Field programmable gate array (FPGA)-based quantum time sequence control method

Abstract

The invention discloses a quantum time sequence control method based on an FPGA (Field Programmable Gate Array), which relates to the technical field of quantum, and comprises the following steps: S1, setting an initial time sequence representing the control parameter estimation implementation size of an initial control signal; s2, performing Fourier transform on the initial control signal along an initial time sequence to obtain a frequency spectrum data distribution map for screening out a predicted abnormal time period; s3, processing the initial control signal by using a Kalman filter to generate a correction time sequence; and S4, using a quantum state evaluation method to evaluate whether the current correction time sequence satisfies electrical noise filtering, if so, completing electrical noise processing of the target quantum system, and if not, returning to the step S2. According to the method, the spectrum characteristics of the electrical noise are identified and analyzed in real time through Fourier transform, the influence of the noise on quantum operation is minimized, and the method has the advantages that the electrical noise is reduced more accurately, and the self-adaptive capacity of the system is improved.

Description

Technical Field

[0001] The present invention relates to the field of quantum technology, and in particular to a quantum timing control method based on FPGA. Background Art

[0002] Quantum timing control refers to the process of achieving the desired quantum state and operation by precisely controlling and managing the evolution of quantum systems over time during quantum computing and quantum information processing, ensuring that quantum gates (basic quantum operation units) are accurately applied in a predetermined order and time. Quantum operations are achieved by applying electromagnetic pulses. The timing and intensity of these pulses are crucial to the control of quantum systems. Designing a suitable pulse sequence to achieve the desired quantum gate operation is one of the core tasks of quantum timing control. FPGA is a programmable logic device with high parallel processing capabilities and flexibility. It is suitable for quantum timing control and can achieve precise pulse sequence generation, clock management and real-time feedback control. Timing control of quantum systems on the FPGA platform can process a large number of signals and computing tasks in parallel, which is very beneficial for controlling multiple quantum bits at the same time and applying complex pulse sequences; at the same time, FPGA can be reprogrammed according to needs, which allows flexible adjustment and optimization of quantum timing control, and provides high-precision clock management functions, which can generate stable and accurate clock signals, suitable for applications requiring precise timing control.

[0003] When the existing quantum timing control is operated on the FPGA platform, due to the complex digital circuit structure and high-speed signal processing inside the FPGA, various electrical noises may be introduced into the quantum system. The main noise sources include internal power supply noise, clock signal noise and switch switching noise. The impact of electrical noise on quantum timing control mainly includes causing control signal distortion, causing timing signal deviation and interfering with the synchronization accuracy between modules inside the FPGA, which affects the accuracy and stability of pulses in the quantum system and interferes with the accurate operation and coherence time of quantum bits. At present, the main method adopted in the prior art for the electrical noise that is prone to occur when using the FPGA platform for quantum timing control is to use power supply filters and decoupling capacitors. Since the frequency response of the filter and decoupling capacitor must match the operating frequency and noise characteristics of the FPGA, the noise characteristics of the FPGA may change with time sequence. If the parameter size is not suitable, it may not be possible to effectively filter out all harmful noise frequency bands; at the same time, the filter and decoupling capacitor will occupy limited PCB space, affecting the compactness and layout design of the overall system. Therefore, a more effective method for real-time calibration and optimization of electrical noise is needed in the prior art. Summary of the invention

[0004] The present invention provides a quantum timing control method based on FPGA, which solves the problem in the prior art that it is difficult to perform accurate and real-time anti-interference processing on electrical noise when performing quantum timing control based on an FPGA platform.

