Quantum control waveform overshoot correction method and system based on frequency response measurement, storage medium, and electronic device

Abstract

Disclosed in the present invention are a quantum control waveform overshoot correction method and system based on frequency response measurement, a storage medium, and an electronic device. The method comprises: setting a pulse parameter and an output mode, loading a pulse signal and outputting by playing same, and acquiring an output waveform and measuring a pulse response of a system; performing data transformation to obtain a frequency response function of the system, a frequency response function for frequency domain compensation, and a time domain waveform function; performing two convolution operations to obtain final correction data; and loading the correction data to correct the waveform, and performing parallel computation in conjunction with a programmable gate array. In the method, two convolution operations are performed to convert edge data into calibration data and then the calibration data is used to correct the waveform, and parallel computation is performed in conjunction with the programmable gate array, so that the corrected waveform is relatively smooth, avoiding the problems of missing edges or overcorrection caused by manual calibration, and effectively improving computational efficiency and calibration accuracy.

Description

A quantum-controlled waveform overshoot correction method, system, storage medium, and electronic device based on frequency response measurement. Technical Field

[0001] This invention relates to the field of quantum computer technology, and in particular to a quantum control waveform overshoot correction method and system based on frequency response measurement. Background Technology

[0002] Controlling the qubits of a superconducting quantum computer requires simultaneous operation through two physical ports: the XY control circuit and the Z control circuit. Microwave pulse signals are input through the XY circuit, while DC pulse signals are input through the Z circuit for bit control and testing. Quantum computing demands that the DC pulse signal in the Z circuit not only have very fast rise and fall times to ensure more qubit gates can be configured within a given decoherence time, but also extremely low overshoot at the pulse edges to guarantee the stability of the qubit resonant cavity frequency.

[0003] Therefore, it is necessary to generate fast and highly flat DC pulse signals to ensure the quality of qubit manipulation and testing. Existing correction methods involve performing pre-distortion processing in the time domain. This involves adjusting the values ​​of edge data points and observing the overshoot changes in the adjusted signal. After multiple adjustments, a DC pulse with smaller overshoot is obtained. Then, calibration data is calculated based on the obtained pulse rising edge data to perform digital pre-distortion processing on the edges of the pulse waveform to be played later. Since the overshoot size after pre-distortion processing depends on the calibration data, which is manually adjusted based on the waveform size displayed on the oscilloscope, multiple parameter adjustments are required when adjusting multiple parameters. These parameters have a synergistic effect during adjustment, necessitating repeated calibrations. This approach cannot guarantee optimal calibration results, has low computational efficiency, and involves a large computational load. Furthermore, when some pulses are not identified or are identified as pulses when they are not, there are problems of missed or over-correction. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a quantum-controlled waveform overshoot correction method and system based on frequency response measurement.

[0005] To achieve the above objectives, the present invention employs the following technical solution: a quantum-controlled waveform overshoot correction method based on frequency response measurement, comprising:

[0006] S1, set the pulse parameters and output mode of the quantum control signal generation system, load the pulse signal and output it for playback, and collect the output waveform to obtain the pulse response data of the quantum control signal generation system;

[0007] S2, perform multi-step data transformation processing on the data points of the system's impulse response to obtain the system's time-domain waveform function;

[0008] S3, set the pulse waveform function, perform the first convolution process on the time-domain waveform function and the pulse waveform function to convert it into calibration data, and perform the second convolution process on the pulse waveform function and the calibration data to obtain a corrected data;

[0009] S4: Load correction data to correct the waveform and perform parallel computation in conjunction with a programmable gate array.

[0010] As a further description of the above technical solution, step S2 includes:

[0011] S21, Perform Fourier transform on the impulse response data of the quantum control signal generation system to calculate the system's frequency response function Y(jω);

[0012] S22, taking the reciprocal of the system's frequency response function Y(jω), we calculate the frequency response function Z(jω) with frequency domain compensation. n );

[0013] S23, Frequency response function Z(jω) for frequency domain compensation n Perform an inverse Fourier transform to calculate the system's time-domain waveform function z(t). n ).

