Quantum random number generator based commercial cryptographic system entropy source quality detection method

Abstract

The present application belongs to the technical field of commercial cipher system test, and particularly relates to a method for detecting the quality of an entropy source of a commercial cipher system based on a quantum random number generator, which comprises: obtaining shot noise of an optical quantum detector at a high sampling rate, removing direct current bias by using sliding average subtraction, and obtaining a quantum fluctuation sequence with zero mean; obtaining a correlation degree by using nonlinear mapping of an exponential function based on the variance ratio and short-range autocorrelation of the sequence, so as to amplify weak non-quantum memory effects; obtaining an evaluation model containing a hyperbolic cosine and a logarithmic function by combining the drift potential of a bias voltage relative to a target value and the aging time of a device, so as to obtain a dynamic entropy loss index; and comparing the entropy loss index with a preset threshold value, and performing adaptive bias current feedback adjustment or triggering hardware fusing according to the comparison result. The present application solves the problem of implicit degradation of an entropy source caused by drift of a detector operating point, and realizes precise closed-loop control of the quality of the entropy source.

Description

Technical Field

[0001] This invention relates to the field of commercial cryptographic system testing technology. More specifically, this invention relates to a method for detecting the entropy source quality of commercial cryptographic systems based on quantum random number generators. Background Technology

[0002] In ensuring the security of commercial cryptographic systems, high-quality random number sequences are the core foundation for encryption protocols, key distribution, and authentication, and are typically generated directly by random number generators. With the continuous improvement of computing power, traditional pseudo-random number generators, due to the deterministic nature of their algorithms, are no longer sufficient to meet the requirements of high-security communication. Therefore, quantum random number generators, which utilize the inherent uncertainty of quantum mechanics, have attracted widespread attention from researchers.

[0003] QRNGs typically utilize photo-quantum detectors (PQDs) to capture the intrinsic randomness of photon arrival time, or measure shot noise generated by charge fluctuations as the primary entropy source. However, in real-world industrial applications, the physical state of PQDs is highly susceptible to factors such as ambient temperature fluctuations, power supply ripple, and device aging, leading to a slow shift in the statistical characteristics of their output signal. Existing entropy source quality inspection methods mostly focus on statistical verification after post-processing of the finished binary sequence.

[0004] This detection method, based on backend data, often suffers from time delays, making it difficult to directly reflect the real-time health of the front-end physical entropy source, and it cannot compensate for drifts in underlying physical parameters in real time. When the detector's operating point deviates from the optimal linear region, classical thermal noise may be mixed into the shot noise, or short-range correlations may occur due to circuit bandwidth limitations, thereby reducing entropy production capacity and posing potential security risks to the cryptographic system. Therefore, how to evaluate the physical purity of the quantum entropy source in real time under complex operating conditions and provide effective monitoring and intervention for its performance degradation has become a key aspect of ensuring the high reliability of cryptographic devices. Summary of the Invention

[0005] To address the technical problem of quantum entropy sources being prone to quality degradation due to environmental and aging factors and lacking real-time closed-loop control methods, this invention provides a method for entropy source quality detection in commercial cryptographic systems based on a quantum random number generator. The method includes: oversampling the shot noise output from an optical quantum detector at a set sampling rate to obtain the original noise sequence and preprocessing it to obtain a quantum fluctuation sequence; obtaining a correlation degree using an exponential function mapping based on the ratio of the observed variance of the quantum fluctuation sequence to the variance of the reference quantum noise, combined with the short-range autocorrelation of the quantum fluctuation sequence; obtaining an entropy loss exponent using a hyperbolic function and a logarithmic function based on the correlation degree and the deviation of the bias voltage from the target voltage; obtaining a bias command based on the entropy loss exponent and the bias voltage; and realizing entropy source quality detection and control based on the entropy loss exponent and the bias command.

[0006] This invention obtains the quantum fluctuation sequence and the correlation degree and entropy loss index by oversampling the shot noise output of the optical quantum detector, and then adjusts the injected current in real time or cuts off the output path, thereby realizing dynamic evaluation and feedback management of the quantum entropy source quality during the operation of commercial cryptographic systems.