[0005] The present invention is achieved through the following technical solutions: A quantum timing control method based on FPGA, the method comprising: Step S1: resetting the initial quantum parameters for completing the target quantum operation in the target quantum system, generating the initial control signal for executing the target quantum operation in the target FPGA system according to the initial quantum parameters, setting the initial time sequence representing the estimated implementation size of the control parameters of the initial control signal, and modulating the initial control signal to adapt to the target quantum system; Step S2: Performing Fourier transformation on the initial control signal along the initial time series to obtain a spectrum data distribution diagram, screening out all time series time points corresponding to the spectrum data indicating noise abnormality based on the abnormal spectrum characteristics of the electrical noise, and marking all the time series time points as predicted abnormal time periods in the form of setting adjacent time points as continuous time periods; Step S3: using a Kalman filter to perform distributed filtering processing on the initial control signal along the predicted abnormal period on the initial time series, updating the filtering results of each predicted abnormal period to the initial time series and generating a corrected time series, wherein the corrected time series represents the addition of compensation adjustment to the control parameters of the initial control signal on the basis of the initial time series; Step S4: In the target FPGA system, the initial control signal is subjected to the target quantum operation along the correction time sequence, and the quantum state evaluation method is used to evaluate whether the current correction time sequence satisfies the initial control signal to complete the electrical noise filtering. If so, the electrical noise processing of the target quantum system is completed; if not, the process returns to step S2.

[0006] Furthermore, the control parameters of the initial control signal include time characteristic parameters and periodic characteristic parameters, and a set of time characteristic parameters and periodic characteristic parameters are contained in the discrete variables at each time node in the initial time series and the correction time series; the time characteristic parameters represent the pulse width and level conversion duration of the initial control signal at the current discrete variable; the periodic characteristic parameters represent the frequency and duty cycle of the current discrete variable of the initial control signal.

[0007] Further, the abnormal spectrum characteristics of the electrical noise are in the form of abnormal states of abnormal noise parameters of the electrical noise on a spectrum data distribution diagram; the abnormal noise parameters include spike noise, broadband noise and power supply ripple; The abnormal state of the spike noise is that there is instantaneous spike spectrum data in the spectrum data on the initial time series; The abnormal state of the broadband noise is that the spectrum data is evenly distributed in at least three consecutive time series; The abnormal state of the power supply ripple is that the ripple exists in a low frequency range in the spectrum data.

[0008] Furthermore, the process of the Kalman filter performing distributed processing on the initial control signal includes: At each predicted abnormal period, a filtering dynamic model of the Kalman filter is initialized based on the time characteristic parameters and the periodic characteristic parameters, a covariance matrix representing the randomness measurement of electrical noise and a Kalman gain algorithm based on the covariance matrix are preset in the filtering dynamic model, a state transfer matrix is ​​set to represent the state change of the initial control signal between adjacent time nodes, and a predictive correction update is performed on the covariance matrix at the current time node based on the abnormal noise parameters of the electrical noise and the predicted abnormal period; The time characteristic parameters and period characteristic parameters are calculated as input values ​​of the Kalman gain algorithm, and the output results of the calculation are set as the time characteristic parameters and period characteristic parameters representing the correction at the current time node; the time characteristic parameters and period characteristic parameters output by the Kalman gain algorithm are updated to each predicted abnormal time period on the initial time series, and used to generate a corrected time series.

[0009] Furthermore, the content of the predictive correction update of the covariance matrix includes: Let the covariance matrix be represented as P, and the time node position on the initial time series be represented as k. Then the covariance matrix at the current time node is represented as P k , the covariance matrix obtained by predicting the state of the current time node based on the previous time node is expressed as P k|k-1 ; Set the process noise matrix Q k Used to represent abnormal noise parameters in electrical noise and set the state transfer matrix F k It is used to indicate the state change of the initial control signal from time node k-1 to k. Then the updated covariance matrix P is k|k-1 The calculation formula is: , in Represents the covariance matrix P for the previous time node k-1 After the state transfer matrix F k The transformation of the predicted covariance matrix P k|k-1 , the process noise matrix Q k Represents the amplitude of the system process noise impact; the state transfer matrix F k The transpose of is used to ensure accuracy when updating the covariance matrix.

