Method, device and storage medium for processing signal crosstalk

Abstract

The application discloses a signal crosstalk processing method and device and a storage medium, relates to the technical field of digital signal transmission processing, and comprises the following steps: acquiring a to-be-tested mixed signal in an equalization system; identifying the crosstalk type of a crosstalk signal by using a first signal processing model, determining a suppression signal of the crosstalk signal by using a second signal processing model; and using the suppression signal in a target joint cost model to offset at least part of the crosstalk signal and determine a residual error signal until a residual error signal meeting certain signal conditions is obtained, so as to solve the technical problems of limited space, high design cost and complexity, slow convergence, poor effect, inability to process random noise, inability to intelligently judge the crosstalk type and adaptively adjust the strategy, accurately distinguish the crosstalk type, improve the accuracy of signal problem identification, be suitable for various complex high-speed signal transmission scenes, and achieve the technical effect of optimizing the balance between performance and cost.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of digital signal transmission processing, and in particular to a signal crosstalk processing method and device and storage medium. BACKGROUND

[0002] In related technologies, to solve the problem of signal crosstalk, the crosstalk can be reduced by increasing the distance between transmission lines or adding shielding; or the crosstalk can be actively cancelled by using time-domain adaptive filtering or fixed frequency-domain notch filtering.

[0003] However, in related technologies, the space available for transmission lines is increasingly limited, and the crosstalk problem cannot be solved simply by increasing the distance; adding shielding not only increases the design cost and complexity, but also is difficult to implement in some scenarios; time-domain filtering converges slowly and has poor effect on strong periodic crosstalk; frequency-domain notch filtering easily causes signal phase distortion and cannot handle random noise. In addition, related technologies also lack intelligent judgment of the type of crosstalk, and cannot adaptively adjust the strategy, which performs poorly in the face of complex mixed-mode crosstalk, and needs to be improved. SUMMARY

[0004] The present application provides a signal crosstalk processing method, device and storage medium to at least solve the problems in related technologies that increasing the distance uses limited space; adding shielding has high design cost and complexity, and limited use scenarios; time-domain filtering converges slowly and has poor effect on strong periodic crosstalk; frequency-domain notch filtering easily causes signal phase distortion and cannot handle random noise. In addition, none of them can intelligently judge the type of crosstalk and adaptively adjust the strategy, resulting in poor performance in the face of complex mixed-mode crosstalk.

[0005] The present application provides a signal crosstalk processing method, comprising: obtaining a to-be-tested mixed signal in an equalization system; inputting the to-be-tested mixed signal into a pre-constructed signal processing model, to identify the crosstalk type of at least one crosstalk signal in the to-be-tested mixed signal by using a first signal processing model in the pre-constructed signal processing model, and determine an inhibition signal of the at least one crosstalk signal by using a second signal processing model in the pre-constructed signal processing model; inputting the inhibition signal, the crosstalk type and the to-be-tested mixed signal into a target joint cost model, to cancel at least part of the crosstalk signal by using the inhibition signal, and determine a residual error signal between the inhibition signal and the crosstalk signal based on the at least part, until a residual error signal meeting a preset signal condition is obtained.

[0006] The application further provides a signal crosstalk processing device, comprising: an acquisition module, configured to acquire a to-be-tested mixed signal in an equalization system; an output module, configured to input the to-be-tested mixed signal into a pre-constructed signal processing model, so as to identify a crosstalk type of at least one crosstalk signal in the to-be-tested mixed signal by using a first signal processing model in the pre-constructed signal processing model, and determine an inhibition signal of the at least one crosstalk signal by using a second signal processing model in the pre-constructed signal processing model; and a generation module, configured to input the inhibition signal, the crosstalk type and the to-be-tested mixed signal into a target joint cost model, so as to cancel at least part of the crosstalk signal by using the inhibition signal, and determine a residual error signal between the inhibition signal and the crosstalk signal based on the at least part, until a residual error signal meeting a preset signal condition is obtained.

[0007] The application further provides an electronic device, comprising: a memory, configured to store a computer program; and a processor, configured to execute the computer program to implement the steps of the signal crosstalk processing method.

[0008] The application further provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the signal crosstalk processing method.

[0009] The application further provides a computer program product, comprising a computer program, and the computer program is executed by a processor to implement the steps of the signal crosstalk processing method.

[0010] By the application, the first signal processing model in the pre-constructed signal processing model is used to identify the crosstalk type of the crosstalk signal in the to-be-tested mixed signal, the second signal processing model is used to determine the inhibition signal of the crosstalk signal, then the inhibition signal is used to cancel at least part of the crosstalk signal in the target joint cost model, and the residual error signal between the inhibition signal and the crosstalk signal is determined, until the residual error signal meeting the certain signal condition is obtained, so that the technical problem that the use space is limited when the spacing is increased, the design cost and complexity of the shielding are high, the use scene is limited, the time domain filtering converges slowly and has poor effect on the strong periodic crosstalk, the frequency domain wave trap easily causes signal phase distortion and cannot process random noise, and the like can be solved. In addition, the crosstalk type cannot be intelligently judged and the strategy cannot be adaptively adjusted, which leads to poor performance in front of complex mixed mode crosstalk. The application achieves the technical effects that the signal quality is comprehensively judged, the crosstalk type is accurately distinguished, the accuracy of signal problem identification is improved, misjudgment and omission are avoided, the influence of crosstalk on signals is reduced to the greatest extent, the signal integrity is improved, the application is suitable for various complex high-speed signal transmission scenes, the dependence on hardware indicators is reduced, the performance and cost are optimized and balanced, and the like. BRIEF DESCRIPTION OF DRAWINGS

[0011] In order to more clearly illustrate the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. Obviously, the drawings described in the following description are only some embodiments of the present application. Based on the drawings in the embodiments, all other drawings obtained by those skilled in the art without creative effort are within the scope of the present application.

[0012] Figure 1 The flow chart of the signal crosstalk processing method according to an embodiment of the present application is shown in FIG. 2.

[0013] Figure 2 The flow chart of determining the signal quality of the to-be-tested mixed signal according to an embodiment of the present application is shown in FIG. 3.

[0014] Figure 3 The flow chart of distinguishing the reflection signal and the crosstalk signal according to an embodiment of the present application is shown in FIG. 4.

[0015] Figure 4 The flow chart of the to-be-tested mixed signal re-detection according to an embodiment of the present application is shown in FIG. 5.

[0016] Figure 5 The flow chart of the calibration stage of the signal crosstalk processing method according to an embodiment of the present application is shown in FIG. 6.

[0017] Figure 6 The flow chart of the normal working stage of the signal crosstalk processing method according to an embodiment of the present application is shown in FIG. 7.

[0018] Figure 7 The block schematic diagram of the signal crosstalk processing device according to an embodiment of the present application is shown in FIG. 8.

[0019] Reference signs:

[0020] Wherein, 10 - signal crosstalk processing device; 100 - first acquisition module, 200 - output module, 300 - first generation module. DETAILED DESCRIPTION

[0021] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of the present application.

[0022] It should be noted that in the description of the present application, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or further includes elements inherent to such a process, method, article or device. The terms "first", "second" and the like in the present application are used to distinguish similar objects, not to describe a specific order or sequence.