[0007] Preferably, the step of oversampling the shot noise output by the optical quantum detector at a set sampling rate to obtain the original noise sequence includes: using... The original noise sequence is obtained by sampling the shot noise output by the photonic quantum detector at a certain sampling rate.

[0008] This invention uses a sampling rate of one gigahertz per second to collect the shot noise output by the optical quantum detector to obtain the original noise sequence, which ensures the capture of nanosecond-level micro-fluctuation characteristics and reduces the loss of quantum randomness due to insufficient sampling bandwidth.

[0009] Preferably, the preprocessing to obtain the quantum fluctuation sequence includes: using a window length of... The moving average subtraction algorithm at a certain point removes the common-mode interference of the original noise sequence to obtain the quantum fluctuation sequence.

[0010] This invention utilizes a moving average subtraction algorithm to process the original noise sequence to obtain a quantum fluctuation sequence with zero mean, thereby removing DC drift in the signal and reducing the impact of common-mode interference on the purity of the entropy source in commercial cryptographic systems.

[0011] Preferably, the correlation degree satisfies the expression: In the formula, express The degree of correlation at any given moment; express The variance observations of the quantum fluctuation sequence window at any given time; Represents the reference quantum noise variance; express The first time in the quantum fluctuation sequence window Voltage amplitude at each point; express The first time in the quantum fluctuation sequence window Voltage amplitude at each point; Indicates the autocorrelation delay step size; Indicates the zero constant; Indicates the sensitivity coefficient; Indicates the amplification factor; Represents the natural exponential function; Represents the absolute value symbol.

[0012] Preferably, the reference quantum noise variance is obtained by calibrating at the optimal time before the system leaves the factory: placing the device in a standard temperature chamber, adjusting the bias voltage to the target voltage, and continuously collecting the quantum fluctuation sequence for 1 minute in an environment without external strong light interference, and calculating the average value of its variance as the reference quantum noise variance.

[0013] Preferably, the sensitivity coefficient is set to .

[0014] Preferably, the entropy loss exponent satisfies the expression: In the formula, express The entropy loss index at time t; express The degree of correlation at any given moment; express The bias voltage at any given time; Indicates the target voltage; This represents the voltage half-width of the linear operating region; Indicates the potential energy index; Indicates the attenuation coefficient; express The system's cumulative runtime at any given time; Indicates the detector's design life; Represents the hyperbolic cosine function; This represents the natural logarithm function.

[0015] This invention combines correlation degree and the degree of deviation of bias voltage from target voltage with reference to cumulative operating time to obtain entropy loss index, and evaluates the random output state of optical quantum detectors throughout their entire life cycle, reflecting the cumulative impact of physical device aging and operating point drift on entropy source performance.

[0016] Preferably, the detector's design life is derived from the reliability data sheet provided by the detector manufacturer.

[0017] Preferably, the bias instruction satisfies the expression: In the formula, express The offset instruction at the moment; Indicates the initial bias current; Indicates proportional gain; Represents a symbolic function; express The bias voltage at any given time; Indicates the target voltage; Represents the absolute value symbol.

[0018] This invention utilizes the square root of the voltage deviation to obtain the bias command and performs negative feedback adjustment on the bias current injected into the photonic quantum detector. When the deviation is large, the possibility of system oscillation is reduced by gain attenuation, ensuring that the operating point of the commercial cryptographic system can be stably locked at the center position.

[0019] Preferably, the entropy source quality detection and control based on entropy loss index and bias command includes: adjusting the bias current of the injected photonic quantum detector using bias command in response to the entropy loss index being lower than a preset warning threshold and higher than a preset blocking threshold; and cutting off the quantum entropy source output path and switching to a backup pseudo-random source in response to the entropy loss index being lower than the blocking threshold.

[0020] The beneficial effects of this invention are as follows:

[0021] This invention achieves direct evaluation of the entropy source operation status of commercial cryptographic systems at the physical level by acquiring the shot noise characteristics of optical quantum detectors and obtaining the entropy loss index, thereby reducing the risk of implicit randomness failure caused by environmental fluctuations.