[0010] Furthermore, the Kalman gain algorithm is calculated based on an unscented Kalman filter and uses propagation state distribution sampling to improve accuracy.

[0011] Furthermore, the calculation process of the Kalman gain algorithm includes: The time characteristic parameters and period characteristic parameters at the target time node are set as the state variable parameters L of the Kalman filter, and the time characteristic parameters and period characteristic parameters are respectively brought into L for calculation; and the observation model H is set in the Kalman filter, and the observation data is represented as Z k , let the Kalman gain variable be U, and the measurement noise matrix be R, Using the state transition matrix F k Predict the state variable parameters of the current time node: , Use the observation model H and the measurement noise matrix R to calculate the Kalman gain variable U at the current time node: , Combined with the observed data, the corrected state variable parameter L is calculated k|k : .

[0012] Furthermore, the content of the quantum state evaluation method includes: presetting a reference quantum state for reference and an ideal fidelity value of the reference quantum state, using a correction time sequence to perform quantum operations in a target FPGA system, and then measuring the fidelity value of the quantum state, calculating the measurement fidelity value between the actual quantum state and the reference quantum state, and when the measurement fidelity value is greater than or equal to the ideal fidelity value, determining that the current correction time sequence has no electrical noise; when the measurement fidelity value is greater than or equal to the ideal fidelity value, determining that the current correction time sequence still has electrical noise.

[0013] Furthermore, the calculation method of the measurement fidelity value includes: presetting a reference density matrix of a reference quantum state, measuring the actual density matrix of the actual quantum state multiple times based on a quantum state tomography method, and obtaining the measurement fidelity value based on the similarity between the actual density matrix and the reference density matrix.

[0014] Furthermore, the actual density matrix is ​​a semi-positive definite matrix whose eigenvalues ​​are all negative.

[0015] Compared with the prior art, the present invention uses Fourier transform to identify and analyze the spectral characteristics of electrical noise in real time, and uses a Kalman filter to dynamically correct abnormal time periods to ensure that the impact of noise on quantum operations is minimized. It has the advantages of more accurately reducing the electrical noise of the FPGA platform and improving the adaptive ability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings: Figure 1 It is a flowchart of the present invention. DETAILED DESCRIPTION

[0017] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention. Example

[0018] like Figure 1 As shown, this embodiment is a quantum timing control method based on FPGA, and the method includes: Step S1: resetting the initial quantum parameters for completing the target quantum operation in the target quantum system, generating the initial control signal for executing the target quantum operation in the target FPGA system according to the initial quantum parameters, setting the initial time sequence representing the estimated implementation size of the control parameters of the initial control signal, and modulating the initial control signal to adapt to the target quantum system; Step S2: Performing Fourier transformation on the initial control signal along the initial time series to obtain a spectrum data distribution diagram, screening out all time series time points corresponding to the spectrum data indicating noise abnormality based on the abnormal spectrum characteristics of the electrical noise, and marking all the time series time points as predicted abnormal time periods in the form of setting adjacent time points as continuous time periods; Step S3: using a Kalman filter to perform distributed filtering processing on the initial control signal along the predicted abnormal period on the initial time series, updating the filtering results of each predicted abnormal period to the initial time series and generating a corrected time series, wherein the corrected time series represents the addition of compensation adjustment to the control parameters of the initial control signal on the basis of the initial time series; Step S4: In the target FPGA system, the initial control signal is subjected to the target quantum operation along the correction time sequence, and the quantum state evaluation method is used to evaluate whether the current correction time sequence satisfies the initial control signal to complete the electrical noise filtering. If so, the electrical noise processing of the target quantum system is completed; if not, the process returns to step S2.

