JP2010271888A - Physical random number generator - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a physical random number generator which inexpensively generates random numbers of high quality at high speed. <P>SOLUTION: The physical random number generator includes: a laser diode 2 which is a coherent light source; a laser driver circuit 1 which makes the laser diode 2 drive and emit light; a finger skin rough surface 12 which makes light from the laser diode 2 reflect irregularly; an image sensor array 4 which receives a laser beam which has caused mutual interference by irregular reflection after projecting on skin of the skin rough surface 12; an image sensor module 5 which changes an analog voltage value of each pixel concerning a speckle pattern formed on a light receiving surface of the image sensor array 4 into a digital pixel value; a memory section 6 in which the pixel value detected by the image sensor module 5 is accumulated; a memory read/write section 7 for reading accumulated data string serially from the memory section 6; a processing section 8 which equalizes appearance ratio of "1" and "0" of the digital data read in the memory read/write section 7; and an MCU 9 which controls a series of processings of data flow until to reach a request bit length. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a physical random number generation device, and more particularly to a physical random number generation device that generates a physical random number using a coherent signal.

  In recent years, in network communication represented by e-commerce and the like, confidentiality of data to be communicated is required. Specifically, for example, data such as personal information is communicated in an encrypted state. An encryption key is used for data encryption. For generation of an encryption key, it is necessary to use a random number that cannot be predicted.

  In the fields of financial simulation and scientific simulation, it is desired to use high-quality random numbers with high uniformity in order to improve the accuracy of simulation.

  Here, a pseudo-random number generated by an algorithm, that is, a calculation according to a mathematical formula is known. Since pseudo-random numbers are generated according to a calculation algorithm, the output random number sequence inevitably has periodicity and uneven distribution. In particular, if the purpose of use is encryption, unpredictability is required for simulation, and uniformity is required in the simulation, so that periodicity and uneven distribution occur.

  As a result, in security applications, encryption keys based on pseudo-random numbers are vulnerable to attacks based on the periodicity prediction results. In addition, in simulation applications, pseudorandom numbers have periodicity, which causes uneven distribution of data. For this reason, only a limited improvement in the accuracy of simulation results can be expected.

  In recent years, physical random numbers generated using randomness resulting from physical phenomena in the natural world have attracted attention. The generation of physical random numbers does not require initial values and algorithms that were necessary in the case of pseudo-random number generation.

  In the case of physical random numbers, randomness due to natural phenomena is not reproducible, and a random number sequence generated based on the randomness is a random sequence that is difficult to predict and does not have reproducibility. Therefore, physical random numbers satisfy “reproducibility” and “unpredictability”, which are essential requirements for encryption key generation in security applications, and random numbers with high randomness are also in simulation applications. Can be provided.

  Patent Documents 1 and 2 disclose a method of generating a “physical random number”. This type of prior art employs a method called a “circuit noise method”, and includes a noise source, a high-sensitivity preamplifier, a filter, a weak signal amplifier circuit, and an A / D converter. ing.

  Patent Document 3 discloses a technique for generating a random number using a radio signal including radio waves emitted from radio stars.

JP 2008-165500 A JP 2008-299595 A Patent No. 3876324

  However, the circuit noise type random number generator uses the weak level of thermal noise generated by a single component of the electronic circuit as a random number generator, so a high-sensitivity preamplifier must be placed immediately after the thermal noise source. At a certain point, the number of parts is relatively large, and this increases the cost of the random number generation device.

  In addition, in the circuit noise type random number generation device, when ambient noise (radiation noise from other devices, etc.) is mixed in the immediate vicinity of the noise source, the quality of randomness of the generated random number is deteriorated. In order to avoid this, a troublesome process in which the noise source must be strengthened against ambient noise is also required.

  In addition, the circuit noise type random number generation device needs to include an A / D converter that can sample at high speed in order to improve random number generation efficiency. This is another factor that inevitably increases the cost of the random number generation device.

  Furthermore, when generating random numbers using radio signals including radio waves emitted from radio stars, it is necessary to prepare a large-scale antenna facility. This also contributes to the cost increase of the apparatus.

  In view of such a problem, an object of the present invention is to provide a random number generator capable of generating high-quality physical random numbers at a low cost.

