A waveform optimization method and device based on channel crosstalk, equipment and medium

Abstract

The application provides a waveform optimization method and device based on channel crosstalk, equipment and medium, used in the technical field of quantum waveform optimization, including: respectively applying test pulses to each control channel, and determining the test signal amplitude output by each quantum bit port according to the channel mapping relationship; determining the local crosstalk matrix corresponding to each preset frequency according to the plurality of test signal amplitudes corresponding to each control channel, and applying a preset waveform to the target quantum bit port to obtain a corresponding target waveform; based on the local crosstalk matrix, compensating the target waveform to obtain a target compensation waveform, and after outputting the target compensation waveform, determining a target residual signal based on the corresponding response waveform and the target waveform; based on the target residual signal, determining a plurality of signal characteristic values, and adjusting the local crosstalk matrix according to the signal characteristic values and the respective corresponding preset threshold values; in this way, by compensating the quantum channel crosstalk, the accuracy of the output waveform is improved.

Description

Technical Field

[0001] This application relates to the field of quantum waveform optimization technology, and in particular to a waveform optimization method, apparatus, device and medium based on channel crosstalk. Background Technology

[0002] As the number of qubits in quantum computing systems continues to increase, control systems typically need to simultaneously output multiple microwave or radio frequency pulses to achieve single-qubit gate, double-qubit gate, and parallel gate operations. In actual measurement and control links, due to factors such as transmission line coupling, packaging parasitic parameters, device bandwidth limitations, analog front-end non-ideals, and multi-channel synchronous output errors, crosstalk of varying degrees often exists between different qubit control channels.

[0003] Existing waveform optimization methods include improving hardware routing, shielding structures, or packaging designs to reduce physical coupling; performing amplitude and phase calibration or pulse shaping on a single control channel; or using empirical compensation methods to correct specific crosstalk in certain scenarios.

[0004] However, existing waveform optimization methods suffer from poor compensation for channel crosstalk. Summary of the Invention

[0005] This application provides a waveform optimization method, apparatus, device, and medium based on channel crosstalk, to solve the problem that existing waveform optimization methods have poor compensation effect on channel crosstalk.

[0006] Firstly, this application provides a waveform optimization method based on channel crosstalk, the method comprising: Based on multiple preset frequencies, test pulses are applied to each control channel respectively, and the amplitude of the test signal output by each quantum bit port is determined according to the channel mapping relationship. Based on the amplitude values ​​of multiple test signals corresponding to each control channel, the local crosstalk matrix corresponding to each preset frequency is determined, and a preset waveform is applied to the target quantum bit port to obtain the corresponding target waveform; Based on the local crosstalk matrix, the target waveform is compensated to obtain the target compensated waveform. After outputting the target compensated waveform, the target residual signal is determined based on the corresponding response waveform and the target waveform. Based on the target residual signal, multiple signal feature values ​​are determined, and the local crosstalk matrix is ​​adjusted according to the signal feature values ​​and their corresponding preset thresholds.

[0007] In some embodiments of this application, the local crosstalk matrix corresponding to each preset frequency is determined based on the amplitude values ​​of multiple test signals corresponding to each control channel, including: For each preset frequency, the corresponding matrix rows are determined based on the port sequence corresponding to the quantum bit port, and the corresponding matrix columns are determined based on the channel sequence corresponding to the control channel. Based on the quantum bit ports and control channels corresponding to the amplitudes of each test signal, the corresponding rows and columns of the target matrix are determined, the corresponding elements of the target matrix are obtained, and the element values ​​corresponding to the elements of the target matrix are determined based on the amplitudes of the test signals, thus obtaining the local crosstalk matrix corresponding to the preset frequency.

[0008] In some embodiments of this application, the compensation includes full-frequency domain compensation and frequency domain segmented compensation. Based on the local crosstalk matrix, the target waveform is compensated to obtain the target compensated waveform, including: Determine the corresponding compensation method; When the compensation method is full-frequency domain compensation, the average value is calculated based on all local crosstalk matrices to obtain the average crosstalk matrix, and the average crosstalk matrix is ​​determined as the target crosstalk matrix corresponding to the target waveform. When the compensation method is frequency domain segmented compensation, Fourier decomposition is performed on the target waveform to obtain the waveform components corresponding to each frequency band, and the target crosstalk matrix corresponding to each waveform component is determined based on the local crosstalk matrix. The target compensation waveform is determined based on the target crosstalk matrix, its corresponding transpose matrix, waveform, and regularization parameters.

[0009] In some embodiments of this application, the target compensation waveform is determined based on the target crosstalk matrix and its corresponding transpose matrix, waveform, and regularization parameters, including: Based on the target crosstalk matrix and its corresponding transpose matrix, waveform, and regularization parameters, the corresponding initial compensation waveform is calculated, and the initial compensation waveform is subjected to amplitude limiting processing based on a preset amplitude limiting threshold. Gaussian smoothing is applied to the initial compensation waveform after amplitude limiting to obtain the target compensation waveform.

[0010] In some embodiments of this application, multiple signal feature values ​​are determined based on the target residual signal, including: The target residual signal is sampled to obtain the residual value corresponding to each sampling point, and the corresponding root mean square error value is calculated based on the residual value. Based on the target waveform and the target compensated waveform, the uncompensated signal energy value and the compensated signal energy value of the output of the non-target quantum bit port are determined respectively, and the corresponding energy suppression ratio is calculated based on the ratio between the uncompensated signal energy value and the compensated signal energy value. Based on the root mean square error, energy suppression ratio, and target amplitude corresponding to the target compensation waveform, determine the corresponding signal characteristic values.