[0022] This invention utilizes nonlinear functions to simulate the dynamic process of detector performance degradation and obtains an evaluation index that reflects the decline in entropy production capability, providing a judgment criterion with physical mechanism support for the security defense mechanism of commercial cryptographic systems.

[0023] This invention improves the fault tolerance of commercial cryptographic systems under complex operating conditions by executing hierarchical control logic that includes parameter compensation and hardware circuit breaking, adjusting the bias current to return the entropy source performance to the center of the linear working area when there is a slight deviation. Attached Figure Description

[0024] Figure 1 This is a flowchart illustrating the entropy source quality detection method for commercial cryptographic systems based on quantum random number generators in this invention;

[0025] Figure 2 This is a schematic diagram illustrating the changes in the correlation degree of a quantum fluctuation sequence;

[0026] Figure 3 This is a schematic diagram illustrating the dynamic monitoring of the entropy loss index;

[0027] Figure 4 This is a schematic diagram illustrating the effect of adaptive feedback adjustment of bias voltage. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0030] This invention discloses a method for detecting the entropy source quality of commercial cryptographic systems based on quantum random number generators, referring to... Figure 1 This includes steps S1 to S4:

[0031] S1. Oversample the shot noise output by the optical quantum detector at a set sampling rate to obtain the original noise sequence and perform preprocessing to obtain the quantum fluctuation sequence.

[0032] It should be noted that quantum random number generators rely on the intrinsic uncertainty of photon arrival time to generate entropy. Their core signal originates from the shot noise of the optical quantum detector. The microscopic fluctuations of this noise are typically on the nanosecond scale. Ordinary low-speed acquisition links or circuits with low-pass filtering characteristics smooth out these crucial high-frequency quantum features, causing the acquired signal to degenerate into classical thermal noise. Therefore, this invention requires directly acquiring the original signal through a high-bandwidth signal link and removing common-mode interference to preserve pure quantum properties.

[0033] Specifically, the present invention uses The sampling rate is used to oversample the shot noise output of the optical quantum detector to obtain the original noise sequence. A window length of [value missing] is then used. The moving average subtraction method is used to process the original noise sequence, that is, to subtract the mean within the window from the current sampling point, remove common-mode interference and DC drift, and obtain a quantum fluctuation sequence with zero mean.

[0034] S2. Based on the ratio of the variance observation value of the quantum fluctuation sequence to the baseline quantum noise variance, and combined with the short-range autocorrelation of the quantum fluctuation sequence, the correlation degree is obtained by using an exponential function mapping.

[0035] It should be noted that an ideal quantum entropy source should possess perfect memorylessness, meaning the current signal state is completely independent of its historical state. However, when the operating point of a photonic quantum detector drifts into the nonlinear region due to temperature or aging, a weak short-range autocorrelation occurs within the signal. This autocorrelation signifies the loss of quantum randomness. To accurately capture this subtle degradation, this invention employs a model based on an exponential decay mechanism. Utilizing the nonlinear amplification properties of the exponential function, it imposes a severe numerical penalty on any minute anomalies in autocorrelation, thereby transforming the implicit mechanistic failure into an explicit numerical decrease.

[0036] Specifically, this invention calculates the correlation degree that characterizes the purity of a sequence based on quantum fluctuation sequences, and the correlation degree satisfies the expression:

[0037]

[0038] In the formula, express The degree of correlation at any given moment; express The variance observations of the quantum fluctuation sequence window at any given time; Represents the reference quantum noise variance; express The first time in the quantum fluctuation sequence window Voltage amplitude at each point; express The first time in the quantum fluctuation sequence window Voltage amplitude at each point; Indicates the autocorrelation delay step size; Indicates the zero constant; Indicates the sensitivity coefficient; Indicates the amplification factor; Represents the natural exponential function; Represents the absolute value symbol.