[0014] As a further description of the above technical solution, the frequency response function Y(jω) of the system is calculated using the following formula:

[0015] Wherein, Y(jω) n Let y(n) be the discrete form of the frequency response function Y(jω), where y(n) is the equally spaced discrete sample data of the acquired input signal, N is the total number of data samples, k and n are positive integers from 0 to N-1, and ω is the frequency response angular frequency parameter. n denoted as equally spaced discrete values ​​of the angular frequency parameter, where e is the natural constant and j is the imaginary unit.

[0016] As a further description of the above technical solution, the frequency response function Z(jω) of the frequency domain compensation n The algorithm formula is as follows:

[0017] Where N is the total number of data samples, and n is a positive integer from 0 to N-1.

[0018] As a further description of the above technical solution, the time-domain waveform function z(t) of the system n The algorithm formula is as follows:

[0019] Where z(t) n) represents the discrete data points of the time-domain waveform obtained after transformation, t n Represents a time series, ω n Represents a frequency sequence, Z(jω) n ) is the frequency response function for frequency domain compensation of the supplementary filter.

[0020] As a further description of the above technical solution, step S3 includes:

[0021] S31, regarding the pulse waveform function x(t) n ) and time-domain waveform function z(t n The first convolution process is performed to calculate the rising edge function y(t). n );

[0022] S32, take the rising edge function y(t) n The edge data points are converted into calibration data c. M ;

[0023] S33, the pulse waveform function x(n) and time-domain waveform function z(t) to be output. n A second convolution process is performed to calculate the final corrected data y(k).

[0024] As a further description of the above technical solution, the rising edge function y(t) n The algorithm formula is as follows:

[0025] Wherein, y(t) n ) for t n Waveform amplitude at time, pulse waveform function x(t) k In ) x(t1), x(t2), ..., x(t) m ) is 0, x(t) m+1 ), ...x(t) N-1 ) is 1, 0 <m<N-1。

[0026] As a further description of the above technical solution, the calibration data c M The algorithm formula is as follows: c1 = A1 c2 = A2 - A1 … c M =A M -A M-1

[0027] The edge data are A1, A2…A M There are M data points in total, and A1 = 0, ..., A k ≠0, ..., A M =1, A k This represents the magnitude of the k-th data point in the edge data.

[0028] As a further description of the above technical solution, the algorithm formula for the corrected data y(k) is as follows:

[0029] Where x(n) is the pulse waveform function to be output, n is the index of the waveform sequence, and c(kn) is the calibration data.

[0030] It also includes a quantum-controlled waveform overshoot correction system based on frequency response measurement, said correction system being used to perform the correction method as described in any one of the above technical solutions, comprising:

[0031] The output module generates a quantum-controlled waveform, loads an extremely narrow pulse, and then plays it out.

[0032] The measurement and display module acquires pulse waveforms and the pulse response of the measurement system, saves them as data points, and sends them to the calculation and processing module to display the pulse waveforms before and after correction.

[0033] The processing and calculation module calculates the system frequency response function and obtains corrected data through convolution processing of the frequency response function;

[0034] The transceiver control module receives the set waveform and sampling parameters and sends them to the generation output module. It also receives the correction data from the processing and calculation module and performs parallel calculations and waveform corrections in conjunction with the programmable gate array.

[0035] The above technical solution has the following advantages or beneficial effects:

[0036] This invention acquires the output waveform of a quantum control signal generation system and measures its impulse response. After data transformation, it obtains the frequency response function, the frequency response function with frequency domain compensation, and the time-domain waveform function. Then, it performs two convolution processes to convert edge data into calibration data. The calibration data is then used to correct the waveform. Parallel computation is performed in conjunction with a programmable gate array, resulting in a smoother corrected waveform. This avoids the problems of missed edges or over-correction that occur during manual calibration, effectively improving computational efficiency and calibration accuracy. Attached Figure Description

[0037] Figure 1 shows the calibration flowchart of the existing correction method;

[0038] Figure 2 is a second calibration flowchart of the existing correction method;

[0039] Figure 3 shows the pulse edge diagram after calibration using the existing correction method;

[0040] Figure 4 is a flowchart of the correction method proposed in this invention;

[0041] Figure 5 is a flowchart of the correction method proposed in this invention (II).