[0011] In some embodiments of this application, the local crosstalk matrix is ​​adjusted according to the signal feature values ​​and their corresponding preset thresholds, including: Determine whether each signal feature value is less than its corresponding preset threshold; If at least one signal feature value is not less than the corresponding preset threshold, then the signal feature value not less than the corresponding preset threshold is determined as the target feature value, and the local crosstalk matrix is ​​adjusted based on the target feature value; If all values ​​are less than 1, then no adjustment is made to the local crosstalk matrix.

[0012] In some embodiments of this application, the local crosstalk matrix is ​​adjusted based on the target eigenvalues, including: Determine the target feature values; If the target eigenvalue is the root mean square error or the target magnitude, then the regularization parameter corresponding to the local crosstalk matrix is ​​adjusted. If the target characteristic value is the energy suppression ratio, then based on the updated test pulse, the updated local crosstalk matrix corresponding to each preset frequency is determined.

[0013] Secondly, this application provides a waveform optimization device based on channel crosstalk, the device comprising: The test module is used to apply test pulses to each control channel based on multiple preset frequencies, and determine the amplitude of the test signal output by each quantum bit port according to the channel mapping relationship. The determination module is used to determine the local crosstalk matrix corresponding to each preset frequency based on the amplitude of multiple test signals corresponding to each control channel, and apply a preset waveform to the target quantum bit port to obtain the corresponding target waveform; The compensation module is used to compensate the target waveform based on the local crosstalk matrix to obtain the target compensated waveform, and after outputting the target compensated waveform, to determine the target residual signal based on the corresponding response waveform and the target waveform. The adjustment module is used to determine multiple signal feature values ​​based on the target residual signal, and adjust the local crosstalk matrix according to the signal feature values ​​and their corresponding preset thresholds.

[0014] Thirdly, this application provides a computer device, including: a processor, and a memory communicatively connected to the processor; The memory stores instructions that the computer executes; The processor executes computer execution instructions stored in memory to implement the method of this application.

[0015] Fourthly, this application provides a computer-readable storage medium storing program code, which, when executed by a processor, is used to implement the method of this application.

[0016] Compared with existing technologies, the method of this application applies test pulses to each control channel based on multiple preset frequencies, thereby accurately acquiring the test signal amplitude of each qubit port according to the mapping relationship between the channel and the qubit port, and realizing the accurate characterization of crosstalk coupling characteristics between channels at different frequencies. It uses a constructed local crosstalk matrix to compensate for crosstalk in the target waveform and generates a target compensated waveform. After output, the target residual signal is obtained by comparing the actual response waveform with the ideal target waveform. This achieves the elimination of frequency-varying crosstalk interference between multiple qubit control channels based on the target compensated waveform, restoring the true response deviation of the qubit. Multiple signal feature values ​​are extracted from the target residual signal, and the feature values ​​are compared with corresponding preset thresholds to adaptively adjust the local crosstalk. The crosstalk matrix is ​​optimized to correct crosstalk model mismatch caused by environmental and device drift. It can introduce MIMO signal processing methods from the communication field into quantum measurement and control waveform generation, which helps to handle multi-channel crosstalk problems with a unified model. It can also achieve multi-channel joint pre-compensation through crosstalk matrix modeling, which can handle complex coupling scenarios more effectively than single-channel independent correction. It can generate anti-phase cancellation waveforms at the digital baseband layer, which can reduce the dependence on additional hardware modifications and facilitate deployment to existing measurement and control platforms. It can simultaneously take into account the target waveform accuracy, non-target crosstalk suppression and hardware output constraints, improving the engineering feasibility of the solution. It is suitable for multi-qubit parallel control scenarios and can improve gate fidelity, control stability and system scalability to a certain extent. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] Figure 1 A flowchart illustrating a waveform optimization method based on channel crosstalk provided in this application embodiment; Figure 2 A flowchart illustrating another waveform optimization method based on channel crosstalk provided in this application embodiment; Figure 3 A coupling diagram illustrating a waveform optimization method based on channel crosstalk provided in an embodiment of this application; Figure 4 A schematic diagram of a waveform optimization device based on channel crosstalk provided in an embodiment of this application; Figure 5 This is a structural block diagram of an apparatus for performing a waveform optimization method based on channel crosstalk according to an embodiment of this application. Detailed Implementation

[0019] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0020] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0021] In existing technologies, the following approaches are typically used to reduce the impact of crosstalk: reducing physical coupling by improving hardware wiring, shielding structures, or packaging designs; performing amplitude and phase calibration or pulse shaping on a single control channel; and using empirical compensation methods to correct specific crosstalk in some scenarios.

[0022] In practical applications, although the above-mentioned solutions can reduce crosstalk to some extent, they still have obvious shortcomings. Hardware isolation methods are highly dependent on the system structure, have high modification costs, and are difficult to maintain consistent performance as the scale of qubits increases. Single-channel compensation methods usually assume that the influence of other channels is weak, making it difficult to handle the coupling problem when multiple channels are excited simultaneously. Empirical corrections lack a unified mathematical model, making it difficult to adapt to different operating points, different combinations of parallel gates, and different quantum chips. At the same time, existing solutions often treat waveform generation, crosstalk compensation, and control optimization separately, resulting in unstable compensation results and limiting the improvement of gate fidelity and system scalability.

