varying the timer count reference period in accordance with the value of n.
t _{k + 1} = −1 / n * log (e ^-ntk −p) where t ₀ = 0
In the above equation, when each t _k is calculated when n = 5.822436, threshold values from 1 to 255 can be calculated, as shown in the following table.

この表のデータを閾値として、計測データがどの閾値間に入るかを求めれば、その結果が８ビット長の一様乱数となる。

If the data in this table is used as a threshold value to determine which threshold value the measurement data falls between, the result is a uniform random number with a length of 8 bits.

  There are two methods for determining which threshold value the measurement data is between. One is a method of comparing magnitudes by binary search after measurement is completed. The other is a method of checking which area the current measurement value is in during the measurement process. Each is shown below.

Binary search method This conversion is 16-bit measurement data, and the converted data is 8 bits long, which is a perfect condition for binary search, and the program can be greatly simplified. The following are examples of algorithms specialized for this project.
BitIndicator = 10000000b;
Result = BitIndicator;
Repeat
If MeasuredValue <DataTable (Result) then Result = Result xor BitIndicator;
BitIndicator = RightLotate (BitIndicator, 1bit);
Result = Result xor BitIndicator;
Until BitIndicator. Bit7 = "1";
In this conversion, 16-bit comparison, bit processing (xor), and data table reference are performed eight times.

Judgment in the measurement process The threshold value table has a numerical value increasing in the order of 0 to 255. Therefore, each time the pulse interval is measured, comparison with the data table makes the process very simple, and the conversion to uniform randomization is completed when the measurement is completed (next pulse). The response will be very good. Unlike the method using the binary search described above, a certain period of time is not restricted by the arithmetic processing. When a microcomputer is used for arithmetic processing, it may work effectively. The specific algorithm is as follows.
In this method, it is assumed that data 1-255 of the above table is stored in 0-254.
<Initial processing>
Result = 0; top of table <every prescaler overflow>
if MeasuredValue ＝ DataTable (Result) then Result ＝ Result + 1;
In this way, the processing is extremely simple and the threshold value search processing is performed in a distributed manner, so the load on the processing capability of the arithmetic processing unit is also very small.

  Next, an example of a processing procedure in the device / method for generating uniform random numbers and probabilities according to the present invention will be described in more detail.

  The method described here has three modes, a normal (stand-alone) mode, a demo mode, and a multi-module mode, and has the following features.

  The normal mode is a one-to-one connection mode with the host, and outputs an exponential distributed random number sequence and a uniform random number sequence for each pulse measurement, and outputs a probability determination result in response to an input of a probability case read command. The demo mode is a mode prepared for demonstration, and outputs an exponential distribution random number sequence, a uniform distribution random number sequence, and a probability determination result for each pulse measurement. Multi-mode allows multiple modules to be connected, increasing the speed of random number generation, and exponential distribution random number sequence, uniform distribution random number sequence, and probability judgment result are output according to command input, respectively .

  10A to 10C are flowcharts showing an example of an operation (algorithm) when the device / method of the present invention is processed using a microcomputer.

First, as shown in FIG. 10A, initialization operations such as clearing of RAM, initial setting of various registers, and calling of default values of the flash memory are performed, and the process jumps to the main routine.

  In the main routine, first, it is checked whether or not there is new measurement time data that has not been processed (step S110). If there is, the data is called on the condition that it is not the multi mode (step S120) (S130). ). Next, the prescaler value of the timer for measuring the time between pulses (pulse time interval, pulse interval) is set (step S140), and the timer is cleared (step S150). When an RPG pulse is detected (step S160), the timer is started (step S170), and each time an RPG pulse is detected (step S180), a measured value is output from the timer (step S190) and stored as an exponential distribution random value. (Step S200). Next, it is converted into a uniformly distributed random number (step S210), and the probability (hit and miss) is determined (step S220). Next, it is determined whether or not the operation mode is the demo mode (step S230). If the operation mode is the demo mode, the exponential distribution random number is transmitted unconditionally (regardless of whether or not the random number output is selected) to the host (step S260). Then, a uniformly distributed random number is transmitted (step S270), and a probability determination value is transmitted (step 280).

  When not in the demo mode (that is, in the normal mode), transmission is performed when the output of the exponential distribution random number is selected (step S240), and when the output of the uniform distribution random number is selected (step S250) Send to.

  On the other hand, in the measurement of the CPS value (number of pulses), as shown in FIG. 10B, first, a default value is set (step S310), and when an RPG pulse is detected (step S320), a CPS measurement time setting timer is initialized ( Step S330) and a measurement time are set (Step S340). Next, the pulse number counter is cleared (step S350), and each time RPG is detected (step S360), the measured values are integrated (step S370). This is repeated until the set measurement time (for example, 64 seconds, 1 hour, etc.) is reached. When the measurement set time is reached (step S380), the CPS value of the set measurement time is calculated (step S390), and the pulse time interval (pulse interval) measurement reference time is calculated (step S400). Next, new CPS data is written to the flash memory (step S410), and measurement is restarted (step S420).

  Note that the measurement processes shown in FIGS. 10A and 10B are performed simultaneously in parallel, and the time between pulses (pulse interval) is measured between all pulses.

  In the embodiment of the present invention, as shown in FIG. 10C, when serial data (command) is received (step S500), the command processing routine is started only when there is new data (step S510).

  The command processing routine includes an operation condition setting command process, an internal status read command process, a probability setting command process, an operation mode command process, as described in a program that describes details of a process procedure (algorithm) described later. The module address and designated address are compared.

The program shown below describes the processing algorithm inside the microcomputer when the method for generating uniform random numbers and probabilities according to the present invention is processed using the microcomputer.