[0042] Figure 6 is a schematic diagram of parallel convolution calculation in a programmable gate array in the correction method proposed in this invention;

[0043] Figure 7 is a schematic diagram of the waveform amplitude in the correction method proposed in this invention;

[0044] Figure 8 is a schematic diagram of the correction system proposed in this invention;

[0045] Figure 9 is a comparative schematic diagram of the overshoot correction before and after the correction method proposed in this invention.

[0046] Legend: 1. Output module; 2. Measurement and display module; 3. Processing and calculation module; 4. Transmit and receive control module. Detailed Implementation

[0047] 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 embodiments of the present invention, and not all embodiments. 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.

[0048] The existing correction method suppresses and corrects overshoot by adjusting the amplitude of the pulse edge waveform at various points. The specific technical solution is shown in Figure 1: First, an uncorrected pulse signal is output through a waveform generator. At this time, the output pulse signal has a large overshoot. The specific value of the overshoot is then measured by a broadband oscilloscope. Subsequently, the calibration data is modified to make the pulse edge smoother, and then the calibration data is loaded into the waveform generator. According to the overshoot change trend, the amplitudes of multiple calibration data are traversed from large to small. After multiple corrections, the optimal calibration data can finally be obtained.

[0049] The specific calibration process is shown in Figure 2: Assume there are 6 adjustable points on the pulse waveform edge, namely [A0, A1, A2, A3, A4, A5]. Without correction, A0 equals 0, and A1, A2, A3, A4, and A5 equal 1. First, decrease A5. After observing the overshoot decrease on the oscilloscope, continue decreasing A5 until the overshoot increases. The value obtained at the lowest overshoot point is the undetermined value of A5. Decrease A4 until A4 < A5. During the adjustment process, observe the overshoot change displayed on the oscilloscope and record the value at the lowest overshoot point. Similarly, calibrate A3, A2, A1, and A0 respectively using the above method. However, adjusting each calibration point will affect the overall overshoot, and adjusting each point individually will also affect the data suppression effect of other points. Therefore, the above process needs to be repeated 3-5 times to adjust the overall correction effect of each point.

[0050] As shown in Figure 3, the final rising edge function is: y = 1 - e -bt ;

[0051] Where y is the waveform data amplitude, t is the time, and b is the rise edge control coefficient, used to control the rise edge speed. Since the frequency response of each waveform output device is different, the overshoot obtained by generating the pulse edge strictly according to the formula curve will also be different, and the rise edge is relatively slow. Therefore, when manually calibrating the edge according to the rise trend of the waveform, the calibration efficiency is low, the calibration accuracy is difficult to guarantee, and the waveform overshoot correction result obtained is not the best result.

[0052] The specific steps for applying calibration data to the pulse signal include: first, calculating the amplitude difference between each data point and its adjacent points; when the amplitude difference is greater than or equal to a certain threshold, it is considered that a pulse edge has appeared at that position; then, the amplitude at that edge position is replaced with the previously calibrated data. This process requires calculation point by point, which is inefficient and computationally intensive. If the threshold is set too high, it may skip edges with low pulse amplitude differences, thus missing the correction result. If the threshold is set too low, it may identify normal signals with low pulse amplitude differences as edges, resulting in noise overcorrection.

[0053] This application provides a quantum control waveform overshoot correction method and system based on frequency response measurement. It solves the problems that the calibration effect is affected by human factors and cannot guarantee the optimization of the calibration result. At the same time, the process of applying the calibration data to the pulse data for pre-distortion after calibration has low computational efficiency, large computational load, and some pulses are not identified or are identified as pulses when they are not pulses, resulting in overcorrection.