[0023] Figure 1 This is a flowchart illustrating a waveform optimization method based on channel crosstalk, provided as an embodiment of this application. Figure 1 As shown, this waveform optimization method based on channel crosstalk may include the following steps: S110. Based on multiple preset frequencies, test pulses are applied to each control channel respectively, and the amplitude of the test signal output by each quantum bit port is determined according to the channel mapping relationship.

[0024] The preset frequency refers to a pre-determined frequency range that covers the actual working bandwidth of the multi-qubit control channel. In practical applications, the frequency domain can be divided into four frequency bands with boundaries of [0, 120) MHz, [120, 260) MHz, [260, 500) MHz, and [500, 1000] MHz. Since crosstalk between quantum measurement and control channels is frequency-dependent, the channel coupling strength, amplitude, and phase attenuation characteristics vary greatly at different frequencies. By dividing the preset frequency, the entire working bandwidth can be split into multiple independent frequency bands for individual calibration.

[0025] The channel mapping relationship is a pre-established fixed binding relationship between the control output channel and the quantum bit control port, which clearly defines which quantum bit port each physical control output channel drives, thereby establishing the identity correspondence between the physical channel and the quantum bit, and clarifying which control channel's excitation will affect which quantum bit port, thus avoiding confusion and mismatch between channels and bit ports.

[0026] The test signal amplitude is the amplitude of the response signal induced and output by each quantum bit port after a test pulse is applied to a designated control channel at a preset frequency.

[0027] Based on this, by applying test pulses to each control channel within different test frequency ranges, the amplitude of the corresponding quantum bit response terminal of each control channel is collected according to the channel mapping relationship.

[0028] S120. Based on the amplitude values ​​of multiple test signals corresponding to each control channel, determine the local crosstalk matrix corresponding to each preset frequency, and apply a preset waveform to the target quantum bit port to obtain the corresponding target waveform.

[0029] The local crosstalk matrix is ​​a multi-input multi-output coupling coefficient matrix constructed according to the channel mapping relationship within a single preset frequency, based on the excitation of each control channel and the measured test signal amplitude of each qubit port. The matrix dimension is consistent with the number of control channels and the number of qubit ports. Each element of the matrix represents the signal coupling gain and crosstalk strength of a certain control channel to a certain qubit port at the current frequency. Thus, by constructing the matrix, the crosstalk characteristics between channels within the corresponding frequency band can be quantified so as to compensate for crosstalk in the future.

[0030] The target qubit port refers to the control input port corresponding to the qubit designated to perform a quantum gate operation within the current control cycle. In other words, it is one or more qubit control ports that need to be precisely applied with control waveforms to achieve quantum control in a quantum manipulation task. The remaining qubit ports that do not require this manipulation and only need to suppress crosstalk interference are non-target qubit ports, which are selected from all qubit ports as the target ports for this manipulation.

[0031] The preset waveform refers to the pre-designed ideal baseband pulse waveform, including Gaussian waveforms, DRAG waveforms, and other special waveforms that meet the requirements of quantum bit energy level transitions and gate operations. The preset waveform is then applied to the target quantum bit port to obtain the expected target waveform.

[0032] The target waveform is a multi-channel dimension ideal desired waveform vector formed by loading the corresponding preset waveform onto the selected target qubit port and setting the non-target qubit ports to zero, based on the preset waveform; it represents the standard response waveform that the desired qubit port can receive without crosstalk and without error.

[0033] Based on this, in practical applications, three main pulse center positions can be set within one control cycle, located near the 380th, 980th, and 1420th sampling points, respectively. The number of sampling points is set to 2048, and the sampling rate is set to 2.0 GHz. Ideal baseband control waveforms are generated for the target qubit ports Q1 and Q3, where port Q1 uses a Gaussian envelope waveform, and port Q3 uses a DRAG-type waveform with derivative correction terms; the corresponding pulse amplitudes are configured in proportions of 1.00, 0.72, and 0.88. The desired response of non-target qubit ports is set to zero, thus obtaining the target waveform.

[0034] S130. Based on the local crosstalk matrix, the target waveform is compensated to obtain the target compensated waveform. After outputting the target compensated waveform, the target residual signal is determined based on the corresponding response waveform and the target waveform.

[0035] The target compensation waveform refers to the actual driving waveform generated and output by the control channel after pre-distorting and pre-cancelling the ideal target waveform according to the local crosstalk matrix. This realizes the coupling relationship modeled by the crosstalk matrix, performs reverse pre-distortion on the target waveform in advance, and cancels the inherent crosstalk coupling after the waveform is transmitted through the actual channel, so that the actual received waveform of the quantum bit is close to the ideal target waveform.

[0036] The response waveform refers to the time-domain induced waveform actually acquired at each quantum bit port after the target compensation waveform is output by the control channel, through the physical transmission link and channel crosstalk coupling, thus reflecting the real waveform transmission result.

[0037] The target residual signal refers to the error timing signal obtained by performing point-by-point difference calculation between the response waveform and the target waveform under the same clock and the same sampling point, thereby characterizing the instantaneous deviation of the actual received waveform from the ideal expected waveform.

[0038] Based on this, the target waveform is compensated by a local crosstalk matrix to obtain the target compensated waveform. After outputting the target compensated waveform, the target residual signal between the response waveform and the target waveform is calculated based on the response waveforms actually acquired at each quantum bit port.

[0039] S140. Based on the target residual signal, determine multiple signal feature values, and adjust the local crosstalk matrix according to the signal feature values ​​and their corresponding preset thresholds.