In order to solve the above problems, the physical random number generation device of the present invention provides:
A means for generating a signal whose signal level is dynamic (such as the optical unit 10 in FIG. 1);
Processing means (such as the processing unit 8 in FIG. 1) for reducing the bias of the signal level caused by the generating means;
Instructing means (such as MCU (Micro Controller Unit) 9 in FIG. 1) for instructing the bit length of a random number generated by being processed by the processing means.

  The processing means can perform a uniformization process such as generating an m-bit (n ≠ m) data string based on an n-bit bit sequence constituting a signal having a dynamic signal level. A Neumann collector may be employed for the uniformization process.

  Further, the processing means can perform a filter process for clarifying the level of the signal level. For example, when the generating means generates an optical signal, a Laplacian filter can be used, and when the generating means generates a sound signal, a bandpass filter can be used.

  Further, the processing means can perform differential processing for clarifying the level of the signal level. The difference process performs a difference process on a specific element of a signal temporally or spatially. Specifically, for example, when the generating means generates an optical signal, a method of storing the level of the optical signal in a memory or the like and taking a difference between adjacent pixels can be employed.

  The signal is a coherent signal. Therefore, any signal including a wave component such as an optical signal, a radio wave signal, and a sound signal may be used. As an example, laser light can be used. In this case, the reflected light obtained when the rough object is irradiated with laser light or the transmission light obtained by applying an external force to the optical fiber that is the laser light transmission path is used as the signal. Can do.

As a specific configuration of the physical random number generator of the present invention,
A laser diode which is a coherent light source;
A laser driver circuit for driving and emitting the laser diode;
An irregular reflection type optical path composed of a rough surface object that irregularly reflects light from the laser diode in the course of spatial propagation;
An image sensor array that receives laser light that has been diffusely reflected after the projection on the rough surface and caused mutual interference;
An image sensor module for converting an analog voltage value of each pixel formed on the light receiving surface of the image sensor array and receiving a speckle pattern into a digital pixel value;
A memory unit for temporarily storing pixel values detected by the image sensor;
A memory read / write unit that sequentially reads / writes a stored data string from the memory unit;
A uniformization processing unit for equalizing the appearance rate of “1” (mark) and “0” (space) of digital data read from the memory unit via a memory read / write unit;
And control means for controlling a series of processing of the data flow between the units until the required bit length is reached.

Alternatively, as a specific configuration of the physical random number generator of the present invention,
A laser diode which is a coherent light source;
A laser driver circuit for driving and emitting the laser diode;
An optical waveguide including an optical fiber that propagates light from the laser diode;
A piezoelectric element that randomly changes the interference state by applying a disturbance to the optical waveguide;
“Pseudo Noise (PN) pattern generation circuit for electrically driving the piezoelectric element randomly,“ disturbance application type optical path ”,
An image sensor array for receiving output light from the output end of the optical fiber;
An image sensor module which is formed on the light receiving surface of the image sensor array and converts an analog voltage value for each pixel which has received a speckle pattern into a digital pixel value;
A memory unit for temporarily storing pixel values detected by the image sensor;
A memory read / write unit that sequentially reads / writes a stored data string from the memory unit;
A uniformization processing unit for equalizing the appearance rate of “1” (mark) and “0” (space) of digital data read from the memory unit via a memory read / write unit;
And control means for controlling a series of processing of the data flow between the units until the required bit length is reached.

  However, as will be described later with reference to the embodiment, instead of the “disturbance application type optical path”, a rough surface of the finger skin can be used. In this case, since it is not necessary to provide a PN pattern generation circuit, a physical random number independent of the PN pattern can be obtained, and the physical random number generation device can be made inexpensive because the PN pattern generation circuit is not provided.

  According to the present invention, since it is not necessary to prepare a high-sensitivity preamplifier, an A / D converter, an antenna facility, or the like that causes a cost increase, the random number generation device can suppress the cost increase. In addition, since there is no factor that degrades the quality of randomness of random numbers, it is possible to provide a physical random number generation device that can obtain physical random numbers with excellent randomness.

  The best mode for carrying out the present invention will be described below in detail with reference to the drawings.

  FIG. 1 is a schematic explanatory diagram of a physical random number generator according to an embodiment of the present invention. The physical random number generation device according to the present embodiment is roughly divided into an optical unit 10 that emits and receives light and a data processing unit 20 that processes a signal from the optical unit 10. Hereinafter, the physical random number generation device of the present embodiment will be described according to the operations of the optical unit 10 and the data processing unit 20.