<Processing algorithm inside microcomputer>
Microcomputer: C8051F330
Operating frequency: 24.5MHz
System clock: 1/1

1. Port assignment, interrupt function

P0.0: Pulse input
P0.4: Serial output
P0.5: Serial input
P1.3: LED output

External interrupt 0: Pulse input detection Timer 1: UART transfer rate generation Timer 2: Used to measure time between pulses
: Change the prescaler value by CPS Timer 3: For creating a fixed time (CPS measurement time) Serial interrupt: Used for data reception

2. Timer

Timer 0: Not used
Timer 1: RS232C Baud Rate Generator
Timer1 Condition {
pre-scale = 1/1
8 bit Reload Mode
Reload Value = 96h
}
Timer 2: Prescaler for pulse count
Timer2 Condition {
16 bit Reload Mode
pre-scale = 1/1
Reload Value = t _pres
}
Timer 3: Timer for CPS count
Timer3 Condition {
16 bit Reload mode
Pre-scale = 1/12
Reload Value = 40833; Overflow = 20 mS
}
PCA: (1) Watch Dog Timer

3. memory

SerialBuffer (0..15) as M [8]; Serial receive buffer
IntervalTime (0..3) as M [17]; Buffer for storing pulse measurement time
SerialWPointer as M [8]; Pointer that indicates the write position when serial receive data is written to the receive buffer in the interrupt processing routine.
SerialRPointer as M [8]; Pointer to the position of the next serial data to be processed in the serial receive buffer in the main routine.
IntervalWPointer as M [8]; Pointer to the position where the pulse measurement time is written next in the interrupt processing routine.
IntervalRPointer as M [8]; Pointer to the position in the buffer of the pulse measurement time to be processed next in the main routine.

IntervalCounter as M [17]; Counter that counts the pulse time interval in the interrupt processing routine.
CPSSecCounter as M [8]; 1 second count (25 times x 20mS)
CPSTimeCounter as M [8]; 64 seconds count
CPSCounter as M [16]; CPS counter
CPS as M [16]; CPS value
Prob (0..7) as M [16]; Probability setting
PValue (0..7) as M [16]; Threshold after probability logarithm operation
PSetting as M [16]; Probability setting input value
PArg as M [8]; Probability setting argument
BEXP as M [8]; Exponent part of operand in logarithmic calculation
CycleCounter as M [8]; Used in the arithmetic unit, counter for the number of repetitions
BitIndicator as M [8]; Used in division operation
NOPTimeCount as M [8]; Counter for measuring the number of pulses per hour
RandomEXP as M [24]; Final result of exponential random numbers
UniformedData as M [8]; Final result for uniformly distributed random numbers
CPSForceData as M [16]; CPS forced setting value
CPSMeasureTime as M [8]; Set value of CPS measurement time
NumberOfPulse as M [24]; Number of pulses per hour
CPSReadOut as M [16]; Area for passing CPS measurement value from interrupt to main
CPSACTimeCount as M [8]; Count of CPS integration count
CPSAccumulation as M [24]; CPS measurement time readout integration buffer
; Accumulate the measured value for 64 seconds until the set measurement time is reached
LastCPSAccumulation as M [24]; Past memory of CPS measurement time
NOPCountBuffer as M [24]; 1 hour pulse count buffer
NumberOfNewData as M [8]; Number of serial received raw data

; Bit area

InternalFlag as M [8] = {
CPSDetected as M [1]; CPS count end flag
LogASignMinus as M [1]; Sign of Log (A)
CountRestartF as M [1]; Count restart
DataFlip as M [1]; Uniform random number wrapping
NOPCaptureF as M [1]; 1 hour measurement completed
SerialReceived as M [1]; Serial reception flag
CPSForceF as M [1]; bit7: CPS forced setting}

OperationMode as M [8] = {; Operation mode
UniformDataOutF as M [1]; Uniformly distributed random number output
EXPDataOutF as M [1]; Exponential distribution random number output
DemoModeF as M [1]; Demo mode
MultiModeF as M [1]; Multi mode
}

StatusFlag as M [8] = {; Error status
CPSTooLittleF as M [1]; No CPS detected
CPSTooManyF as M [1]; CPS excess detection
IntervalTimeOF as M [1]; Measurement value buffer overflow
SerialBufferOF as M [1]; Serial buffer overflow}

ProbFlag as M [8] = {; Probability judgment (win / fail)
PJudge0 as M [1]; Judgment of probability number 0
PJudge1 as M [1]; Judgment of probability number 1
PJudge2 as M [1]; Judgment of probability number 2
PJudge3 as M [1]; Judgment of probability number 3
PJudge4 as M [1]; Judgment of probability number 4
PJudge5 as M [1]; Judgment of probability number 5
PJudge6 as M [1]; Judgment of probability number 6
PJudge7 as M [1]; Judgment of probability number 7}

InternalFlag1 as M [8] = {; Internal flag

AddressExist as M [1]; Multi mode with address
AddressSelected as M [1]; When address is selected in multi-mode}

; Constant

CPSMAX as # 100 * 64/100 pulses or more per second
CPSMIN as # 1 * 64/1 pulse or less per second
ConstEXPData as # 01h
ConstUniformedData as # 02h
ConstSecCount as # 50
ConstTimeCount as # 64
ConstNOPCount as # 56
SAck as # 05h
SNak as # 50h
ConstSPFilter as # 10101111b
ConstITFilter as # 11001111b

4). Initial processing

/ Initial processing /
R0 = 00h; Clear RAM
Repeat
RAM (R0) = 00h
R0 + = 1
until R0 = 0;
SerialWPointer = #SerialBuffer; Set the buffer start address
SerialRPointer = #SerialBuffer;
IntervalWPointer = #IntervalTime; Set the buffer start address
IntervalRPointer = #IntervalTime;
NOPTimeCount = #ConstNOPCount;
InitializeSFR; 8051SFR setting
LoadFlashData; Reads data from the flash unit. The default value is read for the first time
CalcTPreScaler; Calculation and setting of pull scale value of Timer2
CountRestart; Memory initialization and interrupt enable for measurement restart
jump to Main;