[0040] Among them, the signal eigenvalues ​​are quantitative indicators extracted from the residual signal and used to evaluate the quality of the compensation effect. Each eigenvalue characterizes the performance of the control system from different dimensions. In practical applications, the signal eigenvalues ​​may include the root mean square error (RMSE) of the target channel, the crosstalk suppression ratio relative to the uncompensated baseline, and the peak value of the output waveform after compensation, thereby quantifying the time-domain residual, crosstalk magnitude, and amplitude exceeding limits.

[0041] The preset threshold is a pre-determined critical value used to judge whether the compensation effect corresponding to each signal feature value meets the requirements. This allows for defining a qualified range for each performance indicator. If the feature value is within the threshold range, the system compensation effect is qualified; if it exceeds the threshold, the performance is deemed substandard and the waveform needs to be corrected.

[0042] Based on this, in practical applications, the actual quantum control channel, microwave link, and device characteristics are not permanently fixed. Various factors such as changes in ambient temperature, device temperature drift, and aging can cause the originally calibrated local crosstalk matrix to no longer match the actual crosstalk characteristics. Therefore, based on the target residual signal, multiple signal feature values ​​can be determined, and based on the signal feature values ​​and their corresponding preset thresholds, it can be determined whether the current crosstalk matrix is ​​no longer accurate, thereby adjusting the local crosstalk matrix.

[0043] Based on the feasible implementation of S120 described above, this application further provides a method for determining the local crosstalk matrix corresponding to each preset frequency based on the amplitude values ​​of multiple test signals corresponding to each control channel, including: For each preset frequency, the corresponding matrix rows are determined based on the port sequence corresponding to the quantum bit port, and the corresponding matrix columns are determined based on the channel sequence corresponding to the control channel. Based on the quantum bit ports and control channels corresponding to the amplitudes of each test signal, the corresponding rows and columns of the target matrix are determined, the corresponding elements of the target matrix are obtained, and the element values ​​corresponding to the elements of the target matrix are determined based on the amplitudes of the test signals, thus obtaining the local crosstalk matrix corresponding to the preset frequency.

[0044] Among them, the port sequence is an ordered numbering sequence formed by arranging all the control ports of the qubits in a fixed order; for example, if the qubit ports are Q1, Q2, Q3, and Q4 in sequence, they form the port sequence [Q1, Q2, Q3, Q4] in a fixed order, and the matrix rows are determined according to the corresponding sequence.

[0045] A channel sequence is an ordered numbered sequence formed by arranging all control output channels in a fixed order; for example, if the control channels are CH1, CH2, CH3, and CH4 in sequence, the fixed order forms the channel sequence [CH1, CH2, CH3, CH4], thus determining the corresponding matrix column.

[0046] The target matrix row is the uniquely determined matrix row position in the local crosstalk matrix for a specific qubit port based on its sorting number in the port sequence, while the target matrix column is the uniquely determined matrix column position in the local crosstalk matrix for a specific control channel based on its sorting number in the channel sequence. For example, port Q1 corresponds to the first row of the matrix, port Q2 corresponds to the second row of the matrix, channel CH1 corresponds to the first column of the matrix, and channel CH2 corresponds to the second column of the matrix.

[0047] The target matrix element is a unique grid position element determined in the local crosstalk matrix after cross-locating the target matrix rows (i.e., qubit ports) and target matrix columns (i.e., control channels).

[0048] Based on this, by using the quantum bit port and control channel corresponding to the test signal amplitude, the corresponding target matrix rows and columns are determined respectively, and based on the specific value of the test signal amplitude, the corresponding matrix element values ​​are determined to obtain each target matrix element, thereby constructing the crosstalk matrix.

[0049] Based on the feasible implementation of S130 described above, this application further provides compensation including full-frequency domain compensation and frequency domain segmented compensation. Based on the local crosstalk matrix, the target waveform is compensated to obtain the target compensated waveform, including: Determine the corresponding compensation method; When the compensation method is full-frequency domain compensation, the average value is calculated based on all local crosstalk matrices to obtain the average crosstalk matrix, and the average crosstalk matrix is ​​determined as the target crosstalk matrix corresponding to the target waveform. When the compensation method is frequency domain segmented compensation, Fourier decomposition is performed on the target waveform to obtain the waveform components corresponding to each frequency band, and the target crosstalk matrix corresponding to each waveform component is determined based on the local crosstalk matrix. The target compensation waveform is determined based on the target crosstalk matrix, its corresponding transpose matrix, waveform, and regularization parameters.

[0050] Among them, the full-frequency domain compensation method refers to using a globally constant crosstalk matrix, that is, the average value of all local crosstalk matrices, to characterize the channel coupling characteristics in the entire frequency band, and calculate the compensation waveform based on this. Therefore, it is only necessary to average all local crosstalk matrices to obtain a single matrix, without the need to perform Fourier decomposition and segment-by-segment solution of the waveform, which is suitable for low-latency and fast control scenarios.

[0051] Frequency domain segmented compensation refers to dividing the entire frequency band into multiple sub-bands, each using an independent local crosstalk matrix. Compensation components are calculated separately for each sub-band, and then the components are spliced ​​together to reconstruct a time-domain compensated waveform. In quantum multi-channel MIMO systems, the self-coupling and cross-channel crosstalk intensity vary significantly at different frequencies. By using independent matrix modeling for each frequency band, the true crosstalk characteristics of each frequency point can be accurately matched, solving the problem of insufficient accuracy of single average matrix modeling.

[0052] The regularization parameter is a regularization coefficient introduced when solving the compensation waveform, used to balance the trade-off between compensation accuracy and output energy / numerical stability.