[0039] In the formula, the variance ratio term on the left side reflects the relative level of the current noise intensity; the exponential term on the right side acts as a purity filter. When non-zero short-range autocorrelation appears in the quantum fluctuation sequence, the autocorrelation term is amplified by the exponential term. When amplified by a power, as the independent variable of a negative exponent, it forces... The function's output value drops sharply and non-linearly, leading to a decrease in the final correlation. This reduces the sensitivity of non-quantum memory effects, thereby enabling sensitive detection.

[0040] It should be further added that the present invention will use the autocorrelation delay step size Set as To detect the non-quantum memory effect of nearest neighbors; to prevent zero constant Values Sensitivity coefficient Set as Amplification index Set as . The settings effectively suppress random disturbances caused by background thermal noise, ensuring that the detection results primarily reflect changes in quantum properties. This invention places the equipment in a standard temperature chamber before the system leaves the factory. In this process, the bias voltage is adjusted to the target voltage, and under conditions without external strong light interference, a quantum fluctuation sequence is continuously collected for 1 minute. The average value of its variance is calculated as the reference quantum noise variance. .

[0041] For example, Figure 2 This is a schematic diagram illustrating the changes in correlation of a quantum fluctuation sequence. The diagram shows that in the initial stage, the quantum entropy source maintains good memorylessness, and the correlation exhibits normal statistical fluctuations around a baseline value. As the sampling process progresses, the detector's operating point deviates due to environmental influences, leading to short-range autocorrelation in the sequence. Because an exponential function-based mapping mechanism is used, the correlation is significantly penalized by these subtle changes and decreases rapidly, demonstrating a high sensitivity to the loss of quantum properties.

[0042] S3. Based on the correlation, and combined with the degree of deviation of the bias voltage from the target voltage, the entropy loss index is obtained using hyperbolic and logarithmic functions.

[0043] It should be noted that the operating state drift of a photonic quantum detector exhibits nonlinear characteristics similar to potential well escape in physics. Specifically, in the initial stage when the bias voltage deviates from the center point, the entropy decays relatively slowly, but once the deviation exceeds the critical boundary of the linear region, the entropy drops abruptly. Furthermore, device aging also leads to an increase in background noise. To accurately assess the current effective entropy production capability, this invention introduces a hyperbolic cosine function to simulate this nonlinear saturation cutoff characteristic and combines this with runtime to evaluate the impact of aging.

[0044] Specifically, this invention calculates the entropy loss index in conjunction with the current bias voltage, and the entropy loss index satisfies the expression:

[0045]

[0046] In the formula, express The entropy loss index at time t; express The degree of correlation at any given moment; express The bias voltage at any given time; Indicates the target voltage; This represents the voltage half-width of the linear operating region; Indicates the potential energy index; Indicates the attenuation coefficient; express The system's cumulative runtime at any given time; Indicates the detector's design life; Represents the hyperbolic cosine function; This represents the natural logarithm function.

[0047] In the formula, when When the bias voltage is within the linear operating region near the target voltage (i.e., the deviation is less than half the voltage width), the hyperbolic cosine function is close to 1, keeping the entropy loss exponent at a high level. Once the bias voltage deviation exceeds half the voltage width, the hyperbolic cosine function increases exponentially, causing the first part of the value to decay rapidly. This simulates the sharp drop in entropy after quantum shot noise enters the saturation region, with the potential energy exponent determining the steepness of this drop. The logarithmic term gradually increases with the cumulative system operating time, thus lowering the final entropy loss exponent. This reflects the irreversible loss of entropy source quality due to the rise in background noise caused by device aging over time.

[0048] It should be further added that, in this invention, the potential energy index... Set as To ensure that the model accurately reflects the physical saturation characteristics of the detector; attenuation coefficient Set as This ensures that aging factors have a cumulative weighted impact on the entropy loss index throughout the entire life cycle. and It is determined during the factory calibration stage based on the voltage-entropy characteristic curve (V-ECurve) of the optical quantum detector; the detector's design life comes from the reliability data sheet provided by the detector manufacturer.