/ InitializeSFR /
p0.4 = OutputMode;
P1.3 = OutputMode;
Timer1 = Timer1Condition;
Timer2 = Timer2Condition;
Timer3 = Timer3Condition;
PCA = PCACondition;
Timer2INTEnable = true;
Timer3INTEnable = true;

/ LoadFlashData /; Read data written to flash memory
; The default value is stored in the flash data section
; Write and put.
CPS = FlashCPS;
CPSMeasureTime = FlashCPSMT;
NumberOfPulse = FlashNofP;
StatusFlag = FlashStatus;
OperationMode = FlashOPMode;
LastCPSAccumulation = FlashLastCPSAccumulation;
PROB (o..7) = FlashPROB (0..7);
Pvalue (0..7) = FlashPT (0..7);

5. Main routine

ConstEXPData as “01h”;
ConstUniformedData as “02h”;

Repeat
if IntervalWPointer <> IntervalRPointer then / New unprocessed measurement
Is there time data?
if not (MultiModeF) then / Do not process in multimode
ReadIntervalTime; / Read new measurement time
ConvertToUniform; / Conversion to uniformly distributed random numbers
CalcProb; / Probability (per / out) judge
if DemoMode
then / Unconditionally send in demo mode
SendRandomEXP; / Send exponential random numbers
SendRandomUniform; / Send uniformly distributed random numbers
SendPJudge; / Send probability judgment value
else / In normal mode, send only when random number output is selected
if EXPDataOutF then SendRAndomEXP;
if UniformDataOutF then SendRAndomUniform;
endif
endif
endif
if CPSDetected then
CPSDetected = false; Processing when measurement of CPS value (64 seconds) ends
AccumulateCPS; / until the measurement time is set
/ Accumulating CPS values for 64 seconds
CPSACTimeCount-= 1;
if CPSACTimeCount = 0 then / When the measurement set time is reached
CalcCPS; / CPS value converted to average value for 16 seconds
if not (CPSForceF) then CalcTPeScaler; / Calculate prescaler value
WriteToFlash; / Write new CPS data to flash
CountRestart; / Restart measurement
endif;
endif
if NumberOfNewData <> 0 then CommandHandling; / Receive serial data
/ Start command processing routine only when there is new data
endif
until forever;

6). Interrupt handling routine

/ External interrupt (INT0): Pulse input interrupt (p0), rising edge detection /

if CountRestartF
then
CountRestartF = false; / Only start the measurement timer when measurement restarts
Start Timer2;
else
IntervalTime (IntervalWPointer) = IntervalCounter;
/ Write measurement results to measurement data buffer
IntervalWPointer = (IntervalWPointer + 1) and #ConstITFiler;
/ Update pointer
if IntervalWPointer = IntervalRPointer then IntervalTimeOF = true;
/ Buffer overflow check
IntervalCounter = 0;
CPSCounter + = 1; / CPS counter count up
endif
Return from Int;

/ Serial interrupt /
if SerialReceiveINTF then / Process only at receive interrupt. Ignore transmission interrupts
SerialReceiveINTF = false;
SerialBuffer (SerialInputPointer) = SerialReceivedData;
/ Write serial receive data to receive buffer
SerialWPointer = (SerialWPointer + 1) and #ConstSPFilter;
/ Update serial write pointer
NumberOfNewData + = 1;
if NumberOfNewData> = 32 then SerialBufferOF = true;
/ Buffer overflow check
endif
Return from Int;

/ Timer2 interrupt /

IntervalCounter + = 1; / Count up the time measurement counter between pulses
Return from Int;

/ Timer3 interrupt /
CPSSecCounter-= 1; / 1 second count
if CPSSecCounter = 0
then
CPSSecCounter = #ConstSecCount;
CPSTimeCounter-= 1; / 64 seconds count
if CPSTimeCounter = 0
then
CPSTimeCounter = #ConstTimeCount;
CPSReadOut = CPSCounter;
/ CPS count value read out after 64 seconds
CPSCounter = 0;
CPSDetected = true;
/ CPS detection flag set
end if
endif
Return form Int

7). subroutine

/ CountRestart / Restart pulse time measurement.
/ When CPS is updated, when flash is written, initial state
CPSSecCounter = #ConstSecCounter;
CPSTimeCounter = #ConstTimeCount;
CPSACTimeCount = CPSMeasureTime;
CPSCounter = 0;
CPSAccumulation = 0;
IntervalCounter = 0;
CountRestartF = true;
CPSDetected = false;

/ ReadIntervalTime / Reads the measurement time from the time measurement value buffer
RandomEXP = IntervalTime (IntervalRPointer);
IntervalReadPointer = (IntervalReadPointer + 4) and #ConstITFilter;

/ AccumulateCPS / Accumulate CPS measurement value
/ Integrate the measured values every 64 seconds to obtain the CSP value for the set measurement time
/ CSP measurement value every hour is also obtained here
/ The measured value every hour is obtained by integrating the measured value every 64 seconds 56 times and obtaining the accumulated value of 3584 seconds.
/ The next 64 seconds measurement is ¼ and added. (3584 + 64 * 0.25) = 3600
CPSAccumulation = CPSAccumulation + CPSReadOut;
/ Accumulation of CPS value Accumulation of number of pulses for 1 hour from here
if NOPCaptureF then CPSReadOut = CPSReadOut / 4;
NOPCountBuffer = NOPCountBuffer + CPSReadOut;
if NOPCaptureF
then / 1 hour measurement completed
NumberOfPulse = NOPCountBuffer;
NOPCaptureF = false;
if DemoModeF then SendNOP;
WriteToFlash;
CountRestart;
else
NOPTimeCount-= 1;
if NOPTimeCount = 0 then
NOPCaptureF = true;
NOPTimeCount = #ConstNOPCount;
endif
endif