[0023] In order to enable those skilled in the art to better understand the present application, the present application will be further described in detail below in combination with the drawings and specific embodiments.

[0024] Before introducing the signal crosstalk processing method proposed in the embodiments of the present application, the related technical terms involved in the embodiments of the present application are explained.

[0025] It can be understood that, with the rapid development of electronic information technology, the signal transmission rate is increasing at an alarming rate. From hundreds of megahertz, gigabit transmission, to gigabit and even higher speed, the data transmission efficiency is significantly improved. However, with the increase of signal rate, signal integrity problems have gradually become the key factors restricting high-speed transmission performance. Among them, signal attenuation, signal reflection and signal crosstalk are three typical signal integrity problems.

[0026] In terms of signal attenuation, as the signal frequency increases, the energy loss of the signal in the transmission medium is intensified. For example, in the wiring of the printed circuit board, the skin effect of the copper foil makes the high-frequency signal current mainly concentrated on the surface of the conductor, resulting in increased resistance and increased signal energy loss; the dielectric loss of the dielectric material also absorbs signal energy, causing signal attenuation. In the related art, the code interval crosstalk problem caused by signal attenuation can be solved by improving the electrical grade of the printed circuit board material, using materials with low dielectric constant and low dielectric loss, and combining with equalization technology. Code interval crosstalk refers to the fact that due to the attenuation and time delay of the signal in the transmission process, the signal waveform of the current code element and the signal waveform of the subsequent code element overlap each other, thereby affecting the correct judgment of the signal. Equalization technology compensates for the attenuated signal, adjusts the amplitude and phase of the signal, reduces the code interval crosstalk, and ensures the accuracy of the signal.

[0027] In terms of signal reflection, when the signal encounters impedance mismatching in the transmission process, such as length change of the transmission line, via, connector, etc., part of the signal energy will be reflected back to the source. The reflected signal and the original signal superimpose, which will cause waveform distortion and affect the signal quality. In the related art, the impedance and resistance matching can be optimized by simulation to better control the reflection problem. In the design stage, the impedance of the transmission line is accurately calculated and optimized by using electromagnetic simulation software to ensure the impedance continuity of the signal in the transmission process and reduce the reflection. In addition, the digital feedback equalizer can also process the reflected signal within a certain time, and the reflected signal is compensated in the reverse direction through the feedback mechanism to improve the integrity of the signal.

[0028] In terms of signal crosstalk, crosstalk refers to when multiple signals are transmitted in adjacent transmission lines at the same time, due to electromagnetic coupling effect, the energy of one signal will be coupled to the adjacent transmission line, causing interference to other signals. In the related art, the crosstalk is mainly reduced by increasing the distance between the transmission lines or adding shielding. Increasing the distance can reduce the electromagnetic coupling strength between signals, however, the related technology faces great challenges. Under the current trend of high-density design, electronic devices pursue smaller size and higher integration, and the space left for the transmission line is increasingly limited, which cannot solve the crosstalk problem by simply increasing the distance. Although adding shielding can isolate the electromagnetic interference between signals, it will increase the design cost and complexity, and it is difficult to implement in some space-limited scenarios. The related equalization algorithm mainly compensates for signal attenuation, and has little effect on improving the crosstalk problem.

[0029] Embodiments of the present application provide a signal crosstalk processing method. The method is described in detail in combination with the execution process of the signal crosstalk processing method.

[0030] Specifically, Figure 1 A flowchart of a signal crosstalk processing method according to an embodiment of the present application is provided.

[0031] As Figure 1 shown, the signal crosstalk processing method includes the following steps:

[0032] In step S101, a to-be-tested mixed signal in an equalization system is obtained.

[0033] It can be understood that in the embodiments of the present application, the to-be-tested mixed signal can be understood as multiple signals being transmitted in adjacent transmission lines at the same time, but due to electromagnetic coupling effect, the energy of one signal will be coupled to the adjacent transmission line, causing interference to other signals. It can be multiple digital signals, or multiple analog signals, or a mixed signal of digital signals and analog signals. The specific setting can be made by those skilled in the art according to the actual situation, and the present application does not make specific limitations.

[0034] Further, the embodiment of the present application can obtain the to-be-tested mixed signal in the equalization system through an oscilloscope, a spectrum analyzer and a multimeter, and the present application does not make specific limitation.

[0035] In some embodiments, the embodiment of the present application can obtain the to-be-tested mixed signal in the equalization system.

[0036] Optionally, in one embodiment of the present application, before the to-be-tested mixed signal is input into the pre-constructed signal processing model, the method further comprises: determining a measured value and a target value corresponding to at least one signal parameter of the to-be-tested mixed signal based on the at least one signal parameter; calculating a difference value between the measured value and the target value, and calculating a ratio value of the difference value and the target value based on the difference value, so as to determine an index margin value corresponding to the at least one signal parameter according to the ratio value; judging whether the index margin value is less than a preset threshold value; if the index margin value is less than the preset threshold value, obtaining a signal attenuation value of the to-be-tested mixed signal, and detecting whether the signal attenuation value is less than a preset attenuation value; when it is detected that the signal attenuation value is less than the preset attenuation value, determining that there is at least one crosstalk signal in the to-be-tested mixed signal, and allowing the to-be-tested mixed signal to be input into the pre-constructed signal processing model; if the index margin value is greater than or equal to the preset threshold value, determining that there is no at least one crosstalk signal in the to-be-tested mixed signal, and prohibiting the to-be-tested mixed signal from being input into the pre-constructed signal processing model.

[0037] It can be understood that, before the to-be-tested mixed signal is input into the pre-constructed signal processing model, the embodiment of the present application can determine whether there is a crosstalk signal in the to-be-tested mixed signal by calculating an index margin value of different signal parameters of the to-be-tested mixed signal, wherein certain conditions can be set by a person skilled in the art according to actual conditions, and the present application does not make specific limitation.

[0038] The signal parameter can be but is not limited to a signal amplitude, a signal-to-noise ratio, a bit error rate, an eye diagram, etc., and the present application does not make specific limitation.

[0039] In some embodiments, the embodiments of the present application can first acquire measured values and target values corresponding to different signal parameters of the to-be-tested mixed signal, then calculate the difference between the measured values and the target values, and based on the difference, calculate the ratio of the difference and the target value, so as to determine the index margin value corresponding to the different signal parameters according to the ratio, and judge whether the index margin value is less than a certain threshold value; if it is less than, acquire the signal attenuation value of the to-be-tested mixed signal, and detect whether the signal attenuation value is less than a certain attenuation value, when it is detected that the signal attenuation value is less than a certain attenuation value, identify the crosstalk signal, and allow the to-be-tested mixed signal to be input to the pre-constructed signal processing model, otherwise, it is determined that there is no crosstalk signal in the to-be-tested mixed signal, and the to-be-tested mixed signal is prohibited from being input to the pre-constructed signal processing model. The certain margin value and the certain attenuation value can be set by a person skilled in the art according to the actual situation, and the present application does not make specific limitations. The embodiments of the present application can be combined with Figure 2 As shown in the figure, the signal quality of the to-be-tested mixed signal is determined, and the main content can be:

[0040] Step S201: Acquire the to-be-tested mixed signal.