[0019] The target quantum operation may include quantum gate operation, quantum state preparation and evolution, quantum entanglement and quantum correction in specific applications. The initial quantum parameters mainly include quantum bit state, quantum gate parameters and quantum operation sequence. The process of generating the initial control signal according to the initial quantum parameters is a conventional technology in the art, which mainly includes processes such as defining the quantum bit state, generating driving voltage and current, designing PFGA logic circuit and control signal verification. The initial time series indicates how to apply the amplitude and frequency of the control signal at different time points, for example, the initial control signal can be increased or decreased at a specific time point; it also indicates the preliminary estimated values ​​of signal modulation and control parameters, including a description of the change pattern of the control signal over time, including possible waveforms, periodic changes, amplitude changes, etc., and the estimated time sampling frequency of the control signal. The main process of this implementation method is: generate control signals according to preset initial quantum parameters, and set time series to modulate these signals to prepare for the operation of the quantum system; obtain the spectrum data of the signal through Fourier transform, identify the abnormal noise spectrum, and identify the abnormal period through the abnormal spectrum; use the Kalman filter to filter the predicted abnormal period to generate a correction time series to compensate for the influence of electrical noise; use the correction time series to perform quantum operations and evaluate whether the electrical noise is successfully filtered. If the requirements are not met, return to the second step for reprocessing. Using the Kalman filter to compensate for the control signal in real time can effectively cope with noise changes, enhance the system's adaptability to noise interference, and reduce the interference of electrical noise on the operation of the quantum system through precise noise processing, thereby improving the accuracy and reliability of quantum operations.

[0020] Furthermore, as a feasible implementation method, the control parameters of the initial control signal include time characteristic parameters and periodic characteristic parameters, and a set of time characteristic parameters and periodic characteristic parameters are included in the discrete variables at each time node in the initial time series and the correction time series; the time characteristic parameters represent the pulse width and level conversion duration of the initial control signal at the current discrete variable; the periodic characteristic parameters represent the frequency and duty cycle of the current discrete variable of the initial control signal.

[0021] More preferably, the abnormal spectrum characteristics of the electrical noise are in the form of abnormal states of abnormal noise parameters of the electrical noise on a spectrum data distribution diagram; the abnormal noise parameters include spike noise, broadband noise and power supply ripple; The abnormal state of the spike noise is that there is instantaneous spike spectrum data in the spectrum data on the initial time series; The abnormal state of the broadband noise is that the spectrum data is evenly distributed in at least three consecutive time series; The abnormal state of the power supply ripple is that the ripple exists in a low frequency range in the spectrum data.

[0022] The pulse width refers to the duration of the control signal within a single pulse cycle, which affects the time resolution of the signal and the accuracy of quantum operations. The level conversion duration refers to the time required for the control signal to switch from one level to another, which affects the rise time and fall time of the signal. The duty cycle in the periodic characteristic parameter indicates the proportion of the high level of the signal in one cycle, which affects the energy distribution and average power of the signal. The spike noise usually represents short-term high-intensity interference, which may have an instantaneous impact on the operation of the quantum system; after Fourier transform, the prominent peaks in the spectrum graph are observed and these high-frequency spikes are identified, indicating abnormal spike noise. The broadband noise means that the spectrum data is evenly distributed over at least three consecutive time series, showing as noise in a wide frequency range. Broadband noise covers a wide frequency range and usually affects all parts of the system; a wide and flat noise distribution in the frequency range is observed in the spectrum graph during detection, indicating that the system is widely interfered within these frequency ranges. The power ripple indicates that there are regular ripples in the low-frequency range of the spectrum data, which usually represent low-frequency periodic interference. The power ripple is usually caused by the instability of the power supply. During the detection, the periodic fluctuations in the low-frequency area of ​​the spectrum diagram are mainly observed, and these features are identified to indicate the existence of power ripple. By defining the time characteristic parameters and periodic characteristic parameters of the initial control signal in detail, and combining the specific abnormal form of electrical noise, the influence of noise on quantum timing control can be accurately detected and corrected.