[0049] For example, Figure 3 This is a schematic diagram illustrating the dynamic monitoring of the entropy loss index. The diagram shows the evolution of the entropy loss index throughout the entire testing cycle. The entropy loss index comprehensively assesses the cumulative effects of physical correlation decay, voltage drift, and long-term device aging. When the bias voltage deviates from the linear operating region and the correlation decreases, the entropy loss index shows a significant downward trend and approaches the warning threshold. Subsequently, under the action of feedback control, as the physical state of the detector is repaired, the index rebounds and stabilizes within the safe range, ensuring the quality of random output from commercial cryptographic systems.

[0050] S4. Obtain the bias command based on the entropy loss index and bias voltage, and realize entropy source quality detection and control based on the entropy loss index and bias command.

[0051] It should be noted that, to ensure the security of commercial cryptographic systems, active intervention must be carried out before the entropy source completely loses its quantum properties. This invention establishes a hierarchical closed-loop control mechanism based on the entropy loss exponent. Through a nonlinear feedback adjustment expression, the system can prevent overshoot when the bias deviation is large and achieve precise locking when the deviation is small, thereby stabilizing the detector's operating point within the optimal linear region.

[0052] Specifically, this invention compares the entropy loss index with preset warning thresholds and blocking thresholds. When the entropy loss index is within the warning range, a bias command applied to the detector is calculated, the bias command satisfying the expression:

[0053]

[0054] In the formula, express Offset instruction at time (unit: ); Indicates the initial bias current (unit: ); Represents proportional gain (unit: ); This represents a sign function, meaning it takes the value 1 when the independent variable is greater than 0, -1 when the independent variable is less than 0, and 0 when the independent variable is equal to 0. express Bias voltage at time (unit: ); Indicates target voltage (unit: ); Represents the absolute value symbol.

[0055] In the formula, the sign function determines the direction of current adjustment, ensuring that the current is always adjusted in the direction that reduces voltage deviation; the square root term in the expression... It provides nonlinear gain characteristics, ensuring that when the deviation between the bias voltage and the target voltage is small, the adjustment amount is relatively large, which can provide sufficient strength to accurately lock the operating point at the center; when the deviation is large, the square root term limits the growth rate of the adjustment amplitude, preventing the system from overshooting or oscillating due to drastic changes in current. The proportional gain is used to balance the physical dimensions at both ends of the expression and control the overall adjustment intensity.

[0056] Furthermore, the present invention implements a hierarchical control strategy. When the entropy loss index is lower than the warning threshold and higher than the blocking threshold, the bias current injected into the quantum detector is changed by using the calculated bias command, and the bias voltage of the detector is physically adjusted so that it returns to the center of the linear working region. When the entropy loss index is lower than the blocking threshold, the entropy source is determined to be physically failed, and the hardware fuse logic is directly triggered to cut off the output path of the quantum entropy source and switch to the backup pseudo-random source.

[0057] It should be further noted that the warning threshold in this invention is set as follows: The blocking threshold is set to This strategy ensures that the system can identify the phase transition process from quantum supply to circuit saturation and take timely measures; This ensures that the detector reaches the target voltage at the initial moment. The required drive current is determined by open-loop scanning during the system power-on initialization phase. For proportional gain Its value determines the response speed and stability of the feedback regulation. If... If the value is too large, the system is prone to oscillation around the target point; if... If the value is too small, thermal drift cannot be suppressed in time. This invention uses the Ziegler-Nichols method for tuning and considers the nonlinear gain characteristics of the square root term to ultimately determine the proportional gain. This means that when the voltage deviation is At that time, the feedback current adjustment amount is This is sufficient to cover the typical thermal drift range.

[0058] For example, Figure 4 This diagram illustrates the effect of adaptive feedback adjustment of the bias voltage. The figure shows the process where the bias voltage gradually deviates from the target voltage due to thermal drift during operation of the quantum detector. After triggering the hierarchical control logic, the system uses bias commands to perform negative feedback adjustment on the injected current. Thanks to its nonlinear gain characteristics, the bias voltage can quickly converge towards the center point when the deviation is large, and remains stable when approaching the target value, ultimately reducing the implicit degradation risk to the entropy source caused by operating point drift.