[0041] Step S202: Calculate the corresponding index margin value based on different signal parameters.

[0042] Step S203: Judge whether the index margin value is less than a certain threshold value.

[0043] Wherein, if it is less than, execute step S205; otherwise, execute step S204.

[0044] Step S204: No need to process.

[0045] Step S205: Acquire the signal attenuation value of the to-be-tested mixed signal.

[0046] Step S206: Judge whether the signal attenuation value is less than a certain attenuation value.

[0047] Wherein, if it is less than, execute step S207; otherwise, execute step S208.

[0048] Step S207: Identify the crosstalk signal.

[0049] Step S208: Modify the current design.

[0050] Wherein, the embodiments of the present application can improve the electrical grade of the printed circuit board, or use a signal regenerator or other relay device, which can be set by a person skilled in the art according to the actual situation, and the present application does not make specific limitations.

[0051] For example, the embodiment of the present application processes the to-be-tested mixed signal input into the equalization system, collects the signal amplitude, signal-to-noise ratio, bit error rate, eye diagram and the like of the to-be-tested mixed signal in real time, takes the signal amplitude as an example, acquires the measured value of the voltage of the to-be-tested mixed signal through a high-precision voltage measurement circuit, compares it with the target amplitude value, calculates the amplitude margin, takes the 20% margin value as a certain threshold, judges whether the amplitude margin is less than 20%, if less than 20%, further analyzes the cause of the signal quality problem, if greater than or equal to 20%, it indicates that the reflection, crosstalk and other factors in the current system have little effect on the signal quality, which is not enough to cause the signal quality problem, no additional processing is needed, and the to-be-tested mixed signal is prohibited from being input into the pre-constructed signal processing model.

[0052] It should be noted that the embodiment of the present application can also take the signal-to-noise ratio as an example, measure and calculate the useful signal power and noise power in the signal through a dedicated signal analysis chip, compare it with the target power value, calculate the power margin, and then judge whether the power margin is less than 20% to analyze the signal quality; it can also comprehensively consider multiple signal parameters to calculate the margin value of the to-be-tested mixed signal after passing through the equalization system, which can be set by the person skilled in the art according to the actual situation, and the present application does not make specific limitations.

[0053] Further, in the case where the index margin value is less than 20%, the signal attenuation value is further analyzed, and if it is determined through detection and analysis that the signal attenuation value in the current design is greater than or equal to a certain attenuation value, the equalization system cannot effectively compensate for such a large attenuation, and the design can be modified. The specific measures include improving the electrical grade of the printed circuit board, selecting a printed circuit board material with lower dielectric loss and better electrical performance to reduce signal attenuation during transmission; or using a signal regenerator and other relay equipment, which can reshape, amplify and clock restore the attenuated signal, enhance the signal strength, compensate for signal attenuation, and thus improve the signal quality.

[0054] In addition, if it is determined that the signal attenuation value in the current design is less than a certain attenuation value, that is, the attenuation value is within a reasonable standard interval after the equalization system processing, but the signal quality still has problems, and the equalization algorithm itself is excluded, it can be determined that there is a reflection or crosstalk problem in the current system, which needs to be further distinguished and processed.

[0055] Among them, the embodiment of the present application can analyze the signal from two dimensions of frequency domain and time domain to accurately distinguish the reflection signal and the crosstalk signal, combined with Figure 3 As shown in FIG. 6, the main content can be:

[0056] Step S301: distinguish the reflection signal and the crosstalk signal.

[0057] Step S302: frequency domain discrimination.

[0058] In the embodiment of the application, the Fourier transform is first performed on the mixed signal to be measured to convert the time domain signal into a frequency domain signal and obtain the frequency spectrum data of the signal. By analyzing the periodicity and frequency distribution of the frequency spectrum data, it is determined whether the signal is a reflection signal or a crosstalk signal. If the frequency spectrum presents obvious periodicity and the period is related to the electrical length and other parameters of the transmission line, it is determined to be a reflection signal. If the frequency spectrum is a narrow-band noise signal of fixed frequency, it is preliminarily determined that the interference source is crosstalk of a periodic signal. If the frequency spectrum is a wide-band noise signal without periodicity, it is preliminarily determined that the interference source is crosstalk of a non-periodic signal.

[0059] Step S303: time domain discrimination.

[0060] In the embodiment of the application, the time domain waveform of the signal is monitored in real time, and the time domain waveform of the mixed signal to be measured is collected by using a high-speed oscilloscope and other devices. It is observed whether there is a waveform with time delay repetition and attenuation. If there is, it is a reflection signal. If there is a periodic interference component in the waveform but no attenuation, it is periodic crosstalk. If there is no regular interference waveform, it is non-periodic crosstalk.

[0061] Step S304: input to a pre-constructed signal processing model.

[0062] In the embodiment of the application, the reflection signal and the crosstalk signal are identified through comprehensive analysis of the frequency domain and the time domain, and the mixed signal to be measured is allowed to be input to the pre-constructed signal processing model.

[0063] Before the mixed signal to be measured is input to the pre-constructed signal processing model, the embodiment of the application can determine whether there is a crosstalk signal in the mixed signal to be measured according to the index margin value of different signal parameters of the mixed signal to be measured and the signal attenuation value of the mixed signal to be measured. When there is a crosstalk signal, the mixed signal to be measured is allowed to be input to the pre-constructed signal processing model. Otherwise, the mixed signal to be measured is prohibited from being input to the pre-constructed signal processing model. The crosstalk signal is identified in advance to ensure the signal quality of the input signal processing model, improve the accuracy of signal processing, and optimize the resource utilization efficiency.

[0064] Optionally, in an embodiment of the present application, before the to-be-tested mixed signal is input into the pre-constructed signal processing model, further comprising: obtaining a plurality of reference crosstalk signals; performing fast Fourier transform on the reference crosstalk signals to obtain power spectral densities corresponding to the reference crosstalk signals, and calculating a difference between the power spectral density amplitude and the background noise discrete spectrum peak; determining whether the difference is greater than a first threshold value; if the difference is greater than the first threshold value, determining that the reference crosstalk signal is a periodic crosstalk signal, otherwise, determining that the reference crosstalk signal is an aperiodic crosstalk signal; and constructing a first signal processing model in the signal processing model based on the periodic crosstalk signal and the aperiodic crosstalk signal to identify the crosstalk type of the reference crosstalk signal.

[0065] It can be understood that the embodiment of the present application can identify the crosstalk type of the reference crosstalk signal through the first signal processing model.

[0066] In some embodiments, the embodiment of the present application can perform fast Fourier transform on a plurality of reference crosstalk signals respectively before inputting the to-be-tested mixed signal into the pre-constructed signal processing model, thereby obtaining power spectral densities corresponding to different reference crosstalk signals, identifying the crosstalk type of the reference crosstalk signal by calculating the difference between the power spectral density amplitude and the background noise discrete spectrum peak, and constructing the first signal processing model.

[0067] For example, the embodiment of the present application can determine the crosstalk type by performing fast Fourier transform on the reference crosstalk signal and detecting the discrete spectrum peak, wherein when the difference between the power spectral density corresponding to the reference crosstalk signal and the background noise discrete spectrum peak is greater than a first threshold value, it is determined that the reference crosstalk signal is a periodic crosstalk signal, and a center frequency estimation value is output, otherwise, it is an aperiodic crosstalk signal, wherein the first threshold value can be set by a person skilled in the art according to the actual situation, and the present application does not make specific limitations.