[0023] Further, as a feasible implementation manner, the process of the Kalman filter performing distributed processing on the initial control signal includes: At each predicted abnormal period, a filtering dynamic model of the Kalman filter is initialized based on the time characteristic parameters and the periodic characteristic parameters, a covariance matrix representing the randomness measurement of electrical noise and a Kalman gain algorithm based on the covariance matrix are preset in the filtering dynamic model, a state transfer matrix is ​​set to represent the state change of the initial control signal between adjacent time nodes, and a predictive correction update is performed on the covariance matrix at the current time node based on the abnormal noise parameters of the electrical noise and the predicted abnormal period; The time characteristic parameters and period characteristic parameters are calculated as input values ​​of the Kalman gain algorithm, and the output results of the calculation are set as the time characteristic parameters and period characteristic parameters representing the correction at the current time node; the time characteristic parameters and period characteristic parameters output by the Kalman gain algorithm are updated to each predicted abnormal time period on the initial time series, and used to generate a corrected time series.

[0024] The time characteristic parameters and period characteristic parameters describe the time characteristics of the control signal. In the filtering dynamic model, these parameters are initialized to accurately simulate the behavior and changes of the control signal. The covariance matrix is ​​used to measure the randomness of the noise, that is, the variance and correlation of the noise in the signal. The preset covariance matrix allows the filter to understand the characteristics of the noise so as to make effective estimates. The Kalman gain calculation formula based on the covariance matrix is ​​used to optimize the performance of the filter and determine how to combine new observation data with previous estimates to minimize the estimation error. The state transfer matrix is ​​used to describe the state change of the initial control signal between adjacent time nodes, which is used to help the filter predict future states and adjust the control signal. According to the output calculated by the Kalman gain algorithm, the time characteristic parameters and period characteristic parameters are updated to the current time node respectively to reflect the corrected control signal characteristics. The Kalman filter is used to estimate the system state in time series data. It minimizes the error variance by combining model prediction and actual observation. As a specific application, preferably, the Kalman gain algorithm is calculated based on an unscented Kalman filter, and propagation state distribution sampling is used to improve accuracy.

[0025] Further, as a feasible implementation, the content of the predictive correction update of the covariance matrix includes: Let the covariance matrix be represented as P, and the time node position on the initial time series be represented as k. Then the covariance matrix at the current time node is represented as P k , the covariance matrix obtained by predicting the state of the current time node based on the previous time node is expressed as P k|k-1 ; Set the process noise matrix Q k Used to represent abnormal noise parameters in electrical noise and set the state transfer matrix F k It is used to indicate the state change of the initial control signal from time node k-1 to k. Then the updated covariance matrix P is k|k-1 The calculation formula is: , in Represents the covariance matrix P for the previous time node k-1 After the state transfer matrix F k The transformation of the predicted covariance matrix P k|k-1 , the process noise matrix Q k Represents the amplitude of the system process noise impact; the state transfer matrix F k The transpose of is used to ensure accuracy when updating the covariance matrix.

[0026] The covariance matrix P k Represents the variance and covariance of the estimated error of the control signal at the current moment. The covariance matrix Pk|k-1 1 represents the prediction covariance matrix of the current time node obtained based on the state prediction of the previous time node, reflecting the estimation of the prediction error. The state transfer matrix F k It is used to predict the error estimate of the current time node by applying the covariance matrix of the previous time node, describing the state change of the initial control signal from time node k−1 to time node k, and is used to transform the covariance matrix of the previous time node to the current time node. The noise matrix Q k The matrix representing the process noise contains abnormal noise parameters, which reflects the influence of electrical noise and is used to adjust the uncertainty in state prediction. Indicates that the covariance matrix P of the previous time node k-1 Through the state transfer matrix F k Transformed to the prediction covariance matrix P of the current time node k|k-1 The state transfer matrix F k Used to describe the dynamic changes of system states, the state transfer matrix F k The transpose of ensures the accuracy of the transformation. In the predictive correction update process of the covariance matrix, the state transfer matrix F k and the process noise matrix Q k The effective application of enables the filter to accurately predict the error at the current time node and adjust the control signal to adapt to the electrical noise. In this way, the Kalman filter realizes dynamic correction of the signal, enhances the noise processing ability of the system, and ensures the accuracy of quantum operations.

