CN118466902B - Ultra-miniature anti-ripple quantum random number chip - Google Patents

Abstract

The invention relates to the technical field of optical secret communication, and discloses an ultra-micro anti-ripple quantum random number chip, wherein a photosensitive pixel array is used for converting optical signals irradiated on each pixel into electric signals and outputting an electric signal matrix, and each element of the electric signal matrix corresponds to a voltage digital signal converted from photocurrent accumulated by the corresponding pixel in exposure time; the data preprocessing algorithm comprises the steps of dividing an electric signal matrix into a plurality of blocks with the same element number, and differentiating elements of each block and any one other block to obtain a new electric signal matrix serving as original random bits. Compared with the prior art, the invention detects photon number fluctuation of the light source by utilizing the photosensitive pixel array, eliminates ripple signals in the light source driving signals by difference, not only can greatly improve the integration level of the chip, but also can improve the entropy content of the quantum entropy source and improve the randomness and safety of random numbers.

Description

An ultra-micro ripple-resistant quantum random number chip

Technical Field

The invention relates to the technical field of optical secure communication, and in particular to an ultra-micro ripple-resistant quantum random number chip.

Background Art

In modern society, random numbers are widely used in many fields such as simulation and cryptography. According to different generation principles, random numbers can be divided into two categories: pseudo-random numbers and true random numbers. Since pseudo-random numbers are generally generated by algorithms, with the increasingly urgent threat of quantum computing, pseudo-random numbers will become predictable, so their security cannot be guaranteed. Quantum random number generator (QRNG) is a new technology that uses the intrinsic characteristics of quantum physics to generate physical true random numbers. For example, quantum random number generators based on quantum optical principles such as quantum vacuum state noise and quantum phase noise of laser spontaneous radiation generate completely unpredictable random numbers. Therefore, they have true randomness and are also the most studied and mature quantum random number generation schemes. At present, quantum random number generators are mostly built with discrete optical components, which are large in size, complex in structure, poor in stability, high in cost, and difficult to mass produce. Therefore, integrating optical devices is an inevitable development trend. Quantum random number generators based on vacuum state quantum noise and laser phase noise can be chip-based, but they all require lasers, detectors or interferometers, which are large in size and complexity.

Patent CN107066236B proposes a scheme to generate quantum random numbers by using the fluctuation of the number of photons of a light source. This scheme uses the fact that the number of photons emitted by a light source follows the Poisson distribution, and the number of photons emitted per unit time is random and cannot be accurately predicted. This quantum effect is usually called "quantum noise" or "shot noise", which is caused by the quantum fluctuation characteristics of the light field. The quantum entropy source of this scheme is very simple to implement, and only requires an LED to be irradiated to a photon sensor (such as a CCD camera, etc.). However, for any light source drive signal, such as the LED drive current, there is inevitably a ripple current, which will cause the light intensity of the LED to fluctuate randomly with the change of the ripple current. This part of the fluctuation may be controlled by the eavesdropper, so the photon number fluctuation of the light source itself and the light intensity fluctuation caused by the ripple are superimposed and cannot be distinguished. If the influence of the light intensity fluctuation noise caused by the ripple signal on the quantum random number chip is ignored, that is, this part of the noise is also regarded as quantum noise, as the quantum entropy source of the quantum random number chip, the size of the quantum entropy will be overestimated, which greatly reduces the security of the random number chip.

Summary of the invention

In view of the above defects in the prior art, the present invention proposes an ultra-micro ripple-resistant quantum random number chip.

The technical solution of the present invention is achieved in this way:

An ultra-micro anti-ripple quantum random number chip, including a driving module, a light source, a photosensitive pixel array and a post-processing module.

The driving module is used to generate a driving signal to drive the light source to generate a light signal, and the amplitude of the driving signal is greater than the working threshold of the light source;

The light source is used to generate a light signal with a Poisson distribution of photons;

The photosensitive pixel array is used to convert the light signal irradiated on each pixel into an electrical signal and output an electrical signal matrix, each element of which corresponds to a voltage digital signal converted from the photocurrent accumulated by the corresponding pixel during the exposure time and contains quantum noise of random fluctuations in the number of photons;

The post-processing module is used to process the electrical signal matrix by using a data pre-processing algorithm and a randomness extraction algorithm to obtain a random bit string after randomness extraction;

The data preprocessing algorithm includes dividing the electric signal matrix into a plurality of blocks with an equal number of elements, and performing a difference between the elements of each block and any other block to obtain a new electric signal matrix as the original random bits.