[0068] Further, the embodiment of the present application can construct the first signal processing model based on the periodic crosstalk signal and the aperiodic crosstalk information, and then identify the crosstalk type of the reference crosstalk signal by using the first signal processing model.

[0069] The embodiment of the present application compares the difference between the power spectral density amplitude and the background noise discrete spectrum peak by performing fast Fourier transform on the reference crosstalk signal, determines the crosstalk type of the reference crosstalk signal, and then constructs the first signal processing model, taking the difference between the power spectral density amplitude and the background noise discrete spectrum peak as the judgment standard, which can effectively exclude the interference of background noise, avoid misjudging noise as crosstalk signal, improve the accuracy of crosstalk detection, construct the first signal processing model based on the periodic crosstalk signal and the aperiodic crosstalk signal, so that the model can be specifically used for identifying different types of crosstalk signals, improve the identification ability of the model to crosstalk signals, and better adapt to various environments.

[0070] Optionally, in an embodiment of the present application, before the to-be-tested mixed signal is input into the pre-constructed signal processing model, further comprising: acquiring a plurality of probe crosstalk signals; extracting time domain features and frequency domain features of the probe crosstalk signals based on the crosstalk types of the probe crosstalk signals; simultaneously performing time domain processing and frequency domain processing on the time domain features and the frequency domain features to obtain a first suppression signal of the probe crosstalk signals in the time domain and a second suppression signal of the probe crosstalk signals in the frequency domain; and calculating a suppression signal of the probe crosstalk signals based on the first suppression signal and the second suppression signal to construct a second signal processing model in the signal processing model.

[0071] It can be understood that the embodiment of the present application can use the second signal processing model to process the probe crosstalk signal in parallel, that is, to calculate the suppression signal of the probe crosstalk signal according to the crosstalk type of the probe crosstalk signal, and then eliminate the probe crosstalk signal.

[0072] In some embodiments, before the to-be-tested mixed signal is input into the pre-constructed signal processing model, the embodiment of the present application can extract the time domain features and the frequency domain features of the probe crosstalk signal based on the crosstalk type of the probe crosstalk signal, and then simultaneously perform time domain processing and frequency domain processing to obtain the first suppression signal of the probe crosstalk signal in the time domain and the second suppression signal of the probe crosstalk signal in the frequency domain, and then calculate the suppression signal of the probe crosstalk signal, and then construct the second signal processing model in the signal processing model.

[0073] Among them, the embodiment of the present application can use corresponding processing methods according to different types of crosstalk signals: for periodic crosstalk signals, call frequency domain notch branch, such as parameter configurable notch filter; for non-periodic crosstalk signals, according to edge synchronous and asynchronous conditions, call time domain adaptive filtering branch, such as adaptive finite impulse response filter based on least mean square algorithm, it should be noted that the frequency domain notch branch and the time domain adaptive filtering branch are parameter adjustable, which can be optimized and adjusted according to the actual crosstalk signal to achieve the best crosstalk suppression effect.

[0074] It should be noted that in the embodiment of the present application, the frequency domain notch branch is realized by a double second-order infinite impulse response filter with adjustable center frequency, bandwidth and depth, which is responsible for deep suppression of specific periodic interference components; the time domain adaptive filtering branch is realized by a 32-order finite impulse response filter using the least mean square algorithm, and its filter coefficient vector is W, which is responsible for tracking and canceling wideband random interference.

[0075] The embodiment of the application can extract the time domain features and the frequency domain features of the detection crosstalk signal based on the crosstalk type of the detection crosstalk signal, and then obtain the suppression signals of the detection crosstalk signal in the time domain and in the frequency domain, so as to construct a second signal processing model. Through comprehensive feature extraction and utilization, the crosstalk signal is described from multiple dimensions, the characteristics of the crosstalk signal can be more comprehensively and accurately grasped, a solid foundation is provided for subsequent accurate processing, the double processing is complementary to each other, the crosstalk signal can be more effectively suppressed, the accuracy of signal processing is improved, and the model is better adapted to the crosstalk suppression demand in a specific scene, the pertinence and effectiveness of the model are improved, and the generalization ability of the model is enhanced.

[0076] Optionally, in an embodiment of the application, based on the first suppression signal and the second suppression signal, the suppression signal of the detection crosstalk signal is calculated, including: based on the first suppression signal and the second suppression signal, determining a first weight coefficient corresponding to the first suppression signal and a second weight coefficient corresponding to the second suppression signal; calculating a first product between the first suppression signal and the first weight coefficient, and a second product between the second suppression signal and the second weight coefficient, respectively; and summing the first product and the second product to obtain the suppression signal.

[0077] In some embodiments, the embodiment of the application can obtain the suppression signal by determining the first weight coefficient of the first suppression signal, the first product between the first suppression signal and the first weight coefficient, the second weight coefficient of the second suppression signal, and the second product between the second suppression signal and the second weight coefficient.

[0078] For example, the embodiment of the application can use a weighted synthesizer to fuse the first suppression signal and the second suppression signal according to the dynamic weight coefficients , to generate the suppression signal , wherein, is the first suppression signal, is the second weight coefficient, is the second suppression signal, is the first weight coefficient.

[0079] The embodiment of the application can obtain the suppression signal by the first suppression signal, the first weight coefficient corresponding to the first suppression signal, the second suppression signal, and the second weight coefficient corresponding to the second suppression signal. By assigning different weight coefficients to the two suppression signals and summing them, the advantages of time domain and frequency domain processing can be integrated, the detection crosstalk signal can be more comprehensively suppressed, the signal suppression effect is optimized, the introduction of the weight coefficient makes it possible to flexibly adjust the contribution degree of the first suppression signal and the second suppression signal in the final suppression signal according to actual needs, better adapt to crosstalk signals of different types and intensities, improve the pertinence and effectiveness of signal suppression, and enhance the adaptability and robustness of the model.

[0080] In step S102, the to-be-tested mixed signal is input into the pre-constructed signal processing model, so as to identify the crosstalk type of at least one crosstalk signal in the to-be-tested mixed signal by using a first signal processing model in the pre-constructed signal processing model, and determine the suppression signal of the at least one crosstalk signal by using a second signal processing model in the pre-constructed signal processing model.

[0081] It can be understood that the pre-constructed signal processing model in the embodiment of the present application includes the first signal processing model and the second signal processing model, wherein the first signal processing model can identify the crosstalk type of the crosstalk signal, and the second signal processing model can determine the suppression signal of the crosstalk signal.

[0082] Further, in some embodiments, the embodiment of the present application can input the to-be-tested mixed signal into the pre-constructed signal processing model, and then identify the crosstalk type of the crosstalk signal by using the first signal processing model, and determine the suppression signal of the crosstalk signal by using the second signal processing model.