/ CalcCPS / Calculation of CPS value for 64 seconds
CPSAccumulation = (CPSAccumulation + LastCPSAccumulation) / 2;
LastCPSAccumulation = CPSAccumulation;
Breg = CPSMeasureTime;
if Breg.bit0 <> “1” then
repeat
CPSAccumulation = CPSAccumulation / 2;
RightShift (b, 1bit);
until Breg.bit0 = “1”;
endif;
CheckCPSMAXMIN; Check the maximum and minimum CPS values
if not (CPSTooLittleF or CPSTooManyF) then CPS = CPSAccumulation;
if DemoModeF then SendCPS;
WriteToFlash:
CountRestart;

/ CheckCPSMAXMIN /
/ Check whether the CPS value exceeds the specified range
/ If the CPS value is too large, the Timer2 interrupt cycle will be very short,
/ Be careful as the interrupt will always be applied and processing will not be performed.
Breg = # 0;
If CPSAccumulation> = #CPSMAX then Breg.bit1 = “1”;
If CPSAccumulation <#CPSMIN then Breg.bit0 = “1”;
if StatusFlag <> Breg then
StatusFlag = Breg;
SendStatus;
endif;

/ SendRandomEXP / Serial output of exponential random numbers
SerialOutput = #ConstEXPData;
if RandomEXP [23..16] = “0”
then
SerialOutput = RandonEXP;
Else
SerialOutput = #FFFFh;
endif

/ SendRandomUniform / Uniformly distributed random number serial output
SerialOutput = #ConstUniformData;
SerialOutput = UniformedData;

/ SendCPS / CPS value serial output
/ When reading internal status and CPS value changes in demo mode
SerialOutput = # 11h;
SerialOutput = CPS;

/ SendCPSMT / CPS measurement time serial output
/ When reading internal status
SerialOutput = # 12h;
SerialOutput = CPSMeasureTime;

/ SendNOP / Serial output of the number of pulses per hour
/ When the internal status is read and the CPS value changes per hour in demo mode
SerialOutput = # 13h;
SerialOutput = NumberOfPulse;

/ SendStatus / Serial output of status information
/ When internal status (CPS error) changes and when reading internal status
SerialOutput = # 14h;
SerialOutput = StatusFlag;

/ SendVersion / Serial output of version information
/ When reading internal status
SerialOutput = # 15h;
i = 0;
repeat
SerialOutput = FlashVer (i);
i + = 1;
until FlashVer (i) = “FFh”;

/ SendProbSetting / Serial output of probability setting value
/ When reading internal status
SerialOutput = SerialCommand;
SerialOutput = PROB (Command and “FFh”);

/ SendPJudge / Probability judgment value serial output
/ When a probabilistic read command is received, the internal status is displayed for each pulse input in the demo mode.
When a read command for / is received
SerialOutput = # 28h;
SerialOutput = ProbFlag;

/ CalcProb / Judges / declines from exponential random number
i = 0;
ProbFlag = “0”;
Repeat
if PValue (i) <= RandomEXP then ProbFlag [bit i] = “1”;
i + = 1;
until i = “8”;