Preferably, the data preprocessing algorithm divides the electrical signal matrix into two blocks with equal numbers of elements, and differentiates the corresponding elements of the two blocks to obtain a new electrical signal matrix as the original random bits.

Preferably, an optical diffuser is provided between the light source and the photosensitive pixel array for uniformly distributing the light signal spatially on the photosensitive pixel array.

Preferably, the light source comprises one or more laser diodes LD.

Preferably, the light source comprises one or more light emitting diodes (LEDs).

Preferably, the photosensitive pixel array is a CCD sensor or a CMOS sensor.

Preferably, the randomness extraction algorithm is a von Neumann correction algorithm: the adjacent 2 bits 01 in the original random bits are taken as 0, 10 as 1, and consecutive 00 and 11 are discarded to obtain an extracted random bit string.

Preferably, the randomness extraction algorithm is a Toeplitz matrix algorithm based on fast Fourier transform: construct a Toeplitz matrix, multiply it with the original random bits to obtain an extracted random bit string.

Compared with the prior art, the present invention has the following beneficial effects:

The present invention proposes an ultra-micro anti-ripple quantum random number chip, which uses a photosensitive pixel array to detect the fluctuation of the photon number of a light source and eliminates the ripple signal in the light source driving signal by difference. This can not only greatly improve the integration of the chip and reduce its size to the mm³ level, but also improve the entropy content of the quantum entropy source and the randomness and security of the random number.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a schematic diagram of the structure of an ultra-micro anti-ripple quantum random number chip according to the present invention;

FIG2 is a structural schematic diagram of a first embodiment of an ultra-micro anti-ripple quantum random number chip of the present invention;

FIG3 is a schematic diagram of the area division of the electrical signal matrix of the photosensitive pixel array of the present invention;

FIG4 is a histogram of the electrical signal directly output by a single pixel of the present invention;

FIG. 5 is a histogram of the differential output electrical signal of a single pixel of the present invention.

DETAILED DESCRIPTION

The present invention will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present invention.

As shown in FIG1 , an ultra-micro anti-ripple quantum random number chip includes a driving module, a light source, a photosensitive pixel array, and a post-processing module.

The light source is used to generate a light signal with a Poisson distribution of photons;

The driving module is used to generate a driving signal to drive the light source to generate a light signal, and the amplitude of the driving signal is greater than the working threshold of the light source;

The photosensitive pixel array is used to convert the light signal irradiated on each pixel into an electrical signal and output an electrical signal matrix, each element of which corresponds to a voltage digital signal converted from the photocurrent accumulated by the corresponding pixel during the exposure time and contains quantum noise of random fluctuations in the number of photons;

The post-processing module is used to process the electrical signal matrix by using a data pre-processing algorithm and a randomness extraction algorithm to obtain a random bit string after randomness extraction;

The data preprocessing algorithm includes dividing the electric signal matrix into a plurality of blocks with an equal number of elements, and performing a difference between the elements of each block and any other block to obtain a new electric signal matrix as the original random bits.

The specific working principle is as follows:

The light source generates a light signal, and the photon number distribution of the light signal obeys Poisson distribution

,

Where u is the average number of photons in the optical signal, It represents the probability that a light signal contains n photons. Due to the limitations of quantum mechanics, the number of photons emitted per unit time cannot be predicted completely and accurately under any circumstances, that is, the light signal emitted by the light source has quantum fluctuations, which is called quantum noise. By measuring the number of photons in the light signal through a photosensitive device, the quantum fluctuations of the light source can be converted into electrical signal fluctuations and collected, and a bit string containing quantum randomness can be obtained.

Assuming that during the exposure time, the light signal is uniformly irradiated onto the photosensitive pixel array, and the average number of photons absorbed by each pixel is equal, after processes such as photoelectric conversion, signal amplification, and data acquisition, the accumulated photocurrent of each pixel is converted into a voltage digital signal, and an electrical signal corresponding to the number of absorbed photons is output. Finally, an electrical signal matrix corresponding to the photosensitive pixel array is obtained, which contains quantum noise caused by fluctuations in the number of photons emitted by the light source.