[0053] Based on this, in practical applications, Let I be the regularization parameter and I be the identity matrix. Then, when the compensation method is full-frequency domain compensation: ; To compensate for the target waveform, The average crosstalk matrix, This is the transpose of the average crosstalk matrix. The target waveform; When the compensation method is frequency domain segmented compensation: ; For the i-th target compensation waveform component, Let i be the local crosstalk matrix. Let be the transpose of the i-th local crosstalk matrix. Let i be the i-th target waveform component. Based on multiple target compensation waveform components, after frequency domain splicing, the waveform is reconstructed into a time domain compensation waveform through inverse transformation.

[0054] Based on the feasible implementation of S130 described above, this application further provides a method for determining the target compensation waveform based on the target crosstalk matrix and its corresponding transpose matrix, waveform, and regularization parameters, including: Based on the target crosstalk matrix and its corresponding transpose matrix, waveform, and regularization parameters, the corresponding initial compensation waveform is calculated, and the initial compensation waveform is subjected to amplitude limiting processing based on a preset amplitude limiting threshold. Gaussian smoothing is applied to the initial compensation waveform after amplitude limiting to obtain the target compensation waveform.

[0055] Among them, the preset amplitude limiting threshold refers to the maximum output voltage, dynamic range and device tolerance of the control channel hardware drive circuit that is predetermined in advance. It is used to limit the maximum allowable amplitude of the instantaneous sampling point of the compensation waveform. It is the critical standard for distinguishing the normal amplitude of the waveform from the over-limit amplitude. It includes positive amplitude threshold and negative amplitude threshold, which constitute the amplitude constraint boundary of the waveform output.

[0056] Gaussian smoothing refers to a waveform shaping method that uses a Gaussian function as a smoothing convolution kernel to perform time-domain sliding convolution filtering on the initial compensated waveform after amplitude limiting.

[0057] Based on this, after the initial compensation waveform is solved by matrix, it is easy to have instantaneous amplitude overshoot and spike change, which exceeds the linear output range of the power amplifier. By forcibly clamping the over-limit point through the preset amplitude limiting threshold, the waveform signal exceeding the preset amplitude limiting threshold is subjected to amplitude limiting processing to prevent hardware saturation clipping and distortion damage to the device.

[0058] Based on the feasible implementation of S140 described above, this application further provides a method for determining multiple signal characteristic values ​​based on the target residual signal, including: The target residual signal is sampled to obtain the residual value corresponding to each sampling point, and the corresponding root mean square error value is calculated based on the residual value. Based on the target waveform and the target compensated waveform, the uncompensated signal energy value and the compensated signal energy value of the output of the non-target quantum bit port are determined respectively, and the corresponding energy suppression ratio is calculated based on the ratio between the uncompensated signal energy value and the compensated signal energy value. Based on the root mean square error, energy suppression ratio, and target amplitude corresponding to the target compensation waveform, determine the corresponding signal characteristic values.

[0059] Among them, the uncompensated signal energy value is the total signal energy of the response waveform acquired at the non-target qubit port during the entire sampling time when the target waveform is sent directly as the output waveform without any crosstalk compensation; thus, it characterizes the original interference energy of the channel crosstalk coupled to the non-target qubit port without any compensation measures.

[0060] The compensation signal energy value is the total energy of the actual response waveform collected at the same non-target qubit port and within the same sampling time range when the target compensation waveform is output after crosstalk compensation is applied; it represents the residual crosstalk energy remaining at the non-target port after pre-compensation cancellation.

[0061] Based on this, in practical applications, the actual response can be compared with the expected response. The average root mean square error of the target channel, the crosstalk suppression ratio relative to the uncompensated baseline, and the peak value of the compensated output are used as evaluation indicators to obtain multiple signal characteristic values. Among them, the crosstalk suppression ratio is the amount of decrease in the residual energy of the non-target channel relative to the uncompensated state.

[0062] Based on the feasible implementation of S140 described above, this application further provides an adjustment of the local crosstalk matrix according to the signal feature values ​​and their corresponding preset thresholds, including: Determine whether each signal feature value is less than its corresponding preset threshold; If at least one signal feature value is not less than the corresponding preset threshold, then the signal feature value not less than the corresponding preset threshold is determined as the target feature value, and the local crosstalk matrix is ​​adjusted based on the target feature value; If all values ​​are less than 1, then no adjustment is made to the local crosstalk matrix.

[0063] Based on this, by comparing the calculated signal feature values ​​one by one with their respective preset thresholds, it is determined that the current crosstalk matrix is ​​not abnormal when all signal feature values ​​are less than the corresponding preset thresholds; however, if at least one signal feature value is not less than the corresponding preset threshold, it is determined that the current crosstalk matrix is ​​abnormal and needs to be adjusted.

[0064] Based on the feasible implementation of S140 described above, this application further provides an adjustment of the local crosstalk matrix based on the target eigenvalue, including: Determine the target feature values; If the target eigenvalue is the root mean square error or the target magnitude, then the regularization parameter corresponding to the local crosstalk matrix is ​​adjusted. If the target characteristic value is the energy suppression ratio, then based on the updated test pulse, the updated local crosstalk matrix corresponding to each preset frequency is determined.

[0065] Among them, the update test pulse refers to the test signal used to recalculate the crosstalk matrix when the energy suppression ratio fails to meet the standard. It is the same or similar to the test pulse during the initial calibration in terms of type, frequency range and amplitude, and is used to obtain real-time response data under the current state.