/ CommandHandling /. Serial command processing
if (SerialBuffer (SerialRPointer) and “F0h”) = “00h” then / Operating condition setting command
if SerialBuffer (SerialRPointer) = “00h” then
/ NAK is returned when an unassigned code is received
SendNAK;
exit;
endif
if SerialBuffer (SerialRPointer) = “01h” then / CPS value forced setting?
if NumberOfNewData <3 then exit;
/ If no operand (setting data) has been received, do nothing.
RenewSRP;
if (MultiModeF and AddressExist and not (AddressSelected)
then
/ Ignore command if address does not match in multi mode
RenewSRP;
SendNal;
Exit
else
RenewSRP;
CPSForceData = SerialBuffer (SerialRPointer);
/ Read CPS compulsory value
CPSForceF = true;
/ Set CPS forced flag
CalcTPrescaler;
/ Prescaler value calculation (with forced data)
RenewSRP;
SendAck;
endif
if SerialBuffer (SerialRPointer) = “02h” then / CPS forced release
if (MultiModeF and AddressExist and not (AddressSelected)
then
/ Ignore command if address does not match in multi mode
SendAck;
exit
else
CPSForceF = false;
/ Reset CPS forced flag
SendAck;
Exit
endif
if SerialBuffer (SerialRPointer) = “03h” then / CPS measurement time setting
if NumberOfNewData <2 then exit;
/ If no operand (setting data) has been received, do nothing.
RenewSRP;
if (MultiModeF and AddressExist and not (AddressSelected)
then
/ Ignore command if address does not match in multi mode
SendAck;
Exit
else
/ Read and set CSP measurement time
CPSMeasureTime = SerialBuffer (SerialRPointer);
WriteToFlash;
CountRestart;
SendAck;
Exit
endif
if SerialBuffer (SerialRPointer) = “0Fh” / self-destruct
then
if NumberOfNewData <5 then exit;
/ If no operand (setting data) has been received, do nothing.
if (MultiModeF and AddressExist and not (AddressSelected)
then
/ Ignore command if address does not match in multi mode
SerialRPointer + = 4;
NumberOfNewData-= 4;
RenewSRP;
exit
else
RenewSRP;
If SerialBuffer (SerialRPointer) = “86h, 51h, 29h, 78h”
then
/ If regular data string is received, erase flash data
Destroy;
exit;
else
/ Otherwise, ignore the command and return Nak
SerialRPointer + = 3;
NumberOfNewData-= 3;
SendNAk;
exit
endif;
endif
else
SendNAK;
exit;
endif
endif
if (SerialBuffer (SerialRPointer) and “F0h”) = “10h” then / read internal status
if SerialBuffer (SerialRPointer) = (“10h” or “16h” or “17h”) then
/ Returns Nak for invalid data
SendNak;
exit;
endif;
ACC = SerialBuffer (SerialRPointer
Sirent;
if (not (MultiModeF) or (AddressExist and AddressSelected))
then
/ In stand-alone mode and when the addresses match in multi-mode
if ACC = “11h” then / CPS return
SendCPS;
exit;
endif;
if ACC = “12h” then / CSP measurement time return
SendCPSMT;
exit;
endif;
if ACC = “13h” then / 1 hour CPS value return
SendNOP;
exit;
endif;
if ACC = “14h” then / Return status data
SendStatus;
exit;
endif;
if ACC = “15h” then / version return
SendVersion;
exit;
endif;
SendProbSetting; / Return probability setting value
exit;
else
exit;
endif
endif
if (SerialBuffer (SerialRPointer) and “F0h”) = “20h” then / probability setting command
if SerialBuffer (SerialRPointer) = “28h” then
/ Reading probability judgment value
if (not (MultiModeF) or (AddressExist and AddressSelected))
/ then SendPJudge;
/ Address match in stand-alone mode and multi-mode
/ Only return probability judgment value
Sirent;
exit;
endif
if SerialBuffer (SerialRPointer) <“28h” then
if NumberOfNewData <3 then exit;
/ If no operand (setting data) has been received, do nothing.
if (MultiModeF and AddressExist and not (AddressSelected)
then
/ Ignore command if address does not match in multi mode
SerialRPointer + = 2;
NumberOfNewData-= 2;
Sirent;
exit;
else
/ Set probability set value
PArg = SerialBuffer (SerialRPointer) and “00000111b”;
/ Calculation of probability number
RenewSRP;
PROB (PArg) = SerialBuffer (SerialRPointer);
/ Write probability set value
RenewSRP;
CalcPThreshold;
/ Probability threshold calculation
WriteToFlash;
SendAck;
CountRestart;
exit;
endif;
endif
SendNak;
exit;
endif
if (SerialBuffer (SerialRPointer) and “F0h”) = “40h” then / operation mode command
if SerialBuffer (SerialRPointer) <“45h”
then
if (MultiModeF and AddressExist and not (AddressSelected))
then
/ Ignore command if address does not match in multi mode
RenewSRP ;;
AddressSelected = false;
AddressExist = false;
exit;
else
Breg = OperationMode;
OperationMode = SerialBuffer (SerialRPointer) and “00000111b”;
if Breg <> OperationMode then
/ When there is a change in the operation mode setting
WriteToFlash;
CountRestart;
endif;
SendAck;
exit;
endif
else
SendNak;
exit;
endif
endif
if SerialBuffer (SerialRPointer) = “7Fh” then MultiModeF = true;
if SerialBuffer (SerialRPointer) = “7Eh” then MultiModeF = false;
if SerialBuffer (SerialRPointer) = (“7Eh” or “7Fh”) then
/ Multi-mode command reception processing
WriteToFlash;
CountRestart;
Sirent;
exit;
endif
if SerialBuffer (SerialRPointer) <“80h” then
SendNak;
exit;
endif;
if not (MultiModeF) then
SendNak;
exit;
endif;
AddressExist = true;
if (SerialBuffer (SerialRPointer) and “00111111b”) = #ModuleAddress
then AddressSelected = true;
else AddressSelected = false;
/ Comparison of module address and specified address
endif;
if SerialBuffer (SerialRPointer)> = “C0h”
then
/ Address selection for command input in multi mode
RenewSRP ;;
exit;
else
/ Reading random number data in multi mode
if IntervalWPointer = IntervalRPointer
then
/ When there is no random number data
SendNakWhenNoData;
exit;
else
/ When there is random data
if AddressSelected then
/ If an address is selected
ReadIntervalTime;
/ Reading time measurement value
ConvertToUniform;
/ Convert to uniformly distributed random numbers
CalcProb;
/ Probability judgment
if EXPDataOutF then SendRandomEXP;
if UniformDataOutF then SendRandomUniform;
/ Random number output
endif;
Sirent;
endif;
endif;

/ SendNak / Send NAK code
if (not (MultiModeF) or (AddressExist and AddressSelected) then SerialOutput
/ = #SNAK; / Do not send NAK if address does not match in multi mode
AddressSelected = false;
AddressExist = false;
SerialRPointer + = 1;
SerialRPointer = SerialRPointer and #ConstSPFilter;
/ Update serial read pointer
NumberOFNewData-= 1;

/ SendNakWhenNoData / When reading the multi-mode random number, there is no random number
SerialOutput = #SNAK;
AddressSelected = false;
AddressExist = false;
NumberOFNewData-= 1;

/ SendAck / Send ACK code
if (not (MultiModeF) or (AddressExist and AddressSelected) then SerialOutput
/ = #SACK; / If the address does not match in multi mode, do not send ACK
AddressSelected = false;
AddressExist = false;
SerialRPointer + = 1;
SerialRPointer = SerialRPointer and #ConstSPFilter;
/ Update serial read pointer
NumberOFNewData-= 1;

/ RenewSRP / Serial pointer update
SerialRPointer + = 1;
SerialRPointer = SerialRPointer and #ConstSPFilter;
NumberOFNewData-= 1;

/ Sirent / do nothing with invalid data
AddressSelected = false;
AddressExist = false;
SerialRPointer + = 1;
SerialRPointer = SerialRPointer and #ConstSPFilter;
NumberOFNewData-= 1;

/ WriteToFlash / Write data to flash memory
FlashCPS = CPS;
FlashCPSMT = CPSMeasureTime;
FlashNofP = NumberOfPulse;
StatusFlag = FlashStatus;
FlashOPMode = OperationMode;
FlashLastCPSAccumulation = LastCPSAccumulation;
FlashPROB (0..7) = PROB (o..7);
FlashPT (0..7) = Pvalue (0..7);