[0083] Optionally, in an embodiment of the present application, determining the suppression signal of the at least one crosstalk signal by using the second signal processing model in the pre-constructed signal processing model includes: obtaining a coupling response signal generated by the crosstalk signal under no-load of the to-be-tested mixed signal; calculating an initial time-domain impulse response and / or an initial frequency-domain transfer function of the crosstalk signal based on the crosstalk signal and the coupling response signal; initializing the second signal processing model by using the initial time-domain impulse response and / or the initial frequency-domain transfer function to obtain an initialized second signal processing model, and obtaining the suppression signal by using the initialized second signal processing model.

[0084] It can be understood that, when determining the suppression signal by using the second signal processing model, the embodiment of the present application can initialize the second signal processing model by using the initial time-domain impulse response and / or the initial frequency-domain transfer function, and then determine the suppression signal by using the initialized second signal processing model.

[0085] In some embodiments, the embodiment of the present application can obtain the coupling response signal generated by the crosstalk signal under no-load of the to-be-tested mixed signal, and then calculate the initial time-domain impulse response and / or the initial frequency-domain transfer function of the crosstalk signal, and initialize the second signal processing model by using the initial time-domain impulse response and / or the initial frequency-domain transfer function, and then obtain the initialized second signal processing model, and obtain the suppression signal by using the initialized second signal processing model.

[0086] For example, when initializing the second signal processing model, the probe signal generator can be used to generate a corresponding crosstalk signal, and the crosstalk signal can be input into the interference signal link in the calibration mode. The high-precision analog-to-digital converter can synchronously collect signals at the victim signal link end. Since the system has not started normal data transmission at this time, the content on the victim signal link is almost completely the coupling response signal generated by the crosstalk signal through the coupling channel (may also contain a small amount of background noise). Further, the initial time-domain impulse response and / or the initial frequency-domain transfer function can be calculated based on the crosstalk signal and the coupling response signal collected from the victim signal link by using the least square method or the correlation method, as the priori knowledge of the system, to greatly improve the convergence speed of the subsequent adaptive process, and initialize the second signal processing model, such as initializing the filter coefficient vector W of the time-domain adaptive filter branch, such as directly inputting the crosstalk signal and the coupling response signal into the least mean square algorithm to allow it to quickly iterate several times, at this time, W can be used as the initial value; initializing the center frequency and bandwidth of the frequency-domain notch branch, and the present application does not make specific limitations.

[0087] The least square method can be directly calculated in the frequency domain or the time domain for a short sequence. In the frequency domain, the calculation formula can be H_coupling_initial=FFT(d_probe) / FFT(x_probe), where d_probe is the coupling response signal, x_probe is the crosstalk signal, and H_coupling_initial is the initial frequency-domain transfer function.

[0088] The correlation method can calculate the cross-correlation function of the crosstalk signal and the coupling response signal, where the cross-correlation function can directly reflect the impulse response of the channel.

[0089] In addition, in the embodiments of the present application, the crosstalk signal generated by the signal generator should have good autocorrelation characteristics and a wide frequency spectrum, so as to fully characterize the coupling channel. It can be a pseudo-random sequence (such as an m sequence, a pseudo-random binary sequence with a length of 1024 points, which is not limited in the present application), which has a spectrum similar to white noise and can stimulate the response of the channel at all frequencies. It can also be a linear frequency modulation signal (such as a linear frequency modulation signal, which is not limited in the present application), a signal with a linearly changing frequency over time, which can also cover a wide frequency band

[0090] The embodiment of the application can obtain the initial time domain impulse response and / or initial frequency domain transfer function of the crosstalk signal generated under the idle state of the to-be-tested mixed signal, initialize the second signal processing model to obtain the initialized second signal processing model, and obtain the suppression signal by using the initialized second signal processing model, so that the coupling of the crosstalk signal in the actual transmission environment can be directly reflected, the model is more closely fitted to the actual crosstalk scene, the deviation caused by the initialization based on the theoretical assumption or the simplified model is avoided, the precision of the model initialization is improved, the adaptability and generalization ability of the model are enhanced, and the model is adapted to different crosstalk scenes.

[0091] In step S103, the suppression signal, the crosstalk type and the to-be-tested mixed signal are input into the target joint cost model, at least part of the crosstalk signal is cancelled by using the suppression signal, and the residual error signal between the suppression signal and the crosstalk signal is determined based on the at least part until the residual error signal meeting the preset signal condition is obtained. The expression of the target joint cost model can be but is not limited to:

[0092]

[0093] , is a residual error signal; is a coefficient vector of a time domain filter; is a parameter vector of a frequency domain notch filter; is a fusion weight vector of a weighted synthesizer; , , is a regularization factor, represents a variation of the parameters corresponding to adjacent periods.

[0094] In some embodiments, the embodiment of the application can cancel at least part of the crosstalk signal by using the suppression signal based on the target joint cost model, and determine the residual error signal between the suppression signal and the crosstalk signal based on the at least part until the residual error signal meeting a certain signal condition is obtained. The certain signal condition can be set by a person skilled in the art according to the actual situation, and the application does not make specific limitations. The expression of the target joint cost model can be but is not limited to:

[0095]

[0096] , is a residual error signal; is a coefficient vector of a time domain filter; is a parameter vector of a frequency domain notch filter, wherein, is a center frequency, is a gain, is a quality factor.​​ is a fusion weight vector of the weighted combiner, , , is a regularization factor for constraining mutation of parameters and ensuring system stability; represents a variation of the corresponding parameter in the adjacent period.

[0097] In addition, in some embodiments, the expression of the target joint cost model in the embodiments of the present application can be, but is not limited to:

[0098] .

[0099] Optionally, in an embodiment of the present application, the suppression signal, the crosstalk type and the to-be-tested mixed signal are input into the target joint cost model to offset at least part of the crosstalk signal by using the suppression signal, and to determine a residual error signal between the suppression signal and the crosstalk signal based on the at least part, including: calculating an initial residual error signal of the suppression signal and the crosstalk signal; obtaining an initial joint cost value of the to-be-tested mixed signal based on the initial residual error signal; updating parameters of the target joint cost model based on the initial joint cost value, the crosstalk type and a step factor until the updated target joint cost model meets a preset model condition, so as to obtain the residual error signal meeting the preset signal condition by using the updated target joint cost model.

[0100] It can be understood that in the embodiments of the present application, the parameter updating of the target joint cost model adopts the gradient descent method, which can be, but is not limited to, expressed as:

[0101] ,

[0102] wherein, , represents a set of parameters to be optimized , is a corresponding step factor.

[0103] Further, taking the coefficient vector updating of the time domain filter as an example, the expansion thereof can be, but is not limited to, expressed as:

[0104] ,

[0105] wherein, is a time delay input vector composed of reference signals, takes a value of 0.001-0.01, takes a value of 0.01-0.1.

[0106] ​In some embodiments, the embodiments of the present application can calculate an initial residual error signal of the suppression signal and the crosstalk signal, and then obtain an initial joint cost value of the to-be-tested mixed signal, and update the parameters of the target joint cost model based on the initial joint cost value, the crosstalk type and the step factor, until the updated target joint cost model meets a certain model condition, so as to obtain a residual error signal meeting a certain signal condition. The certain model condition can be set by a person skilled in the art according to the actual situation, and the present application does not make specific limitations.

[0107] For example, the embodiments of the present application can collect the residual error signal, dynamically jointly optimize the notch filter parameter θ f, the filter coefficient W and the fusion weight α by minimizing the target joint cost model J(n), and then obtain the residual error signal meeting the certain signal condition.