[0066] Updating the local crosstalk matrix involves using the updated test pulse to re-excite and acquire a new test signal amplitude.

[0067] Based on this, in practical applications, if the root mean square error or the output target amplitude exceeds the standard, only the regularization parameter is adjusted without changing the crosstalk matrix; if the energy suppression ratio exceeds the standard, the data is retested based on the updated test pulse, and an updated local crosstalk matrix is ​​generated to replace the old matrix; thus achieving hierarchical correction with small error parameter tuning and large mismatch modeling.

[0068] Please refer to Figure 2 , Figure 2 A flowchart illustrating another waveform optimization method based on channel crosstalk provided in this application embodiment; as shown Figure 2 As shown, S1 establishes a mapping between the controlled object and the channel, clarifying the one-to-one correspondence between each qubit port and the control channel; S2 performs crosstalk calibration and constructs a crosstalk matrix, quantizing the coupling and crosstalk characteristics between channels at different frequencies based on the measured impulse response; S3 generates the target control waveform vector, clarifying the ideal control waveform of the target qubit and the zero response constraint of the non-target port; S4 performs joint precoding and pre-compensation solving, performing pre-distortion cancellation on the target waveform based on the crosstalk matrix, suppressing the inherent crosstalk of the channel in advance, and triggering regularization, amplitude limiting, or constraint optimization through real-time judgment of matrix ill-conditioning or output exceeding limits; S5 performs waveform shaping and hardware constraint adaptation, constraining the compensation waveform within the hardware output range through amplitude limiting and smoothing processing to ensure waveform continuity and engineering feasibility; S6 performs multi-channel synchronous output and executes quantum control, realizing parallel and precise control of multiple qubits; S7 updates the model parameters based on the measurement results, collects the residual signal between the actual response and the target waveform, extracts the signal feature value and compares it with the preset threshold, and adjusts the regularization parameters or recalibrates the crosstalk matrix differently for different causes of degradation.

[0069] Please refer to Figure 3 , Figure 3 A coupling diagram illustrating a waveform optimization method based on channel crosstalk provided in this application embodiment; as shown. Figure 3 As shown, a multi-input multi-output qubit control channel coupling model based on crosstalk matrix H is constructed. The input control channels (CH1-CH4) serve as excitation ends, inputting control signals into the crosstalk matrix H. The qubit response ends (Q1-Q4) receive the response signals after channel coupling. The elements H(i,j) of the crosstalk matrix H are defined as the coupling weights of the control channel CHj to the qubit Qi. The main diagonal elements characterize the self-response gain of the target channel to the target qubit, while the off-diagonal elements quantize the crosstalk coupling strength between different channels.

[0070] Based on the above steps, it can be seen that this application applies test pulses to each control channel based on multiple preset frequencies, thereby accurately acquiring the test signal amplitude of each quantum bit port according to the mapping relationship between the channel and the quantum bit port, realizing the accurate characterization of crosstalk coupling characteristics between channels at different frequencies; using the constructed local crosstalk matrix to perform crosstalk compensation on the target waveform and generate the target compensation waveform, after outputting, by comparing the actual response waveform with the ideal target waveform, the target residual signal is obtained, realizing the elimination of frequency-varying crosstalk interference between multiple quantum bit control channels based on the target compensation waveform, restoring the true response deviation of the quantum bit; extracting multiple signal feature values ​​through the target residual signal, comparing the feature values ​​with the corresponding preset thresholds and adaptively adjusting the local crosstalk matrix to optimize the crosstalk matrix, thereby correcting the crosstalk model mismatch problem caused by environmental and device drift, introducing the MIMO signal processing method in the field of communication into quantum measurement and control waveform generation, and realizing multi-channel joint pre-compensation through crosstalk matrix modeling, so as to compensate for crosstalk of quantum multi-channels based on the target compensation waveform, improving the accuracy of the output waveform, while reducing the dependence on additional hardware modifications, and facilitating deployment to existing measurement and control platforms.

[0071] Figure 4 This is a schematic diagram of a waveform optimization device based on channel crosstalk, provided as an embodiment of this application. Figure 4 As shown, this waveform optimization device based on channel crosstalk includes: a testing module, a determination module, a compensation module, and an adjustment module; wherein: The test module is used to apply test pulses to each control channel based on multiple preset frequencies, and determine the amplitude of the test signal output by each quantum bit port according to the channel mapping relationship. The determination module is used to determine the local crosstalk matrix corresponding to each preset frequency based on the amplitude of multiple test signals corresponding to each control channel, and apply a preset waveform to the target quantum bit port to obtain the corresponding target waveform; The compensation module is used to compensate the target waveform based on the local crosstalk matrix to obtain the target compensated waveform, and after outputting the target compensated waveform, to determine the target residual signal based on the corresponding response waveform and the target waveform. The adjustment module is used to determine multiple signal feature values ​​based on the target residual signal, and adjust the local crosstalk matrix according to the signal feature values ​​and their corresponding preset thresholds.

[0072] In this embodiment of the application, the determining module can also be specifically used for: For each preset frequency, the corresponding matrix rows are determined based on the port sequence corresponding to the quantum bit port, and the corresponding matrix columns are determined based on the channel sequence corresponding to the control channel. Based on the quantum bit ports and control channels corresponding to the amplitudes of each test signal, the corresponding rows and columns of the target matrix are determined, the corresponding elements of the target matrix are obtained, and the element values ​​corresponding to the elements of the target matrix are determined based on the amplitudes of the test signals, thus obtaining the local crosstalk matrix corresponding to the preset frequency.