[0054] Referring to Figure 4, one embodiment of the present invention provides a quantum-controlled waveform overshoot correction method based on frequency response measurement, comprising:

[0055] S1, set the pulse parameters and output mode of the quantum control signal generation system, load the pulse signal and output it for playback, and collect the output waveform to obtain the pulse response data of the quantum control signal generation system;

[0056] S2, perform multi-step data transformation processing on the data points of the system's impulse response to obtain the system's time-domain waveform function;

[0057] S3, set the pulse waveform function, perform the first convolution process on the time-domain waveform function and the pulse waveform function to convert it into calibration data, and perform the second convolution process on the pulse waveform function and the calibration data to obtain a corrected data;

[0058] S4: Load correction data to correct the waveform and perform parallel computation in conjunction with a programmable gate array.

[0059] Referring to Figure 5, step S1 includes:

[0060] S11 uses a quantum-controlled waveform generation device (such as a waveform generator) to load an extremely narrow pulse with only one sampling point being 1 and the other points being 0. After passing through channel conditioning, filtering, amplification and other circuits, it is played out and output in an equally spaced trigger playback mode.

[0061] S12, use a measuring device (such as a broadband oscilloscope) to measure the output waveform, adjust the trigger and sampling rate of the oscilloscope so that it can completely acquire a pulse waveform, which is the pulse response of the quantum control signal generation system, and save the waveform as data point y(n);

[0062] Referring to Figure 5, step S2 includes:

[0063] S21, Perform Fourier transform on the impulse response data of the quantum control signal generation system to calculate the system's frequency response function Y(jω);

[0064] S22, taking the reciprocal of the system's frequency response function Y(jω), we calculate the frequency response function Z(jω) with frequency domain compensation. n );

[0065] S23, Frequency response function Z(jω) for frequency domain compensation n Perform an inverse Fourier transform to calculate the system's time-domain waveform function z(t). n ).

[0066] In this embodiment, the discrete form of the system's frequency response function Y(jω) is expressed as Y(jω) n The algorithm formula is as follows:

[0067] Where y(n) represents the equally spaced discrete sample data of the acquired input signal, N is the total number of data samples, k and n are positive integers from 0 to N-1, and ω is the frequency response angular frequency parameter. n denoted as equally spaced discrete values ​​of the angular frequency parameter, where e is the natural constant and j is the imaginary unit.

[0068] In this embodiment, the frequency response function Z(jω) of frequency domain compensation n The algorithm formula is as follows:

[0069] Where N is the total number of data samples, and n is a positive integer from 0 to N-1;

[0070] It is important to explain in detail that, in order to ensure that the system can effectively compensate and process the frequency response, the frequency range of the input signal must not exceed the system's maximum bandwidth. This is to avoid adverse effects on the frequency response compensation effect caused by exceeding the frequency range. Specifically, 0 ≤ n ≤ 0.4*N, 0.6*N ≤ n ≤ N. This range corresponds to the maximum bandwidth of the frequency response of the control signal generation system. Exceeding this bandwidth results in a rapid decrease in the frequency response, leading to an excessively large reciprocal value, which in turn affects the compensation effect.

[0071] In this embodiment, the system's time-domain waveform function z(t) n The algorithm formula is as follows:

[0072] Where z(t) n ) represents the discrete data points of the time-domain waveform obtained after transformation, t n Represents a time series, ω n Represents a frequency sequence, Z(jω) n ) is the frequency response function for frequency domain compensation of the supplementary filter.

[0073] Referring to Figure 5, step S3 includes:

[0074] S31, regarding the pulse waveform function x(t) n ) and time-domain waveform function z(t n The first convolution process is performed to calculate the rising edge function y(t). n );

[0075] S32, take the rising edge function y(t) n The edge data points are converted into calibration data c. M ;

[0076] S33, the pulse waveform function x(n) and time-domain waveform function z(t) to be output. n A second convolution process is performed to calculate the final corrected data y(k).