/ ConvertToUniform / Conversion to uniformly distributed random numbers
if RandomEXP [23..16] <> “0”
then
/ When the exponential distribution random number exceeds “FFh”, it is set to “FFh”
UniformedData = “FFh”;
else
/ Convert to uniform distribution by binary search
BitIndicator = 10000000b;
UniformedData = “0”;
repeat
UniformedData = UniformedData xor BitIndicator;
if RandomEXP <DataTable (UniformedData)
then UniformedData = UniformedData xor BitIndicator;
endif;
Carry = “0”;
RightLotateWithCarry (BitIndicator, 1bit);
until Carry = “1”;
endif;

/ CalcTPreScaler / Calculate Timer2 prescaler value from CPS
/ In this calculation, if the number of pulses for 64 seconds (CPS value) is 0-3, the exact prescaler value
/ Is not required. Need to confirm in advance> Cut out with CPSMAX and CPSMIN
a as M32 {
aH as M [16];
aL as M [16];
}
CycleCounter as M [8];

a = 2246Dh;
CycleCounter = 16;
LeftShiftWithCarry (a, 1bit);
repeat
if CPSForceF
then
/ If CPS value is forcibly set, use forcibly set data
if (aH-CPSForceData)> = 0
then
Carry = 1;
aH = aH CPSForceData;
else Carry = 0;
else
if (aH-CPS)> = 0
then
Carry = 1;
aH = aH CPS;
else Carry = 0;
endif;
LeftShiftWithCarry (a, 1bit);
CycleCounter = CycleCounter 1;
until CycleCounter = 0;
Timer2ReloadREG = aL;

/ CalcPThreshold / Calculates the time count threshold for probability determination used internally from the probability setting value
/ Perform logarithmic operation Refer to the design specifications for processing details.
W as M [32] {
W0 as M [8];
W1 as M [8];
W2 as M [8];
W3 as M [8]; / Operation buffer X
}
X as M [32] {
X0 as M [8];
X1 as M [8];
X2 as M [8];
X3 as M [8]; / Operation buffer X
}
Y as M [32] {
Y0 as M [8];
Y1 as M [8];
Y2 as M [8];
Y3 as M [8]; / Operation buffer Y
}
LogA as M [32] {
LogA0 as M [8];
LogA1 as M [8];
LogA2 as M [8];
LogA3 as M [8]; / Operation buffer W
}
Z as M [32] {
Z0 as M [8];
Z1 as M [8];
Z2 as M [8];
Z3 as M [8]; / Operation buffer Z
}
ZZ as M [32] {
ZZ0 as M [8];
ZZ1 as M [8];
ZZ2 as M [8];
ZZ3 as M [8]; / Operation buffer ZZ
}
Result as M [32] {
Result0 as M [8];
Result1 as M [8];
Result2 as M [8];
Result3 as M [8]; / Operation buffer ZZ
}
LogASign as M [1]; / Log (A) sign

CalcAB;
CalcZ;
CalcZZ;
CalcLogA;
CalcLogB;
CalcResult;

/ CalcAB /
B as M [8];

X1 & X2 = P (PArg);
X0 = X3 = 0;
B = “0Fh”;
repeat
LeftShift (X, 1 bit);
B-= 1;
until X0.bit0 = "1"; / Result: A remains in X.

/ CalcZ /
Root2 as “01.6A09E6”;
CycleCounter as M [8]

Y = X; Set A in both / X and Y buffers
X = X-Root2; / X = (a-√2). If the result is negative, take the complement.
if X <0 then
X = 0 X;
LogASign = true;
else
LogASign = false;
end if
Y = Y + Root2; / Y = (a + √2)

CycleCounter = 22; / bit length of divisor y
z = 0; / From here, Z = X / Y calculation
ShiftLeft (X, 2bit)
Repeat
if (XY)> 0
then
Carry = “1”; / Set the least significant bit of the result z to “1”
X = X-Y;
else
Carry = “0”; / Reset the least significant bit of the result z to “0”
end if
ShiftLeftWithC (Z, 1bit);
ShiftLeft (X, 1bit);
CycleCounter = CycleCounter 1;
until CycleCounter = 0; The operation result remains in / Z

/ CalcZZ /
CycleCounter as M [8];

X = Z;
X = Z;
WeqXmulY;
ZZ = W;

/ CalcLogA /
Const10 as “01.000000h”; / 1
Const03 as “00.555555h”; /0.3333333
Const02 as “00.333333h”; /0.2
Const07 as “00.249248h”; /“.0010 0100 1001 0010 0100 1000 ”= 0.1428571
CycleCounter as M [8];

/ ZZ / 7 operation, variable × constant operation, so the bit position where the constant is “0” is a simple shift
/ Adds bit positions where the constant is “1” to reduce processing time
Y = ZZ;
X = 0;
CycleCounter = 7
repeat
ShiftRight (Y (1..3), 3bit);
X (1..3) = X (1..3) + Y (1..3);
CycleCounter-= 1;
until CycleCounter = 0; / result remains in X

X = X + Const02;

Y = ZZ; /zz*(0.2 + zz * 0.1428571)
WeqXmulY;

W = W + Const03;

Y = ZZ; /zz*(0.3333333 + zz * (0.2 + zz * 0.1428571)
X = 0;
WeqXmulY;

W = W + Const10;

Y = Z; / 2 * z * (1 + zz * (0.3333333 + zz * (0.2 + zz * 0.1428571)))
X = W;
WeqXmulY4;
LogA = 2 * W;

/ CalcLogB /
Const05 as “00.800000h”;
ConstLog2 as “0B17217Fh”; / Log (2) * 16 is inserted to simplify the calculation

X = B + 0.5
W = 0;
Y = ConstLog2;