[0073] In this embodiment of the application, the compensation module can also be specifically used for: Determine the corresponding compensation method; When the compensation method is full-frequency domain compensation, the average value is calculated based on all local crosstalk matrices to obtain the average crosstalk matrix, and the average crosstalk matrix is ​​determined as the target crosstalk matrix corresponding to the target waveform. When the compensation method is frequency domain segmented compensation, Fourier decomposition is performed on the target waveform to obtain the waveform components corresponding to each frequency band, and the target crosstalk matrix corresponding to each waveform component is determined based on the local crosstalk matrix. The target compensation waveform is determined based on the target crosstalk matrix, its corresponding transpose matrix, waveform, and regularization parameters.

[0074] In this embodiment of the application, the compensation module can also be specifically used for: Based on the target crosstalk matrix and its corresponding transpose matrix, waveform, and regularization parameters, the corresponding initial compensation waveform is calculated, and the initial compensation waveform is subjected to amplitude limiting processing based on a preset amplitude limiting threshold. Gaussian smoothing is applied to the initial compensation waveform after amplitude limiting to obtain the target compensation waveform.

[0075] In this embodiment of the application, the adjustment module can also be specifically used for: The target residual signal is sampled to obtain the residual value corresponding to each sampling point, and the corresponding root mean square error value is calculated based on the residual value. Based on the target waveform and the target compensated waveform, the uncompensated signal energy value and the compensated signal energy value of the output of the non-target quantum bit port are determined respectively, and the corresponding energy suppression ratio is calculated based on the ratio between the uncompensated signal energy value and the compensated signal energy value. Based on the root mean square error, energy suppression ratio, and target amplitude corresponding to the target compensation waveform, determine the corresponding signal characteristic values.

[0076] In this embodiment of the application, the adjustment module can also be specifically used for: Determine whether each signal feature value is less than its corresponding preset threshold; If at least one signal feature value is not less than the corresponding preset threshold, then the signal feature value not less than the corresponding preset threshold is determined as the target feature value, and the local crosstalk matrix is ​​adjusted based on the target feature value; If all values ​​are less than 1, then no adjustment is made to the local crosstalk matrix.

[0077] In this embodiment of the application, the adjustment module can also be specifically used for: Determine the target feature values; If the target eigenvalue is the root mean square error or the target magnitude, then the regularization parameter corresponding to the local crosstalk matrix is ​​adjusted. If the target characteristic value is the energy suppression ratio, then based on the updated test pulse, the updated local crosstalk matrix corresponding to each preset frequency is determined.

[0078] Figure 5 This is a schematic diagram of the structure of an apparatus for performing a waveform optimization method based on channel crosstalk according to an embodiment of this application. Figure 5 As shown, the device includes: The device may include one or more processors with processing cores, one or more computer-readable storage media such as memory, communication components, etc. The processor, memory, and communication components are connected via a bus.

[0079] In the specific implementation process, at least one processor executes computer execution instructions stored in memory, causing at least one processor to execute the waveform optimization method based on channel crosstalk as described above.

[0080] The specific implementation process of the processor can be found in the above method embodiments, and its implementation principle and technical effect are similar, so it will not be repeated here.

[0081] Furthermore, the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this application can be directly manifested as being executed by a hardware processor, or executed by a combination of hardware and software modules within the processor.

[0082] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0083] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0084] In some embodiments, a computer program product is also provided, including a computer program or instructions that, when executed by a processor, implement the steps in any of the above-described waveform optimization methods based on channel crosstalk.

[0085] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0086] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0087] Therefore, embodiments of this application provide a computer-readable storage medium storing multiple lines of program code that can be loaded by a processor to execute steps in any of the waveform optimization methods based on channel crosstalk provided in embodiments of this application.

[0088] The storage medium may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0089] According to one aspect of this application, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium.

[0090] Since the instructions stored in the storage medium can execute the steps in any of the waveform optimization methods based on channel crosstalk provided in the embodiments of this application, the beneficial effects that any of the waveform optimization methods based on channel crosstalk provided in the embodiments of this application can achieve can be realized. For details, please refer to the previous embodiments, which will not be repeated here.

[0091] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope of this application is indicated by the appended claims.

[0092] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. A waveform optimization method based on channel crosstalk, characterized in that, The method includes: Based on multiple preset frequencies, test pulses are applied to each control channel respectively, and the amplitude of the test signal output by each quantum bit port is determined according to the channel mapping relationship. Based on the amplitude values ​​of multiple test signals corresponding to each control channel, the local crosstalk matrix corresponding to each preset frequency is determined, and a preset waveform is applied to the target quantum bit port to obtain the corresponding target waveform; Based on the local crosstalk matrix, the target waveform is compensated to obtain the target compensated waveform. After outputting the target compensated waveform, the target residual signal is determined based on the corresponding response waveform and the target waveform. The compensation includes full-frequency domain compensation and frequency domain segmented compensation. The step of compensating the target waveform based on the local crosstalk matrix to obtain the target compensated waveform includes: Determine the corresponding compensation method; When the compensation method is the full-frequency domain compensation method, the average value is calculated based on all the local crosstalk matrices to obtain the average crosstalk matrix, and the average crosstalk matrix is ​​determined to be the target crosstalk matrix corresponding to the target waveform. When the compensation method is the frequency domain segmented compensation method, the target waveform is subjected to Fourier decomposition to obtain the waveform components corresponding to each frequency band, and the target crosstalk matrix corresponding to each waveform component is determined according to the local crosstalk matrix. The target compensation waveform is determined based on the target crosstalk matrix and its corresponding transpose matrix, waveform, and regularization parameters. Based on the target residual signal, multiple signal feature values ​​are determined, and the local crosstalk matrix is ​​adjusted according to the signal feature values ​​and their respective preset thresholds; wherein, adjusting the local crosstalk matrix according to the signal feature values ​​and their respective preset thresholds includes: Determine whether each of the signal feature values ​​is less than its corresponding preset threshold; If at least one of the signal feature values ​​is not less than the corresponding preset threshold, then the signal feature value that is not less than the corresponding preset threshold is determined as the target feature value, and the local crosstalk matrix is ​​adjusted based on the target feature value; If all values ​​are less than 1, then the local crosstalk matrix will not be adjusted.