[0027] Further, as a feasible implementation, the calculation process of the Kalman gain algorithm includes: The time characteristic parameters and period characteristic parameters at the target time node are set as the state variable parameters L of the Kalman filter, and the time characteristic parameters and period characteristic parameters are respectively brought into L for calculation; and the observation model H is set in the Kalman filter, and the observation data is represented as Z k , let the Kalman gain variable be U, and the measurement noise matrix be R, Using the state transition matrix F k Predict the state variable parameters of the current time node: , Use the observation model H and the measurement noise matrix R to calculate the Kalman gain variable U at the current time node: , Combined with the observed data, the corrected state variable parameter L is calculated k|k : .

[0028] In the Kalman filter, both the time characteristic parameter and the period characteristic parameter are regarded as state variables, which are used to describe the characteristics of the control signal. k|k-1 represents the predicted state variable at the current time node k, based on the state variable parameter L at the previous time node k−1 k-1|k-1 The prediction covariance matrix P k|k-1 Represents the prediction covariance matrix of the current time node k. The observation model H k Represents the noise matrix R used to map the state variables to the observation data space at the current time node. k Represents the noise characteristics in the observed data. The state variable parameter L k|k It represents the final estimated value after combining the observed data. k ∙P k|k-1 Represents the state variable parameter L k|k-1 The calculated observation value. k ∙(Z k -H k ∙P k|k-1 ) represents the actual observation data Z k and the predicted observation value H k ∙P k|k-1 The difference (residual part) between them is applied to the predicted state to obtain the corrected state variable parameter L k|k In this scheme, the Kalman gain algorithm ensures accurate adjustment of the initial control signal through state variable prediction, gain calculation and state variable correction. The state transfer matrix and observation model are combined with the measurement noise matrix to achieve effective fusion of prediction and actual observation data through gain optimization, thereby improving the system's adaptability to noise and the accuracy of the control signal.

[0029] Furthermore, as a feasible implementation method, the content of the quantum state evaluation method includes: presetting a reference quantum state for reference and an ideal fidelity value of the reference quantum state, using a correction time sequence to perform quantum operations in a target FPGA system, and then measuring the fidelity value of the quantum state, calculating the measurement fidelity value between the actual quantum state and the reference quantum state, and when the measurement fidelity value is greater than or equal to the ideal fidelity value, determining that the current correction time sequence has no electrical noise; when the measurement fidelity value is greater than or equal to the ideal fidelity value, determining that the current correction time sequence still has electrical noise.

[0030] The reference quantum state is a preset quantum state for comparison with the actual quantum state, and the reference quantum state is an ideal target state or expected state. The fidelity value is used to quantify the similarity between the actual quantum state and the reference quantum state. The fidelity value is usually calculated by quantum state measurement and reconstruction methods, for example, usually calculated by the density matrix and fidelity formula of the quantum state, or using quantum state reconstruction technology to obtain the actual fidelity value of the quantum state. The ideal fidelity value represents the fidelity value between the actual quantum state and the reference quantum state under ideal conditions. Usually the ideal fidelity value is 1, indicating that the two are completely consistent. In practical applications, an acceptance range or threshold can be set. If the measured fidelity value of the actual quantum state reaches or exceeds the ideal fidelity value, it means that the influence of electrical noise on the quantum state has been effectively processed and the accuracy of quantum operation has been guaranteed; if the fidelity value fails to reach the ideal fidelity value, it means that the corrected time series is still affected by electrical noise, and the correction process may need to be further optimized. By comparing the fidelity values, the effect of correcting the time series can be effectively evaluated.