[0077] In this embodiment, the rising edge function y(t) n The algorithm formula is as follows:

[0078] Wherein, y(t) n ) for t n Waveform amplitude at time, pulse waveform function x(t) k In ) x(t1), x(t2), ..., x(t) m ) is 0, x(t) m+1 ), ...x(t) N-1 ) is 1, 0 <m<N-1。

[0079] In this embodiment, y(t) is taken. n The edge data points in the data are converted into convolutional data using the following formula, which is the final calibration data c. M The algorithm formula is as follows: c1 = A1 c2 = A2 - A1 … c M =A M -A M-1

[0080] Referring to Figure 7, where the edge data are A1, A2…A M There are M data points in total, and A1 = 0, ..., A k ≠0, ..., A M =1, A k This represents the amplitude of the k-th data point in the edge data. As shown in Figure 7, let the data y(t) n If the edge of the edge starts from y(t3), then there are a total of 5 data points (M=5), where A1=y(t3)=0, A2=y(t4), A3=y(t5), A4=y(t6), and A5=y(t7)=1.

[0081] In this embodiment, when correcting the pulse waveform using calibration data, the pulse waveform and calibration data can be directly convolved to calculate the corrected data y(k). The algorithm formula is as follows:

[0082] Where x(n) is the pulse waveform function to be output, n is the index of the waveform sequence, that is, x(n) is the amplitude of the nth point of the waveform, and c(kn) is the calibration data.

[0083] In this embodiment, through the above convolution calculation, there is no need to determine the position of the rising or falling edge during calibration correction. Therefore, there is no problem of missing edges or over-correction. Furthermore, the convolution method can make good use of the parallel computing advantage of the programmable gate array to achieve the purpose of real-time processing, making the corrected waveform smoother and effectively avoiding the problems of missing edges or over-correction caused by manual calibration.

[0084] Referring to Figure 6, driven by the logic clock, the pulse waveform data points move continuously and are multiplied by the convolution kernel, which can quickly output the convolution result without adding extra processing time due to data calculation. This improves computational efficiency and enables real-time processing, effectively enhancing computational efficiency and calibration accuracy.

[0085] Referring to Figure 8, the present invention also includes a quantum-controlled waveform overshoot correction system based on frequency response measurement. The correction system is used to perform the correction method as described in any of the above technical solutions, including:

[0086] Output module 1 is used to generate quantum control waveforms, which are then played back and output after being loaded with extremely narrow pulses.

[0087] Measurement and display module 2 collects pulse waveforms and measurement system pulse responses, saves them as data points, and sends them to the calculation and processing module to display the pulse waveforms before and after correction.

[0088] The processing and calculation module 3 calculates the system frequency response function and obtains the corrected data through convolution processing of the frequency response function;

[0089] The transceiver control module 4 receives the set waveform and sampling parameters, sends them to the generation output module, receives the correction data from the processing and calculation module, and performs parallel calculation and waveform correction in conjunction with the programmable gate array.

[0090] In this embodiment, the correction system executes the above-mentioned correction method: the output module 1 generates quantum-controlled waveforms, which are then played back and output after being loaded with extremely narrow pulses. These waveforms have specific spectral characteristics and modulation methods. The measurement and display module 2 acquires the pulse waveforms and measures the system's pulse response, saves them as data points, and sends them to the calculation and processing module. After receiving the measurement data, the processing and calculation module 3 performs data processing and calculation to obtain the system's frequency response function. The frequency response function is then subjected to two convolution processes to obtain the correction data. The transceiver control module 4 receives the set waveforms and sampling parameters, sends them to the output module 1 to generate waveforms, and after receiving the correction data, performs parallel calculations with a programmable gate array to correct the waveforms. The waveforms before and after correction are displayed in real-time by the measurement and display module 2 to ensure that the correction effect meets expectations. As shown in Figure 9, the final corrected waveform is relatively smooth, without overshoot, and the correction effect is good. Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for quantum control waveform overshoot correction based on frequency response measurement, characterized in that, include: S1, set the pulse parameters and output mode of the quantum control signal generation system, load the pulse signal and output it for playback, and collect the output waveform to obtain the pulse response data of the quantum control signal generation system; S2, perform data transformation processing on the data points of the system's impulse response to obtain the system's time-domain waveform function; S3, set the pulse waveform function, perform the first convolution process on the time-domain waveform function and the pulse waveform function to convert it into calibration data, and perform the second convolution process on the pulse waveform function and the calibration data to obtain a corrected data; S4 loads correction data to correct the waveform and performs parallel computation in conjunction with a programmable gate array.