Due to the imperfection of the device, the electrical signal matrix will contain other technical noises besides quantum noise, including the intensity fluctuation noise of the light signal emitted by the light source caused by the ripple signal in the light source driving signal, and the electronic noise of the photosensitive pixel array. The latter can be obtained by turning off the light source and directly measuring the electrical noise of each pixel in the photosensitive pixel array. The former is mixed with quantum noise and cannot be measured separately. If the influence of the light intensity fluctuation noise caused by the ripple signal on the quantum random number chip is ignored, that is, all noises except the electrical noise of the photosensitive pixel array are regarded as quantum noise as the quantum entropy source of the quantum random number chip, the quantum entropy will be overestimated. Then the eavesdropper can change the proportion of quantum noise in all noises by controlling the ripple signal, thereby reducing the security of the ultra-micro anti-ripple quantum random number chip, so it is necessary to eliminate the influence of the ripple signal. Through analysis, it can be found that the ripple signal causes the overall intensity fluctuation of the light source, which has the same effect on each pixel and can be eliminated by signal differential operation between pixels.

The number of pixels in the photosensitive pixel array is M×N, so the output electrical signal matrix is

.

Divide the electrical signal matrix R into 2K blocks, where K is a positive integer, and you can get 2K sub-matrices. and Subtracting them will give us a new submatrix , thus we can get the new electrical signal matrix R'. Assuming that the average number of photons detected by each element in the photosensitive pixel array is u, and the number of photons obeys the Poisson distribution, then each element value of the electrical signal matrix obeys the Poisson distribution with a mean of u, and each element of the corresponding new electrical signal matrix R' obeys the Skellam distribution. The probability density distribution can be written as

,

Its mean is 0 and its variance is 2u. The quantum minimum entropy is

,

When the collected data is quantized into b bits, the upper limit of the randomness extraction ratio can be calculated as The minimum value of the upper limit of the extraction ratio calculated from all elements is selected as the extraction ratio of the post-processing algorithm, and the electrical signal matrix can eventually be processed into the final random number sequence.

As shown in FIG2 , an embodiment of the ultra-micro anti-ripple quantum random number chip of the present invention:

The data preprocessing algorithm divides the electric signal matrix into two blocks with equal numbers of elements, and differentiates the corresponding elements of the two blocks to obtain a new electric signal matrix as the original random bits.

An optical diffuser is also provided between the light source and the photosensitive pixel array to make the spatial distribution of the light signal on the photosensitive pixel array more uniform.

The light source includes one or more light emitting diodes (LEDs).

The photosensitive pixel array is a CCD sensor.

The randomness extraction algorithm is a Toeplitz matrix algorithm based on fast Fourier transform: a Toeplitz matrix is constructed and multiplied with the original random bits to obtain an extracted random bit string.

The specific work process includes:

The light source generates a light signal, and the photon number distribution of the light signal obeys Poisson distribution

,

Where u is the average number of photons in the optical signal, It represents the probability that a light signal contains n photons. Due to the limitations of quantum mechanics, the number of photons emitted per unit time cannot be predicted completely and accurately under any circumstances, that is, the light signal emitted by the light source has quantum fluctuations, which is called quantum noise. By measuring the number of photons in the light signal through a photosensitive device, the quantum fluctuations of the light source can be converted into electrical signal fluctuations and collected, and a bit string containing quantum randomness can be obtained.

Assuming that during the exposure time, the light signal is uniformly irradiated onto the photosensitive pixel array, and the average number of photons absorbed by each pixel is equal, after processes such as photoelectric conversion, signal amplification, and data acquisition, the accumulated photocurrent of each pixel is converted into a voltage digital signal, and an electrical signal corresponding to the number of absorbed photons is output. Finally, an electrical signal matrix corresponding to the photosensitive pixel array is obtained, which contains quantum noise caused by fluctuations in the number of photons emitted by the light source.

Due to the imperfection of the device, the electrical signal matrix will contain other technical noises besides quantum noise, including the intensity fluctuation noise of the light signal emitted by the light source caused by the ripple signal in the light source driving signal, and the electronic noise of the photosensitive pixel array. The latter can be obtained by turning off the light source and directly measuring the electrical noise of each pixel in the photosensitive pixel array. The former is mixed with quantum noise and cannot be measured separately. If the influence of the light intensity fluctuation noise caused by the ripple signal on the quantum random number chip is ignored, that is, all noises except the electrical noise of the photosensitive pixel array are regarded as quantum noise as the quantum entropy source of the quantum random number chip, the quantum entropy will be overestimated. Then the eavesdropper can change the proportion of quantum noise in all noises by controlling the ripple signal, thereby reducing the security of the quantum random number chip, so it is necessary to eliminate the influence of the ripple signal. Through analysis, it can be found that the overall intensity fluctuation of the light source caused by the ripple signal has the same effect on each pixel, which can be eliminated by signal differential operation between pixels.

The number of pixels in the photosensitive pixel array is M×N, so the output electrical signal matrix is

.