Claims

1. A method for detecting the entropy source quality of a commercial cryptographic system based on a quantum random number generator, characterized in that, include: The shot noise output by the optical quantum detector is oversampled at a set sampling rate to obtain the original noise sequence and then preprocessed to obtain the quantum fluctuation sequence. Based on the ratio of the observed variance of the quantum fluctuation sequence to the variance of the baseline quantum noise, and combined with the short-range autocorrelation of the quantum fluctuation sequence, the correlation degree is obtained using an exponential function mapping. ; In the formula, express The degree of correlation at any given moment; express The variance observations of the quantum fluctuation sequence window at any given time; Represents the reference quantum noise variance; express The first time in the quantum fluctuation sequence window Voltage amplitude at each point; express The first time in the quantum fluctuation sequence window Voltage amplitude at each point; Indicates the autocorrelation delay step size; Indicates the zero constant; Indicates the sensitivity coefficient; Indicates the amplification factor; Represents the natural exponential function; Indicates the absolute value symbol; Based on the aforementioned correlation, and considering the degree of deviation of the bias voltage from the target voltage, the entropy loss exponent is obtained using hyperbolic and logarithmic functions. ; In the formula, express The entropy loss index at time step; express The bias voltage at any given time; Indicates the target voltage; This represents the voltage half-width of the linear operating region; Indicates the potential energy index; Indicates the attenuation coefficient; express The system's cumulative runtime at any given time; Indicates the detector's design life; Represents the hyperbolic cosine function; Represent the natural logarithm function; The bias command is obtained based on the entropy loss index and the bias voltage, and the entropy source quality detection and control are realized based on the entropy loss index and the bias command.

2. The entropy source quality detection method for commercial cryptographic systems based on quantum random number generators according to claim 1, characterized in that, The process of oversampling the shot noise output by the optical quantum detector at a set sampling rate to obtain the original noise sequence includes: by The original noise sequence is obtained by sampling the shot noise output by the photonic quantum detector at a certain sampling rate.

3. The entropy source quality detection method for commercial cryptographic systems based on quantum random number generators according to claim 1, characterized in that, The preprocessing to obtain the quantum fluctuation sequence includes: Using window length The moving average subtraction algorithm at a certain point removes the common-mode interference of the original noise sequence to obtain the quantum fluctuation sequence.

4. The entropy source quality detection method for commercial cryptographic systems based on quantum random number generators according to claim 1, characterized in that, The reference quantum noise variance is obtained as follows: Calibration is performed at the optimal time before the system leaves the factory: the equipment is placed in a standard temperature chamber, the bias voltage is adjusted to the target voltage, and the quantum fluctuation sequence is continuously collected for 1 minute in an environment without external strong light interference. The average value of its variance is calculated as the benchmark quantum noise variance.

5. The entropy source quality detection method for commercial cryptographic systems based on quantum random number generators according to claim 1, characterized in that, The sensitivity coefficient is set to .

6. The entropy source quality detection method for commercial cryptographic systems based on quantum random number generators according to claim 1, characterized in that, The detector's design life is derived from the reliability data sheet provided by the detector manufacturer.

7. The entropy source quality detection method for commercial cryptographic systems based on quantum random number generators according to claim 1, characterized in that, The bias instruction satisfies the expression: ； In the formula, express The offset instruction at the moment; Indicates the initial bias current; Indicates proportional gain; Represents a symbolic function; express The bias voltage at any given time; Indicates the target voltage; Represents the absolute value symbol.

8. The entropy source quality detection method for commercial cryptographic systems based on quantum random number generators according to claim 1, characterized in that, The method for detecting and controlling entropy source quality based on entropy loss index and bias command includes: In response to the entropy loss index being lower than a preset warning threshold and higher than a preset blocking threshold, the bias current injected into the photonic quantum detector is adjusted using a bias command; in response to the entropy loss index being lower than the blocking threshold, the output path of the quantum entropy source is cut off and switched to a backup pseudo-random source.

Images