2. The revision method of claim 1, wherein, In said step S2, comprising ： S21, Perform Fourier transform on the impulse response data of the quantum control signal generation system to calculate the system's frequency response function Y(jω); S22, the frequency response function Y(jω) of the system is inverted, and the frequency domain compensation frequency response function Z(jω) is calculated n ); S23, the Fourier inverse transform is performed on the frequency response function Z(jω n ) compensated in the frequency domain, and the time-domain waveform function z(t n ) of the system is calculated.

3. The revision method of claim 2, wherein, The frequency response function Y(jω) of the system is given by the algorithm formula as follows: Wherein, Y(jω) n Let y(n) be the discrete form of the frequency response function Y(jω), where y(n) is the equally spaced discrete sample data of the acquired input signal, N is the total number of data samples, k and n are positive integers from 0 to N-1, and ω is the frequency response angular frequency parameter. n denoted as equally spaced discrete values ​​of the angular frequency parameter, where e is the natural constant and j is the imaginary unit.

4. The revision method of claim 2, wherein, The frequency response function Z(jω) of the frequency domain compensation n The algorithm formula is as follows: Where N is the total number of data samples, and n is a positive integer from 0 to N-1.

5. The revision method of claim 2, wherein, The time-domain waveform function z(t) of the system n The algorithm formula is as follows: where z(t n ) is the discrete time-domain waveform data point after conversion, t n represents the time sequence, ω n represents the frequency sequence, and Z(jω n ) is the frequency response function of the compensation of the supplementary filter in the frequency domain.

6. The revision method of claim 1 wherein, In said step S3, comprising ： S31, the first convolution processing is performed on the pulse waveform function x(t n ) and the time-domain waveform function z(t n ), and an up edge function y(t n ) is calculated. S32, take the edge data point of the rising edge function y(t n ) and convert it into calibration data c M ; S33, the pulse waveform function x(n) and the time domain waveform function z(t n ) to be output are subjected to a second convolution process, and final correction data y(k) is calculated.

7. The revision method of claim 6, wherein, The rising edge function y(t n The algorithm formula is as follows: wherein y(t n ) is the waveform amplitude at time t n , x(t k ) is the pulse waveform function, x(t1), x(t2, … x(t m ) are 0 in the pulse waveform function x(t m+1 ), … x(t N-1 ) are 1, and 0 < m < N-1.

8. The revision method of claim 6, wherein, The calibration data c M , algorithm formula as follows: c1=A1 c2=A2-A1 … c M =A M -A M-1 Wherein, the edge data is A1, A2...A M M data, and A1=0,...,A k ≠0,...,A M =1, A k indicates the amplitude of the kth data in the edge data.

9. The revision method of claim 6, wherein, The modified data y(k) is calculated according to the following algorithm: Where x(n) is the pulse waveform function to be output, n is the index of the waveform sequence, and c(kn) is the calibration data.

10. A quantum-controlled waveform overshoot correction system based on frequency response measurement, characterized in that, The correction system is used to perform the correction method as described in any one of claims 1-9, including: Output module (1) generates a quantum control waveform, loads a pulse, and then plays it out. The measurement display module (2) collects pulse waveforms and measurement system pulse responses, saves them as data points, sends them to the calculation and processing module, and displays the pulse waveforms before and after correction. The processing and calculation module (3) calculates the system frequency response function and obtains the corrected data by convolution processing the frequency response function; The transceiver control module (4) receives the set waveform and sampling parameters and sends them to the generation output module. It also receives the correction data from the processing calculation module and performs parallel calculation and waveform correction in conjunction with the programmable gate array.

Images