[0108] Further, the embodiments of the present application can re-calculate the index residual value of the to-be-tested mixed signal after eliminating the crosstalk signal, and the schematic diagram thereof is shown in FIG. 4, and the main content is as follows: Figure 4

[0109] Step S401: Obtain the to-be-tested mixed signal after eliminating the crosstalk signal.

[0110] Step S402: Calculate the index residual value.

[0111] Step S403: Determine whether the index residual value is <20%.

[0112] If yes, execute step S405; otherwise, execute step S404.

[0113] Step S404: End the processing.

[0114] Step S405: Re-determine the crosstalk type of the crosstalk signal.

[0115] Step S406: Optimize the algorithm parameters.

[0116] Step S407: Twice determine the index residual value until the processing is ended.

[0117] It can be understood that if the index residual value of the to-be-tested mixed signal after eliminating the crosstalk signal is ≥20%, it is determined that the signal quality meets the standard, and the processing flow is ended; if the index residual value is <20%, the closed-loop feedback mechanism is triggered, the algorithm parameters (such as the notch filter frequency, the reverse pulse amplitude, etc.) are automatically optimized, the crosstalk processing step is executed again, and then the signal re-enters the secondary detection link, and the cycle iteration is performed until the index residual value is ≥20%, so as to improve the reliability of the signal quality evaluation and enhance the model performance and stability.

[0118] ​The embodiment of the application can calculate the initial residual error signal of the suppression signal and the crosstalk signal, calculate the initial joint cost value of the to-be-tested mixed signal, and then update the parameters of the target joint cost model based on the initial joint cost value, the crosstalk type and the step factor until a certain model condition is met, so as to obtain the residual error signal meeting a certain signal condition, enable the model to continuously adjust its parameters according to the actual cancellation condition, gradually optimize the cancellation effect of the suppression signal on the crosstalk signal, improve the accuracy of signal processing, continuously update the parameters of the target joint cost model, improve the reliability of signal quality evaluation, and enhance the model performance and stability.

[0119] The working principle of the signal crosstalk processing method provided by the embodiment of the application will be introduced below in combination with a specific embodiment.

[0120] Among them, Figure 5 The flowchart of the calibration stage of the signal crosstalk processing method provided by an embodiment of the application.

[0121] Calibration stage.

[0122] Step S501: probe signal generator.

[0123] Step S502: calibration control logic.

[0124] Step S503: transmission end switching switch.

[0125] Step S504: interference signal link.

[0126] Step S505: victim signal link.

[0127] Step S506: system identification unit.

[0128] Step S507: calculate initial time-domain impulse response and / or initial frequency-domain transfer function.

[0129] Among them, the crosstalk signal generated by the probe signal generator can be connected to the interference signal link, and the high-precision analog-to-digital converter can synchronously collect signals at the victim signal link end. Since the system has not started normal data transmission at this time, the content on the victim signal link is almost completely the coupling response signal generated by the crosstalk signal through the coupling channel (may also contain a small amount of background noise). Further, the initial time-domain impulse response and / or initial frequency-domain transfer function can be calculated based on the crosstalk signal and the coupling response signal collected from the victim signal link by the least square method or the correlation method, as the system priori knowledge, for greatly improving the convergence speed of the subsequent adaptive process, and initializing the second signal processing model.

[0130] Figure 6A flowchart of a normal working stage of a processing method of signal crosstalk provided according to an embodiment of the present application.

[0131] Normal working stage.

[0132] Step S601: Obtain a crosstalk signal.

[0133] Step S602: Identify a crosstalk type.

[0134] In the embodiment of the present application, the first signal processing model can be used to identify the crosstalk type.

[0135] Step S603: Determine a suppression signal in parallel processing.

[0136] In the embodiment of the present application, the second signal processing model can be used to determine the suppression signal.

[0137] Further, in the embodiment of the present application, the second signal processing model includes a frequency domain notch branch, a time domain adaptive filtering branch, a weighting combiner, and is initialized by using an initial time domain impulse response and / or an initial frequency domain transfer function.

[0138] Step S604: Initialize the initial time domain impulse response and / or the initial frequency domain transfer function.

[0139] Step S605: Inject a signal into a reducer.

[0140] In the embodiment of the present application, the crosstalk signal of the to-be-tested mixed signal in step S607 can be eliminated by using the suppression signal, and then an initial residual error signal is obtained.

[0141] Step S606: Update a target joint cost model parameter.

[0142] In the embodiment of the present application, an initial joint cost value of the to-be-tested mixed signal can be obtained based on the initial residual error signal, and the parameter of the target joint cost model is updated based on the initial joint cost value, the crosstalk type and a step factor until the updated target joint cost model meets a certain model condition.

[0143] Step S607: Link a victim signal.

[0144] In the embodiment of the present application, the to-be-tested mixed signal is output by the victim signal link.

[0145] In addition, it should be noted that in the normal working stage, the sending end switching switch is switched back to the data source.

[0146] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method.

[0147] According to the signal crosstalk processing method proposed in this application, a first signal processing model in a pre-constructed signal processing model can be used to identify the crosstalk type in the mixed signal under test. A second signal processing model can be used to determine the suppression signal of the crosstalk signal. Then, in the target joint cost model, the suppression signal is used to cancel at least a part of the crosstalk signal, and the residual error signal between the suppression signal and the crosstalk signal is determined until a residual error signal that meets certain signal conditions is obtained. Therefore, it can solve the technical problems of limited space when increasing spacing; high cost and complexity of shielding design, and limited application scenarios; slow convergence and poor effect of time-domain filtering for strong periodic crosstalk; and frequency-domain notch filtering that easily causes signal phase distortion and cannot handle random noise. In addition, it cannot intelligently judge the crosstalk type and adaptively adjust the strategy, resulting in poor performance in the face of complex mixed-mode crosstalk. It achieves the technical effect of comprehensively judging signal quality, accurately distinguishing crosstalk type, improving the accuracy of signal problem identification, avoiding misjudgment and omission, minimizing the impact of crosstalk on the signal, improving signal integrity, and being applicable to various complex high-speed signal transmission scenarios. It also reduces the dependence on hardware indicators and achieves an optimized balance between performance and cost.

[0148] Embodiments of this application also provide a signal crosstalk processing apparatus.

[0149] Figure 7 This is a block diagram of a signal crosstalk processing apparatus provided according to an embodiment of this application.

[0150] like Figure 7 As shown, the signal crosstalk processing device 10 includes: a first acquisition module 100, an output module 200, and a first generation module 300.

[0151] The first acquisition module 100 is used to acquire the mixed signal to be tested in the equalization system.

[0152] The output module 200 is used to input the mixed signal to be tested into a pre-built signal processing model, so as to identify the crosstalk type of at least one crosstalk signal in the mixed signal to be tested using a first signal processing model in the pre-built signal processing model, and to determine the suppression signal of at least one crosstalk signal using a second signal processing model in the pre-built signal processing model.

[0153] The first generation module 300 is configured to input the suppression signal, the crosstalk type and the to-be-tested mixed signal into a target joint cost model, to cancel at least part of the crosstalk signal by using the suppression signal, and to determine a residual error signal between the suppression signal and the crosstalk signal based on the at least part, until a residual error signal satisfying a preset signal condition is obtained.