CycleCounter = 5; /X=(B+0.5)*Log(2)
repeat
if W0.bit3 = “1” then W = W + Y;
ShiftRight (Y, 1bit);
ShiftLeft (W (0..1), 1bit)
until CycleCounter = 0; results remain in / X

/ CalcResult /
ConstN as (“00.2BF7C4h” shl 2) = “0.1010 1111 1101 1111 0001 0000”;
/ 0.1717494 * 4 = 4 / 5.822436

if LogASign
then W = W + LogA
else W = W LogA;

Result = 0;
Y = ConstN;
CycleCounter = 22;
repeat
ShiftLeft (Y (1..3), 1bit)
if Carry = “1” then Result = Result + X;
ShiftRight (X, 1bit);
until CycleCounter = 0; results remain in / Result

/ WeqXmulY /
/ W = X * Y operation (calculations only after the decimal point)
W = 0;
CycleCounter = 24;
repeat
ShiftRight (X (1..3), 1bit);
ShiftLeft (Y (1..3), 1bit)
if Carry = “1” then W = W + X;
until CycleCounter = 0; results remain in / W

/ WeqXmulY4 /
/ W = X * Y (4 digits calculation, X is 4 digits, Y is less than 1)
W = 0;
CycleCounter = 24;
repeat
ShiftRight (X, 1bit);
ShiftLeft (Y, 1bit)
if Carry = “1” then W = W + X;
until CycleCounter = 0; / W results in W

/ FlashROMDataArea / Data area in flash memory.
/ Enter the default value of each data as the initial value

FlashCPS as M [16] = “00C0h”;
FlashCPSMT as M [8] = “01h”;
FlashNofP as M [24] = “FFFFFFh”;
FlashStatus as M [8] = “00h”;
FlashOPMode as M [8] = “00010011b”;
; bit2: Demo mode
; bit1: Exponential distribution random number output
; bit0: Uniformly distributed random number output
FlashLastCPSAccumulation as M [24] = “0000C0h”;
FlashPROB (8) as M [16] = (“0002h”, “0003h”, “0005h”, “0010h”, “0080h”,
/ “0100h”, “1000h”, “8000h”);
FlashPT (8) as M [17] = (“001D8Dh”, “002ED7h”, “00449Eh”, “00CEDEh”,
/ "00EC6Ch", "0162A2h", "01BB4Ah");
ModuleAddress as M [8] = “FFh”;
FlashVer as String = “RGZ-1 Ver 1.00”, “0Dh”, “0Ah”, “2004-05-10”, “0Dh”,
“0Ah”, “CopyRight 2004, RPG Technics & Zixsys Inc.”,
“FFh”;

It is a figure which shows the example of attachment to the IC tag of alpha particle | grain radiator. It is an attachment part enlarged view of alpha particle radiator. It is an enlarged view of the α particle radiator attachment portion (attachment example when the α particle radiator is disc-shaped). It is a figure which shows dripping and sealing example of alpha particle | grain radiator. It is a figure which shows dripping and sealing example of alpha particle | grain radiator. It is a figure which shows dripping and sealing example of alpha particle | grain radiator. 5A and 5B are diagrams showing a case where a pulse generation source is attached to a DIP type IC. It is a block diagram which shows an example of RPG-RFID which concerns on this invention. It is a graph which shows the relationship between the alpha particle emitter-detector distance, and the alpha particle count rate. It is a graph which shows the reduction | decrease in the number of electrons with the increase in the thickness of aluminum. It is a figure which shows an example of an RPG output pulse. It is a flowchart which shows a part of operation | movement of one Example of this invention. It is a flowchart which shows a part of operation | movement of one Example of this invention. It is a flowchart which shows a part of operation | movement of one Example of this invention.

Claims (27)