  First, in the optical unit 10, the laser driver circuit 1 drives the laser diode 2 that is a coherent light source according to a command from the MCU 9. Thus, when the laser diode 2 emits light, the light from the laser diode 2 is projected onto the irregular reflection surface 3. As a result, scattered light due to irregular reflection of the projection light from the irregular reflection surface 3 is obtained.

  The image sensor array 4 receives scattered light from the irregular reflection surface 3 and outputs an electrical signal corresponding to the scattered light to the image sensor module 5. The scattered light received by the image sensor array 4 becomes a speckled pattern called a speckle pattern. The image sensor module 5 converts the output for each pixel from the image sensor array 4 from an analog signal to a digital signal.

  Next, in the data processing unit 20, the memory read / write control unit 7 sequentially writes the signals converted by the image sensor module 5 in the memory unit 6. As a result, a multi-value digital signal corresponding to each pixel of the image sensor array 4 is accumulated in the memory unit 6.

  Further, the memory read / write control unit 7 reads out the data sequence stored in the memory unit 6 after the signals from all the pixels are stored in the memory unit 6, and outputs them to the processing unit 8. For example, the processing unit 8 performs a filtering process, a binarization process, and the like on the data string, and further equalizes the appearance rate of “1” (mark) and “0” (space) of the binary signal. The uniform process is performed. The MCU 9 issues an operation command to each unit until the data length after the equalization processing becomes a desired data length.

  Hereinafter, a physical random number generator of this type will be described in detail with some examples.

(First embodiment)
FIG. 2 is a schematic configuration diagram of the physical random number generation device according to the first exemplary embodiment of the present invention. As shown in FIG. 2, the physical random number generation device according to the present embodiment includes a laser driver circuit 1 that drives a laser diode 2, a laser diode 2 that is a coherent light source, and light from the laser diode 2 that has a finger surface. 12, a lens 11 that collects reflected light from the rough surface 12 of the finger skin on the image sensor array 4, an image sensor array 4 that receives light from the lens 13, and reflected light that is received by the image sensor array 4. And an image sensor module 5 for converting an electric signal based on the analog signal into a digital signal.

  The physical random number generation device according to the present embodiment also includes a memory unit 6 in which signals from the image sensor module 5 are stored, an MCU 9 that controls the operation of the entire physical random number generation device, and physical random numbers output from the MCU 9. A PC personal computer (hereinafter referred to as “PC”) 15 for input and a USB interface circuit 14 for connecting the processing unit 8 and the PC 15 are provided.

  Further, the MCU 9 includes a memory read / write control unit 7 that controls reading / writing of signals to / from the memory unit 6, and a processing unit that reduces the deviation in signal level caused by the optical unit 10 with respect to the signal read from the memory unit 6. 8.

  FIG. 3 is a flowchart showing the operation of the physical random number generator shown in FIG. First, the user turns on the power of the random number generation device (step S1).

  When a random number acquisition request is made from the PC 15 in this state, the request is input to the MCU 9. When receiving this request, the MCU 9 turns on the laser driver circuit 1 (step S2).

  As a result, the laser diode 2 emits laser light (step S3).

  Note that the PC 15 also requests the bit length of the random number to be acquired when making a random number acquisition request.

  Laser light emitted from the laser diode 2 is focused by the lens 11 and projected onto the rough surface 12 of the human finger skin pressed against the opening of the housing. As a result of this laser light being irregularly reflected by the finger skin rough surface 12, a light-and-dark brightness pattern with a low correlation is generated between the two-dimensional coordinate position and the light intensity at that position. This pattern is an optical phenomenon and is called a speckle pattern.

  Here, since the human finger skin itself is only a part of the tissue of the living body, even though it seems to be stationary macroscopically, it is not stationary microscopically. In other words, the human finger skin itself is moving in the time direction because the position, shape, and state of the skin fluctuate randomly due to various uncertain and unstable factors when viewed from the level of the wavelength size of light. However, the light interference state changes, and the luminance at a certain position coordinate has a random distribution. As a result, a random luminance distribution is generated in the coordinate direction and the time direction.