2. The method according to claim 1, characterized in that, The step of determining the local crosstalk matrix corresponding to each preset frequency based on the amplitude values ​​of multiple test signals corresponding to each control channel includes: For each preset frequency, the corresponding matrix row is determined based on the port sequence corresponding to the quantum bit port, and the corresponding matrix column is determined based on the channel sequence corresponding to the control channel; Based on the quantum bit port and the control channel corresponding to each of the test signal amplitudes, the corresponding target matrix rows and columns are determined respectively to obtain the corresponding target matrix elements. Based on the test signal amplitudes, the element values ​​corresponding to the target matrix elements are determined to obtain the local crosstalk matrix corresponding to the preset frequency.

3. The method according to claim 1, characterized in that, The step of determining the target compensation waveform based on the target crosstalk matrix and its corresponding transpose matrix, waveform, and regularization parameters includes: Based on the target crosstalk matrix and its corresponding transpose matrix and waveform, and the regularization parameters, the corresponding initial compensation waveform is calculated, and the initial compensation waveform is subjected to amplitude limiting processing based on a preset amplitude limiting threshold. The initial compensation waveform after amplitude limiting is Gaussian smoothed to obtain the target compensation waveform.

4. The method according to claim 1, characterized in that, The determination of multiple signal feature values ​​based on the target residual signal includes: The target residual signal is sampled to obtain the residual value corresponding to each sampling point, and the corresponding root mean square error value is calculated based on the residual value. Based on the target waveform and the target compensated waveform, the uncompensated signal energy value and the compensated signal energy value output by the non-target quantum bit port are determined respectively, and the corresponding energy suppression ratio is calculated based on the ratio between the uncompensated signal energy value and the compensated signal energy value. The corresponding signal characteristic value is determined based on the root mean square error value, the energy suppression ratio, and the target amplitude corresponding to the target compensation waveform.

5. The method according to claim 1, characterized in that, The adjustment of the local crosstalk matrix based on the target feature value includes: Determine the target feature value; If the target feature value is the root mean square error value or the target magnitude, then the regularization parameter corresponding to the local crosstalk matrix is ​​adjusted. If the target feature value is the energy suppression ratio, then based on the updated test pulse, the updated local crosstalk matrix corresponding to each preset frequency is determined.

6. A waveform optimization device based on channel crosstalk, characterized in that, The device includes: The test module is used to apply test pulses to each control channel based on multiple preset frequencies, and determine the amplitude of the test signal output by each quantum bit port according to the channel mapping relationship. The determining module is used to determine the local crosstalk matrix corresponding to each preset frequency based on the amplitude values ​​of multiple test signals corresponding to each control channel, and apply a preset waveform to the target quantum bit port to obtain the corresponding target waveform; The compensation module is used to compensate the target waveform based on the local crosstalk matrix to obtain the target compensated waveform, and after outputting the target compensated waveform, to determine the target residual signal based on the corresponding response waveform and the target waveform; wherein, the compensation includes a full-frequency domain compensation method and a frequency domain segmented compensation method; The step of compensating the target waveform based on the local crosstalk matrix to obtain the target compensated waveform includes: Determine the corresponding compensation method; When the compensation method is the full-frequency domain compensation method, the average value is calculated based on all the local crosstalk matrices to obtain the average crosstalk matrix, and the average crosstalk matrix is ​​determined to be the target crosstalk matrix corresponding to the target waveform. When the compensation method is the frequency domain segmented compensation method, the target waveform is subjected to Fourier decomposition to obtain the waveform components corresponding to each frequency band, and the target crosstalk matrix corresponding to each waveform component is determined according to the local crosstalk matrix. The target compensation waveform is determined based on the target crosstalk matrix and its corresponding transpose matrix, waveform, and regularization parameters. An adjustment module is used to determine multiple signal feature values ​​based on the target residual signal, and adjust the local crosstalk matrix according to the signal feature values ​​and their respective preset thresholds; wherein, adjusting the local crosstalk matrix according to the signal feature values ​​and their respective preset thresholds includes: Determine whether each of the signal feature values ​​is less than its corresponding preset threshold; If at least one of the signal feature values ​​is not less than the corresponding preset threshold, then the signal feature value that is not less than the corresponding preset threshold is determined as the target feature value, and the local crosstalk matrix is ​​adjusted based on the target feature value; If all values ​​are less than 1, then the local crosstalk matrix will not be adjusted.

7. A computer device, characterized in that, include: One or more processors; Memory; One or more programs, wherein the programs are stored in memory and configured to be executed by one or more processors, the programs being configured to perform the method as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores program code that can be called by a processor to perform the method as described in any one of claims 1 to 5.

Images