[0154] Optionally, in an embodiment of the present application, the method further comprises a second acquisition module, a calculation module, a first judgment module, a first determination module and a first construction module.

[0155] The second acquisition module is configured to acquire a plurality of reference crosstalk signals before inputting the to-be-tested mixed signal into the pre-constructed signal processing model.

[0156] The calculation module is configured to perform fast Fourier transform on the reference crosstalk signal to obtain a power spectral density corresponding to the reference crosstalk signal, and to calculate a difference value between an amplitude of the power spectral density and a background noise discrete spectrum peak.

[0157] The first judgment module is configured to judge whether the difference value is greater than a first threshold value.

[0158] The first determination module is configured to determine that the reference crosstalk signal is a periodic crosstalk signal when the difference value is greater than the first threshold value, and otherwise, to determine that the reference crosstalk signal is an aperiodic crosstalk signal.

[0159] The first construction module is configured to construct a first signal processing model in the signal processing model based on the periodic crosstalk signal and the aperiodic crosstalk signal, to identify a crosstalk type of the reference crosstalk signal.

[0160] Optionally, in an embodiment of the present application, the method further comprises a third acquisition module, an extraction module, a second generation module and a second construction module.

[0161] The third acquisition module is configured to acquire a plurality of probe crosstalk signals before inputting the to-be-tested mixed signal into the pre-constructed signal processing model.

[0162] The extraction module is configured to extract a time domain feature and a frequency domain feature of the probe crosstalk signal based on a crosstalk type of the probe crosstalk signal.

[0163] The second generation module is configured to perform time domain processing and frequency domain processing on the time domain feature and the frequency domain feature at the same time, to obtain a first suppression signal of the probe crosstalk signal in the time domain and a second suppression signal of the probe crosstalk signal in the frequency domain.

[0164] The second construction module is configured to calculate a suppression signal of the probe crosstalk signal based on the first suppression signal and the second suppression signal, to construct a second signal processing model in the signal processing model.

[0165] Optionally, in an embodiment of the present application, the second construction module comprises a determination unit, a first calculation unit and a first generation unit.

[0166] The determination unit is configured to determine a first weight coefficient corresponding to the first suppression signal and a second weight coefficient corresponding to the second suppression signal based on the first suppression signal and the second suppression signal.

[0167] The first calculation unit is configured to calculate a first product between the first suppression signal and the first weight coefficient and a second product between the second suppression signal and the second weight coefficient, respectively.

[0168] The first generation unit is configured to sum the first product and the second product to obtain the suppression signal.

[0169] Optionally, in an embodiment of the present application, the output module 200 comprises an acquisition unit, a second calculation unit and a second generation unit.

[0170] The acquisition unit is configured to acquire a coupling response signal generated by the crosstalk signal under no-load of the to-be-tested mixed signal.

[0171] The second calculation unit is configured to calculate an initial time-domain impulse response and / or an initial frequency-domain transfer function of the crosstalk signal based on the crosstalk signal and the coupling response signal.

[0172] The second generation unit is configured to initialize the second signal processing model by using the initial time-domain impulse response and / or the initial frequency-domain transfer function to obtain an initialized second signal processing model, and acquire the suppression signal by using the initialized second signal processing model.

[0173] Optionally, in an embodiment of the present application, the first generation module 300 comprises a third calculation unit, a third generation unit and an updating unit.

[0174] The third calculation unit is configured to calculate an initial residual error signal of the suppression signal and the crosstalk signal.

[0175] The third generation unit is configured to obtain an initial joint cost value of the to-be-tested mixed signal based on the initial residual error signal.

[0176] The updating unit is configured to update a parameter of a target joint cost model based on the initial joint cost value, the crosstalk type and a step factor until the updated target joint cost model satisfies a preset model condition, so as to acquire a residual error signal satisfying a preset signal condition by using the updated target joint cost model.

[0177] Optionally, in an embodiment of the present application, an expression of the target joint cost model can be but is not limited to:

[0178] ,

[0179] wherein, is a residual error signal; is a coefficient vector of a time-domain filter; is a parameter vector of a frequency-domain notch filter; is a fusion weight vector of a weighted combiner; , , is a regularization factor, represents a variation of a parameter corresponding to adjacent periods.

[0180] Optionally, in an embodiment of the present application, further comprising: a fourth acquisition module, a determination module, a second judgment module, a second determination module and a third determination module.

[0181] The fourth acquisition module is configured to determine a measured value and a target value corresponding to at least one signal parameter based on the at least one signal parameter of the to-be-tested mixed signal before the to-be-tested mixed signal is input to the pre-constructed signal processing model.

[0182] The determination module is configured to calculate a difference value between the measured value and the target value, and calculate a ratio of the difference value and the target value based on the difference value, so as to determine an index margin value corresponding to the at least one signal parameter according to the ratio.

[0183] The second judgment module is configured to judge whether the index margin value is less than a preset threshold value.

[0184] The second determination module is configured to, when the index margin value is less than the preset threshold value, acquire a signal attenuation value of the to-be-tested mixed signal, and detect whether the signal attenuation value is less than a preset attenuation value, and when it is detected that the signal attenuation value is less than the preset attenuation value, determine that there is at least one crosstalk signal in the to-be-tested mixed signal, and allow the to-be-tested mixed signal to be input to the pre-constructed signal processing model.

[0185] The third determination module is configured to, when the index margin value is greater than or equal to the preset threshold value, determine that there is no at least one crosstalk signal in the to-be-tested mixed signal, and prohibit the to-be-tested mixed signal from being input to the pre-constructed signal processing model.

[0186] The features of the embodiments of the signal crosstalk processing device can be referred to the related descriptions of the embodiments of the signal crosstalk processing method, which will not be repeated here.

[0187] The signal crosstalk processing device provided in the embodiment of the present application can identify the crosstalk type of the crosstalk signal in the mixed signal to be measured by using the first signal processing model in the pre-constructed signal processing model, determine the suppression signal of the crosstalk signal by using the second signal processing model, further offset at least part of the crosstalk signal by using the suppression signal in the target joint cost model, and determine the residual error signal between the suppression signal and the crosstalk signal until the residual error signal meeting certain signal conditions is obtained. Therefore, the technical problem that the use space is limited by increasing the spacing, the design cost and complexity of the shielding are high, the use scene is limited, the time domain filtering converges slowly and has poor effect on strong periodic crosstalk, the frequency domain wave trap easily causes signal phase distortion and cannot process random noise, and the like can be solved. In addition, none of them can intelligently judge the crosstalk type and adaptively adjust the strategy, resulting in poor performance in front of complex mixed mode crosstalk. The technical effects of comprehensively judging the signal quality, accurately distinguishing the crosstalk type, improving the accuracy of signal problem identification, avoiding misjudgment and omission, minimizing the influence of crosstalk on the signal, improving the signal integrity, being applicable to various complex high-speed signal transmission scenes, reducing the dependence on hardware indicators, and realizing the optimization balance between performance and cost are achieved.

[0188] The embodiment of the present application further provides an electronic device, including a memory and a processor, the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in the embodiments of the signal crosstalk processing method described above.