  Even if it is not projected onto the finger skin rough surface 12, light traveling in the atmosphere is subject to fluctuations of fine particles in the air or air, so that a random luminance distribution is generated. Note that it is not essential to project to

  The laser light irregularly reflected by the rough finger surface 12 is collected by the lens 13 and received by the image sensor array 14 (step S4).

  The image sensor array 14 converts an optical signal into an electric signal and outputs it to the image sensor module 5. The image sensor module 5 converts the voltage value of each pixel of the electrical signal from an analog signal to a digital signal (step S5).

  The MCU 9 determines whether or not there is a random number generation request from the PC 15 (step S6).

  As a result of the determination, if there is no random number generation request from the PC 15, step S6 is continued. On the other hand, if there is a random number generation request from the PC 15, the memory read / write control unit 7 writes the data obtained by the image sensor module 5 to the memory unit 6 up to the image full size (step S7).

  After the image is written up to the full size in the memory unit 6, the memory read / write control unit 7 reads out the data written in the memory unit 6 to the processing unit 8.

  Here, the signal read from the memory unit 6 has a signal level deviation caused by the optical unit 10. Typically, since the laser diode 2 is a point light source, the light received by the image sensor module 5 is biased such that the light receiving level is higher in the central area than in the peripheral area. That is, there is an unavoidable bias in the signal level due to the nature of each hardware part in the optical unit 10. For this reason, in this embodiment, this bias is reduced to improve the randomness of random numbers.

  For example, the processing unit 8 performs a uniformization process such as acquiring 1 bit or acquiring 0 bit based on two consecutive bit strings (step S8).

  Specifically, in this case, 1 bit of “0” is acquired based on the 2 bit strings of “01”, 1 bit of “0” is acquired based on the 2 bit strings of “01”, 1 bit of “1” is acquired based on the 2 bit strings “10”, and 0 bit is acquired (discarded) based on the 2 bit strings of “00” and 2 bit strings of “11”. . As a result, the occurrence probability of “0” and “1” is equalized in the bit string output from the processing unit 8.

  Since the data before the equalization processing is composed of a bit string that is a random physical phenomenon according to the statistical probability distribution, it has an uneven distribution similar to the statistical distribution. On the other hand, this uneven distribution is eliminated in the data after the equalization processing. In addition, the data after the uniformization process has a probability of occurrence of “0” and “1” in the National Institute of Standards and Technology (NIST) SP800-22, which is the de facto standard for the specification of random numbers to be applied in security technology. It is in line with the requirement that is as uniform as possible (mono bit test).

  However, the process of acquiring one bit based on two consecutive bit strings is merely an example. For example, one bit is acquired based on three consecutive bit strings, or two bit strings are acquired from one bit, for example. Processing may be performed.

  Moreover, the process part 8 can perform a filter process instead of the equalization process or with the equalization process. As in the present embodiment, when the processing target is an optical signal, a Laplacian filter can be used. When the processing target is a sound signal, a band pass filter can be used.

  Furthermore, instead of the uniforming process or the filter process, or together with the uniformizing process or the filter process, a difference process can be performed. The difference process performs a difference process on a specific element of a signal temporally or spatially. Specifically, the signal level difference between adjacent pixels of the image sensor module 5 is taken, or the pixel level of the first row and the first column is used as a reference pixel, for example, and the signal level of other pixels in sequence with the pixel. It is possible to adopt a technique such as taking the difference between the two.

  The MCU 9 issues a data output command to the processing unit 8 after performing the equalization process. In accordance with the processing unit 8 and the output command, the data after the equalization processing is output to the USB interface circuit 14 (step S9).

  Thereafter, the MCU 9 determines whether or not the bit length of the data after the equalization processing and the like output to the USB interface circuit 14 has reached the random number request bit length from the PC 15 (step S10).

  As a result of the determination, if the bit length of the data after the equalization processing or the like has not reached the random number request bit length, the process proceeds to step S8. On the other hand, if the bit length of the data after the equalization processing has reached the random number request bit length, a random number sequence is output from the USB interface circuit 14 to the PC 15 (step S11).

(Second embodiment)
FIG. 4 is a schematic configuration diagram of a physical random number generation device according to the second exemplary embodiment of the present invention. FIG. 4 shows an optical fiber 16 that transmits the light collected by the lens 11, a piezoelectric element 17 for applying an external force to the optical fiber 16, and an electric signal of a PN pattern supplied to the piezoelectric element 17. PN pattern generation circuit 18 for generating In FIG. 4, the same parts as those shown in FIG.