As shown in Figure 3, the electrical signal matrix R is divided into two blocks A and B, and then two sub-matrices can be obtained: and . Submatrix and Subtracting them will give a new electrical signal matrix R'. Assuming that the average number of photons detected by each element in the photosensitive pixel array is u, and the number of photons follows a Poisson distribution, then each element value of the electrical signal matrix follows a Poisson distribution with a mean of u, and each element of the corresponding new electrical signal matrix R' follows a Skellam distribution, and the probability density distribution can be written as

,

Its mean is 0 and its variance is 2u.

As shown in Figures 4 and 5, a single pixel is given The bar graph of the direct output signal and the differential output signal after the exposure shows that the average number of photons detected by each pixel during the exposure time is 199.98, with a variance of 199.65, while the mean value of the differential output electrical signal of the pixel is -0.002, with a variance of 398.69. It can be seen that the variance of the output signal becomes larger after the difference.

The quantum minimum entropy is

,

The larger the variance, the larger the minimum entropy, so the quantum entropy source content can be increased by differentiation. When the collected data is quantized into b bits, the upper limit of the randomness extraction ratio can be calculated as , select the minimum value of the upper limit of the extraction ratio calculated by all elements as the extraction ratio of the post-processing algorithm. The Toeplitz matrix algorithm based on fast Fourier transform is used to extract randomness from the original data of length L, and the residual hash lemma is used to determine the length of the output random sequence.

,

Among them, ε is the information-theoretic security parameter.

Then, a j×L Toeplitz matrix is constructed using a random number seed of j+L-1 bits. By multiplying the original data with the Toeplitz matrix, the extracted random bit string can be obtained, and the final quantum random number is output.

From the various embodiments of the present invention, it can be seen that the present invention proposes an ultra-micro anti-ripple quantum random number chip, which uses a photosensitive pixel array to detect the fluctuation of the photon number of a light source, and eliminates the ripple signal in the light source driving signal by difference. This can not only greatly improve the integration of the chip and reduce its size to the mm³ level, but also improve the entropy content of the quantum entropy source, thereby improving the randomness and security of random numbers.

Claims (8)

1. An ultra-micro anti-ripple quantum random number chip, characterized by comprising a driving module, a light source, a photosensitive pixel array and a post-processing module, The driving module is used to generate a driving signal to drive the light source to generate a light signal, and the amplitude of the driving signal is greater than the working threshold of the light source; The light source is used to generate a light signal with a Poisson distribution of photons; The photosensitive pixel array is used to convert the light signal irradiated on each pixel thereof into an electrical signal and output an electrical signal matrix, each element of which corresponds to a voltage digital signal converted from the photocurrent accumulated by the corresponding pixel during the exposure time and contains quantum noise of random fluctuations in the number of photons; The post-processing module is used to process the electrical signal matrix by using a data pre-processing algorithm and a randomness extraction algorithm to obtain a random bit string after randomness extraction; The data preprocessing algorithm includes dividing the electric signal matrix into a plurality of blocks with an equal number of elements, and performing a difference between the elements of each block and any other block to obtain a new electric signal matrix as the original random bits. 2. The ultra-micro anti-ripple quantum random number chip according to claim 1 is characterized in that the data preprocessing algorithm divides the electrical signal matrix into two blocks with equal numbers of elements, and differentiates the corresponding elements of the two blocks to obtain a new electrical signal matrix as the original random bits. 3. The ultra-micro anti-ripple quantum random number chip according to claim 1 is characterized in that an optical diffuser is further provided between the light source and the photosensitive pixel array for making the light signal spatially uniformly distributed on the photosensitive pixel array. 4 . The ultra-micro anti-ripple quantum random number chip according to claim 1 , wherein the light source comprises one or more laser diodes LD. 5. The ultra-micro anti-ripple quantum random number chip according to claim 1, characterized in that the light source includes one or more light-emitting diodes (LEDs). 6 . The ultra-micro anti-ripple quantum random number chip according to claim 1 , wherein the photosensitive pixel array is a CCD sensor or a CMOS sensor. 7. The ultra-micro anti-ripple quantum random number chip according to any one of claims 1 to 6 is characterized in that the randomness extraction algorithm is a von Neumann correction algorithm: the adjacent 2 bits 01 in the original random bits are taken as 0, 10 as 1, and consecutive 00 and 11 are discarded to obtain the extracted random bit string. 8. The ultra-micro anti-ripple quantum random number chip according to any one of claims 1 to 6, characterized in that the randomness extraction algorithm is a Toeplitz matrix algorithm based on fast Fourier transform: construct a Toeplitz matrix, multiply it with the original random bit, and obtain the extracted random bit string.