[0189] The embodiment of the present application further provides a computer readable storage medium, which stores a computer program, wherein the computer program is configured to execute the steps in the embodiments of the signal crosstalk processing method described above when running.

[0190] In an example embodiment, the computer readable storage medium described above can include, but is not limited to, a U disk, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store computer programs.

[0191] The embodiment of the present application further provides a computer program product, which includes a computer program, and the computer program is executed by a processor to implement the steps in the embodiments of the signal crosstalk processing method described above.

[0192] The embodiment of the present application further provides another computer program product, which includes a non-volatile computer readable storage medium, and the non-volatile computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps in the embodiments of the signal crosstalk processing method described above.

[0193] Those skilled in the art will further appreciate that the units and algorithms described in connection with the examples disclosed herein can be implemented in electronic hardware, computer software, or both. As described above, the disclosure is directed to each individual feature, algorithm, and combination of features and algorithms described. Those skilled in the art will recognize that each individual feature or each combination of features can be implemented in hardware, software, or both. The disclosure is directed to each individual feature, algorithm, and combination of features and algorithms regardless of whether the features, algorithms, and combinations of features and algorithms are implemented in hardware, software, or both. The disclosure is directed to each individual feature, algorithm, and combination of features and algorithms regardless of whether the features, algorithms, and combinations of features and algorithms can be best implemented in hardware, software, or both.

[0194] The above provides a signal crosstalk processing method. The principles and implementation manners of the present application are described by using specific examples. The above description of the examples is only used to help understand the method and the core idea of the present application. It should be pointed out that, for those skilled in the art, some improvements and modifications can be made to the present application without departing from the principles of the present application. These improvements and modifications also fall within the protection scope of the claims of the present application.

Claims

1. A method of processing signal crosstalk, characterized by, The method comprises the following steps: acquiring a to-be-tested mixed signal in an equalization system; inputting the to-be-tested mixed signal into a pre-constructed signal processing model, to identify a crosstalk type of at least one crosstalk signal in the to-be-tested mixed signal by using a first signal processing model in the pre-constructed signal processing model, and to determine a suppression signal of the at least one crosstalk signal by using a second signal processing model in the pre-constructed signal processing model; inputting the suppression signal, the crosstalk type and the to-be-tested mixed signal into a target joint cost model, to offset at least part of the crosstalk signal by using the suppression signal, and to determine a residual error signal between the suppression signal and the crosstalk signal based on the at least part, until a residual error signal meeting a preset signal condition is obtained; wherein the inputting the suppression signal, the crosstalk type and the to-be-tested mixed signal into the target joint cost model, to offset at least part of the crosstalk signal by using the suppression signal, and to determine a residual error signal between the suppression signal and the crosstalk signal based on the at least part, comprises: calculating an initial residual error signal of the suppression signal and the crosstalk signal; obtaining an initial joint cost value of the to-be-tested mixed signal based on the initial residual error signal; updating parameters of the target joint cost model based on the initial joint cost value, the crosstalk type and a step factor, until the updated target joint cost model meets a preset model condition, to obtain the residual error signal meeting the preset signal condition by using the updated target joint cost model; an expression of the target joint cost model is: ， wherein is a residual error signal; is a coefficient vector of the time-domain filter; is a parameter vector of the frequency-domain notch filter; is a fusion weight vector of the weighted combiner; , , is a regularization factor, represents a variation of the corresponding parameter of the adjacent period.

2. The method of claim 1, wherein, before inputting the to-be-tested mixed signal into the pre-constructed signal processing model, further comprising: acquiring a plurality of reference crosstalk signals; performing fast Fourier transform on the reference crosstalk signals to obtain power spectral densities corresponding to the reference crosstalk signals, and calculating a difference between an amplitude of the power spectral density and a background noise discrete spectrum peak; determining whether the difference is greater than a first threshold value; if the difference is greater than the first threshold value, determining that the reference crosstalk signal is a periodic crosstalk signal, otherwise, determining that the reference crosstalk signal is an aperiodic crosstalk signal; based on the periodic crosstalk signal and the aperiodic crosstalk signal, constructing a first signal processing model in the signal processing model to identify the crosstalk type of the reference crosstalk signal.

3. The method of claim 1, wherein, before inputting the to-be-tested mixed signal into the pre-constructed signal processing model, further comprising: acquiring a plurality of probe crosstalk signals; based on a crosstalk type of the probe crosstalk signal, extracting a time domain feature and a frequency domain feature of the probe crosstalk signal; performing time domain processing and frequency domain processing on the time domain feature and the frequency domain feature at the same time, to obtain a first suppression signal of the probe crosstalk signal in the time domain and a second suppression signal of the probe crosstalk signal in the frequency domain; based on the first suppression signal and the second suppression signal, calculating a suppression signal of the probe crosstalk signal, to construct a second signal processing model in the signal processing model.

4. The method of claim 3, wherein, The calculating the suppression signal of the probe crosstalk signal based on the first suppression signal and the second suppression signal comprises: determining a first weight coefficient corresponding to the first suppression signal and a second weight coefficient corresponding to the second suppression signal based on the first suppression signal and the second suppression signal; calculating a first product between the first suppression signal and the first weight coefficient and a second product between the second suppression signal and the second weight coefficient respectively; summing the first product and the second product to obtain the suppression signal.

5. The method of claim 1, wherein, The determining the suppression signal of the at least one crosstalk signal by using the second signal processing model in the pre-constructed signal processing model comprises: obtaining a coupling response signal generated by the crosstalk signal under the empty load of the to-be-tested mixed signal; calculating an initial time-domain impulse response and / or an initial frequency-domain transfer function of the crosstalk signal based on the crosstalk signal and the coupling response signal; initializing the second signal processing model by using the initial time-domain impulse response and / or the initial frequency-domain transfer function to obtain an initialized second signal processing model, and obtaining the suppression signal by using the initialized second signal processing model.

6. The method of claim 1, wherein, Before the to-be-tested mixed signal is input into the pre-constructed signal processing model, the method further comprises: determining a measured value and a target value corresponding to the at least one signal parameter based on the at least one signal parameter of the to-be-tested mixed signal; calculating a difference value between the measured value and the target value, and calculating a ratio of the difference value to the target value based on the difference value, so as to determine an index margin value corresponding to the at least one signal parameter according to the ratio; determining whether the index margin value is less than a preset threshold value; if the index margin value is less than the preset threshold value, obtaining a signal attenuation value of the to-be-tested mixed signal, and detecting whether the signal attenuation value is less than a preset attenuation value, when it is detected that the signal attenuation value is less than the preset attenuation value, it is determined that the at least one crosstalk signal exists in the to-be-tested mixed signal, and the to-be-tested mixed signal is allowed to be input into the pre-constructed signal processing model; if the index margin value is greater than or equal to the preset threshold value, it is determined that the at least one crosstalk signal does not exist in the to-be-tested mixed signal, and the to-be-tested mixed signal is prohibited from being input into the pre-constructed signal processing model.

7. An electronic device, comprising: comprise: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the processing method of the signal crosstalk according to any one of claims 1-6.

8. A computer readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the processing method of the signal crosstalk according to any one of claims 1-6.

Images