By providing a random pulse generation source (hereinafter referred to as RPG) that spontaneously generates a random pulse, measuring the time interval of the random pulse generated from the RPG or measuring a voltage value and converting it into a digital value, a random number A semiconductor device configured to generate and / or generate a probability, comprising:
The RPG includes alpha particles, beta rays (particles) released as a result of nuclear decay, or emitters that emit alpha particles and beta rays,
The RPG further includes a detector for detecting emitted α particles, beta rays (particles), or α particles and beta rays, and the emitter is used as a solution, and this solution is dropped onto the detection surface of the detector. Or a member formed by vapor deposition of this solution or processed into a disk shape by a roll method or the like, facing the detector through a predetermined distance,
Semiconductor device. As the emitter that emits alpha particles, beta rays (particles), or alpha particles and beta rays, all uranium series nuclides, thorium series nuclides, ²¹⁰ Pb- The semiconductor device according to claim 1, wherein all nuclides such as ²¹⁰ Po, ²⁴⁴ Cm or ²⁴¹ Am are used.   3. The semiconductor device according to claim 1, wherein the detector includes a PN-type semiconductor, a PNP-type semiconductor, a PIN photodiode, a phototransistor, or a photodiode.   The RPG further has a shielding wall that shields the emitter, and the shielding wall shields at least the α particle, the beta ray (particle) to be used, or the range of the α particle and the beta ray. The semiconductor device according to claim 1, wherein: Random pulses generated from the RPG, random exponential distribution of predetermined bit length, a semiconductor according to uniform random numbers were generated, and / or any one of claims 1 to 4, characterized in that to generate a probability device. 6. The semiconductor according to claim 5 , wherein a required number of probabilities are obtained from an exponential distribution indicating a probability that the same pulse interval occurs for pulses generated from the RPG, and uniform random numbers are created based on the probabilities. device. 2. When the probability is obtained from the time interval between t1 and t2 using the exponential distribution, the pulse is identified as a pulse with a predetermined probability when t2 is exceeded assuming that t2 is an infinite time. 7. The semiconductor device according to 6 . When creating random numbers from the exponential distribution, the same calculation process is performed even if the average emission rate of the pulse fluctuates by changing the basic period of measurement, and the stable appearance probability of the exponential random number and uniform distribution random number the semiconductor device according to any one of claims 5 to 7, characterized in that to achieve a density distribution. Storing the random number generated from the RPG as an authentication signal or anti clone signal, a semiconductor according the signal inputted from the outside to any one of claims 1 to 8, characterized in that comparing the authentication with the stored signal device. The semiconductor device according to claim 9 , wherein a new random number is rewritten as authentication data at each comparison authentication. The semiconductor device according to claim 9 or 10, characterized in that transmitting and receiving authentication data. 12. The semiconductor device according to claim 11 , wherein the authentication data is transmitted / received by wireless communication, infrared communication, or communication by a circuit connection method using contact. Claims 1 to 12 IC card incorporating a semiconductor device according to any one of, IC tag, PC connectable device or devices embedded module. When generating random numbers from an exponential distribution, the distribution of the exponential random numbers is stabilized by changing the basic period of measurement, and the generation of uniformly distributed random numbers and the generation of probabilities are simultaneously processed. Item 9. The semiconductor device according to Item 8 . 9. The method according to claim 8 , wherein when the random number is generated from the exponential distribution, fluctuations in an operation clock such as an oscillation frequency of hardware such as a microcomputer for measuring time are automatically corrected by changing a basic period of measurement. The semiconductor device described. Measuring the time interval of a random pulse generated from a random pulse generation source (hereinafter referred to as RPG) that spontaneously generates a random pulse or measuring a voltage value and converting it to a digital value;
A random pulse converted to a digital value, generating exponentially distributed random numbers of a predetermined bit length, uniform random numbers, and / or generating probability,
The RPG is an α particle, a beta ray (particle), or an emitter that emits an α particle and a beta ray, and an emitted α particle, a beta ray (particle), or an α particle. And a detector that detects beta rays, the emitter is used as a solution, and the solution is dropped on the detection surface of the detector, or the solution is deposited or processed into a disk shape by a roll method or the like Facing the detector through a predetermined distance,
A method of generating random numbers and / or probabilities. Generating random numbers, uniform random numbers of a predetermined bit length from random pulses, and / or generating probabilities, the required number of probabilities from the exponential distribution indicating the probability that the same pulse interval will occur for random pulses The method according to claim 16 , wherein a uniform random number is generated based on the probability. The step of generating a random number of an exponential distribution having a predetermined bit length, a uniform random number, and / or generating a probability from a random pulse uses the exponential distribution to obtain a probability from a time interval between t1 and t2. 18. The method according to claim 17 , wherein a pulse having a predetermined probability is recognized when t2 time is exceeded assuming that t2 time is infinite time. The step of generating an exponential distribution random number, a uniform random number of a predetermined bit length from random pulses, and / or generating a probability is performed by changing a basic period of measurement when generating a random number from an exponential distribution. , even if the same calculation processing average release rate of the pulse is varied, exponential distribution random number, characterized in that to realize stable appearance probability density distribution of uniformly distributed random numbers to any of claims 16 to 18 The method described. The step of generating an exponential distribution random number, a uniform random number of a predetermined bit length from random pulses, and / or generating a probability is performed by changing a basic period of measurement when generating a random number from an exponential distribution. The method according to any one of claims 16 to 19 , wherein the distribution of exponential distribution random numbers is stabilized, and the generation of uniform distribution random numbers and the generation of probability are processed simultaneously. The step of generating an exponential distribution random number, a uniform random number of a predetermined bit length from random pulses, and / or generating a probability is performed by changing a basic period of measurement when generating a random number from an exponential distribution. 21. The method according to claim 20 , wherein fluctuations in an operation clock such as an oscillation frequency of hardware such as a microcomputer for measuring time are automatically corrected. A program for implementing a method of generating random numbers and / or probabilities using a computer,
The method includes a step of measuring a time interval of a random pulse generated from a random pulse generation source (hereinafter referred to as RPG) that spontaneously generates a random pulse or measuring a voltage value and converting it to a digital value; Generating random numbers of an exponential distribution having a predetermined bit length, uniform random numbers, and / or generating probabilities from the converted random pulses,
The RPG is an α particle, a beta ray (particle), or an emitter that emits an α particle and a beta ray, and an emitted α particle, a beta ray (particle), or an α particle. And a detector that detects beta rays, the emitter is used as a solution, and the solution is dropped on the detection surface of the detector, or the solution is deposited or processed into a disk shape by a roll method or the like Facing the detector through a predetermined distance,
program. Generating random numbers, uniform random numbers of a predetermined bit length from random pulses, and / or generating probabilities, the required number of probabilities from the exponential distribution indicating the probability that the same pulse interval will occur for random pulses 23. The program according to claim 22 , wherein a uniform random number is generated based on the probability. 2. When the probability is obtained from the time interval between t1 and t2 using the exponential distribution, the pulse is identified as a pulse with a predetermined probability when t2 is exceeded assuming that t2 is an infinite time. The program according to 23 . When creating random numbers from the exponential distribution, the same calculation process is performed even if the average emission rate of the pulse fluctuates by changing the basic period of measurement, and the stable appearance probability of the exponential random number and uniform distribution random number The program according to any one of claims 22 to 24, wherein a density distribution is realized. When generating random numbers from an exponential distribution, the distribution of the exponential random numbers is stabilized by changing the basic period of measurement, and the generation of uniformly distributed random numbers and the generation of probabilities are simultaneously processed. Item 26. The program according to any one of Items 22 to 25 . 27. The method according to claim 26 , wherein when generating a random number from an exponential distribution, fluctuations in an operation clock such as an oscillation frequency of hardware such as a microcomputer for measuring time are automatically corrected by changing a basic period of measurement. The program described.

Images