[0031] Further, as a feasible implementation method, the calculation method of the measurement fidelity value includes: presetting a reference density matrix of a reference quantum state, measuring the actual density matrix of the actual quantum state multiple times based on the quantum state tomography method, and obtaining the measurement fidelity value based on the similarity between the actual density matrix and the reference density matrix; wherein, the actual density matrix is ​​a semi-positive definite matrix in which all eigenvalues ​​are negative. Quantum state tomography is a method for quantizing quantum states. By measuring the quantum state multiple times, its density matrix is ​​reconstructed. The process involves the measurement of different bases and the statistical analysis of the measurement results. Fidelity quantifies the similarity between two quantum states. The actual density matrix may have negative eigenvalues ​​in quantum state tomography and needs to be adjusted to a semi-positive definite matrix. This is usually done by discarding negative eigenvalues ​​or adjusting them to zero. When using the formula to calculate the fidelity, ensuring that the actual density matrix is ​​a semi-positive definite matrix can more accurately determine whether the corrected time series effectively reduces electrical noise.

[0032] The reference density matrix is ​​a preset density matrix of an ideal quantum state, which represents the expected state of the target quantum state. The actual quantum state is measured multiple times by quantum state tomography to reconstruct its actual density matrix, and the quantum state tomography estimates the density matrix of the quantum state through a series of measured data.

[0033] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A quantum timing control method based on FPGA, characterized in that: The method includes: Step S1: resetting the initial quantum parameters for completing the target quantum operation in the target quantum system, generating the initial control signal for executing the target quantum operation in the target FPGA system according to the initial quantum parameters, setting the initial time sequence representing the estimated implementation size of the control parameters of the initial control signal, and modulating the initial control signal to adapt to the target quantum system; Step S2: Performing Fourier transformation on the initial control signal along the initial time series to obtain a spectrum data distribution diagram, screening out all time series time points corresponding to the spectrum data indicating noise abnormality based on the abnormal spectrum characteristics of the electrical noise, and marking all the time series time points as predicted abnormal time periods in the form of setting adjacent time points as continuous time periods; Step S3: using a Kalman filter to perform distributed filtering processing on the initial control signal along the predicted abnormal period on the initial time series, updating the filtering results of each predicted abnormal period to the initial time series and generating a corrected time series, wherein the corrected time series represents the addition of compensation adjustment to the control parameters of the initial control signal on the basis of the initial time series; Step S4: In the target FPGA system, the initial control signal is subjected to the target quantum operation along the correction time sequence, and the quantum state evaluation method is used to evaluate whether the current correction time sequence satisfies the initial control signal to complete the electrical noise filtering. If so, the electrical noise processing of the target quantum system is completed; if not, the process returns to step S2.

2. The FPGA-based quantum timing control method according to claim 1, characterized in that: The control parameters of the initial control signal include time characteristic parameters and periodic characteristic parameters, and a set of time characteristic parameters and periodic characteristic parameters are contained in the discrete variables at each time node in the initial time series and the correction time series; the time characteristic parameters represent the pulse width and level conversion duration of the initial control signal at the current discrete variable; the periodic characteristic parameters represent the frequency and duty cycle of the current discrete variable of the initial control signal.

3. The FPGA-based quantum timing control method according to claim 1, characterized in that: The abnormal spectrum characteristics of the electrical noise are in the form of abnormal states of abnormal noise parameters of the electrical noise on a spectrum data distribution diagram; the abnormal noise parameters include spike noise, broadband noise and power supply ripple; The abnormal state of the spike noise is that there is instantaneous spike spectrum data in the spectrum data on the initial time series; The abnormal state of the broadband noise is that the spectrum data is evenly distributed in at least three consecutive time series; The abnormal state of the power supply ripple is that the ripple exists in a low frequency range in the spectrum data.