  FIG. 5 is a flowchart showing the operation of the physical random number generator shown in FIG. First, as in step S1, when the random number generation device is powered on (step S21) and a random number acquisition request is made from the PC 15, the request is input to the MCU 9. When receiving this request, the MCU 9 turns on the laser driver circuit 1 and turns on the PN pattern generation circuit 18 (step S22).

  As a result, the laser diode 2 emits laser light, and the piezoelectric element 17 is driven according to the PN pattern (step S23).

  For this reason, the laser light from the laser diode 2 is focused by the lens 11 and enters the optical fiber 16.

  Here, the incident light of the optical fiber 16 ideally travels while being totally reflected at various reflection angles at the interface between the core and the clad. At this time, since the piezoelectric element 17 applies a random external force to the optical fiber 6, the optical path itself changes.

  Therefore, the incident light of the optical fiber 6 travels through the optical fiber 6, and the irregular reflection angle changes with respect to the time axis. Therefore, a speckle pattern with a luminance distribution having temporal randomness is generated. .

  However, even if an external force is not applied to the optical fiber 6 using the piezoelectric element 17, the structural non-uniformity of the core or the clad constituting the optical fiber 6, the refractive index due to impurities mixed in the core or the clad Due to the non-uniformity, the total reflection surface is microscopically uneven and randomly rough.

  As described above, even if the incident light of the optical fiber 6 travels only in the optical fiber 6, a speckle pattern with a spatially random luminance distribution is generated in the light incident on the image sensor array 14. Become.

  Therefore, the physical random number generation device according to the present embodiment, when an external force is applied to the optical fiber 6 using the piezoelectric element 17, has a speckle pattern with a luminance distribution that has randomness both spatially and temporally. Will be generated.

  The subsequent processing in the physical random number generation device of the present embodiment is the same as steps S6 to S11 (steps S26 to S31).

(Third embodiment)
FIG. 6 is a schematic partially enlarged view of the physical random number generation device according to the third embodiment of the present invention. This physical random number generator is an explanatory diagram of a modification of the physical random number generator shown in FIG. In FIG. 6, the enlarged view of finger skin rough surface 12 vicinity is shown.

  Here, in addition to the laser diode (LD) 2 and the image sensor array 4 described above, a plurality of beam splitters 21 that divide the light emitted from the LD 2 and the light from each beam splitter 21 are applied to the finger skin rough surface 12. A plurality of reflecting prisms 22 that reflect the light so as to irradiate the side surface, an optical attenuator 23 that is provided between the reflecting prism 22 and the rough finger skin surface 12 and can attenuate, for example, 50%, and rough finger skin. A glass plate 25 on which the surface 12 is placed, and an electronic shutter 24 selectively provided between the glass plate 25 and the beam splitter 21 are shown.

  In this embodiment, when generating random numbers, not only the reflected light of the irradiation light from the LD 2 to the rough skin 12 but also the transmitted light obtained when the rough skin 12 is irradiated using the reflecting prism 22 or the like is used. Is also used.

  The transmitted light from the finger skin rough surface 12 can also be used for so-called vein authentication. At this time, the reflected light and the transmitted light do not enter the image sensor array 4 at the same time. Specifically, the electronic shutter 24 is closed during vein authentication so that the reflected light does not enter the image sensor array 4. It is good to.

It is a figure explaining the principle of the random number generator of this invention. It is a figure explaining the structure of Example 1 of this invention. It is a figure explaining the principle of Example 1 of this invention. It is a figure explaining the structure of Example 2 of this invention. It is a figure explaining the principle of Example 2 of this invention. It is a typical partial enlarged view of the physical random number generator of Example 3 of this invention.

Claims (3)

Means for generating a signal whose signal level is dynamic;
Processing means for reducing bias in signal level caused by the generating means;
A physical random number generator comprising: instruction means for instructing a bit length of a random number generated by being processed by the processing means.   The physical random number generation device according to claim 1, wherein the signal is an optical signal from a rough object.   The physical random number generation device according to claim 1, wherein the signal is obtained by applying an external force to a transmission path of the signal.

Images