4. A quantum timing control method based on FPGA according to any one of claims 2 to 3, characterized in that: The process of the Kalman filter performing distributed processing on the initial control signal includes: At each predicted abnormal period, a filtering dynamic model of the Kalman filter is initialized based on the time characteristic parameters and the periodic characteristic parameters, a covariance matrix representing the randomness measurement of electrical noise and a Kalman gain algorithm based on the covariance matrix are preset in the filtering dynamic model, a state transfer matrix is ​​set to represent the state change of the initial control signal between adjacent time nodes, and a predictive correction update is performed on the covariance matrix at the current time node based on the abnormal noise parameters of the electrical noise and the predicted abnormal period; The time characteristic parameters and period characteristic parameters are calculated as input values ​​of the Kalman gain algorithm, and the output results of the calculation are set as the time characteristic parameters and period characteristic parameters representing the correction at the current time node; the time characteristic parameters and period characteristic parameters output by the Kalman gain algorithm are updated to each predicted abnormal time period on the initial time series, and used to generate a corrected time series.

5. The FPGA-based quantum timing control method according to claim 4, characterized in that: The content of the predictive correction update of the covariance matrix includes: Let the covariance matrix be represented as P, and the time node position on the initial time series be represented as k. Then the covariance matrix at the current time node is represented as P k , the covariance matrix obtained by predicting the state of the current time node based on the previous time node is expressed as P k|k-1 ; Set the process noise matrix Q k Used to represent abnormal noise parameters in electrical noise and set the state transfer matrix F k It is used to indicate the state change of the initial control signal from time node k-1 to k. Then the updated covariance matrix P is k|k-1 The calculation formula is: , in Represents the covariance matrix P for the previous time node k-1 After the state transfer matrix F k The transformation of the predicted covariance matrix P k|k-1 , the process noise matrix Q k Represents the amplitude of the system process noise impact; the state transfer matrix F k The transpose of is used to ensure accuracy when updating the covariance matrix.

6. The FPGA-based quantum timing control method according to claim 4, characterized in that: The Kalman gain algorithm is calculated based on an unscented Kalman filter and uses propagation state distribution sampling to improve accuracy.

7. The FPGA-based quantum timing control method according to claim 5, characterized in that: The calculation process of the Kalman gain algorithm includes: The time characteristic parameters and period characteristic parameters at the target time node are set as the state variable parameters L of the Kalman filter, and the time characteristic parameters and period characteristic parameters are respectively brought into L for calculation; and the observation model H is set in the Kalman filter, and the observation data is represented as Z k , let the Kalman gain variable be U, and the measurement noise matrix be R, Using the state transition matrix F k Predict the state variable parameters of the current time node: ， Use the observation model H and the measurement noise matrix R to calculate the Kalman gain variable U at the current time node: ， Combined with the observed data, the corrected state variable parameter L is calculated k|k : 。 8. The FPGA-based quantum timing control method according to claim 1, characterized in that: The quantum state evaluation method includes: presetting a reference quantum state for reference and an ideal fidelity value of the reference quantum state, performing quantum operations using a correction time sequence in a target FPGA system, and then measuring the fidelity value of the quantum state, calculating the measurement fidelity value between the actual quantum state and the reference quantum state, and when the measurement fidelity value is greater than or equal to the ideal fidelity value, judging that the current correction time sequence has no electrical noise; when the measurement fidelity value is greater than or equal to the ideal fidelity value, judging that the current correction time sequence still has electrical noise.

9. The FPGA-based quantum timing control method according to claim 8, characterized in that: The calculation method of the measurement fidelity value includes: presetting a reference density matrix of a reference quantum state, measuring the actual density matrix of the actual quantum state multiple times based on a quantum state tomography method, and obtaining the measurement fidelity value based on the similarity between the actual density matrix and the reference density matrix.

10. The FPGA-based quantum timing control method according to claim 9, characterized in that: The actual density matrix is ​​a semi-positive definite matrix whose eigenvalues ​​are all negative.

Images