A chip test signal analysis processing system and method based on big data

Abstract

The application discloses a chip test signal analysis processing system and method based on big data, relates to the chip test signal analysis technical field, and collects chip test signals, obtains target test signals through signal preprocessing, and sets a period window to obtain a target period signal; corresponding period time point signal channel decomposition processing is carried out based on the target period signal, channel characteristic data of corresponding time point signal data are obtained; a characteristic correlation group of the corresponding time point signal data is constructed in combination with the channel characteristic data, characteristic state data of signals at corresponding time points are analyzed based on the characteristic correlation groups corresponding to the signal data at the time points in a period, and continuous time point signal state running steady state analysis is carried out based on the signal characteristic state data at the time points, so that the chip signal state condition is determined; the application effectively analyzes the real-time running state characteristics of the chip and determines the running steady state condition of the chip signal, and accurate measurement and analysis are carried out.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of chip test signal analysis, and particularly relates to a chip test signal analysis processing system and method based on big data. BACKGROUND

[0002] At present, the semiconductor industry is thriving, especially in the environment of rapid technological development, both hardware and software demands are gradually expanding, so the current environment has extremely large demand for semiconductor chips, which requires high quality and strict standards for semiconductor chips;

[0003] Traditional chip signal testing mostly relies on oscilloscopes and other devices to observe signal waveforms, measure signal amplitude and period, and compare signal data through small-scale data statistical analysis to establish performance standards; However, it cannot handle complex signal characteristics, and it is difficult to extract deep feature information in the face of a large number of digital signals and mixed signals in the current chip signals, and the data size it relies on is too small to fully reflect the real-time performance status of the chip, so the traditional chip signal testing method cannot effectively analyze the real-time running state characteristics of the chip and determine the running steady state of the chip signal, so it cannot be accurately measured and analyzed. SUMMARY

[0004] The purpose of the present application is to provide a chip test signal analysis processing system and method based on big data to solve the problems in the prior art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical scheme:

[0006] A chip test signal analysis processing method based on big data, the method comprising the following steps:

[0007] Collecting chip test signals, obtaining target test signals through signal preprocessing, and setting a cycle window to retrieve target cycle signals from the target test signals; The signal preprocessing includes signal frequency band positioning processing and signal noise reduction processing;

[0008] Based on the target cycle signal, the corresponding cycle time point signal channel decomposition processing is performed to obtain the channel feature data of the corresponding time point signal data; The channel feature data includes channel feature components and instantaneous feature frequencies of the corresponding channel; The feature correlation group of the corresponding time point signal data is constructed in combination with the channel feature data, the feature state data of the corresponding time point signal is analyzed based on the feature correlation group corresponding to the signal data at each time point in the cycle, and the running steady state of the continuous time point signal in the cycle is analyzed based on the feature state data of the signal at each time point, and the chip signal state is determined based on the analysis data.

[0009] The signal frequency band positioning processing is performed by using a signal acquisition device to sample and debug a signal, adjusting a sampling frequency according to a highest frequency of a chip test signal, and obtaining initial signal data of a target test signal in a target frequency interval;

[0010] The signal noise reduction processing is performed by converting the initial signal data of the target test signal collected in time domain to frequency domain, removing interference signal data in the target test signal through a filtering algorithm, and outputting target test signal data after noise reduction.

[0011] The target test signal data after preprocessing is set with a period window for period signal interception, and the intercepted period signal data is labeled to obtain target period signal data;

[0012] The target period signal data corresponding to the label is called, and signal data at each time point in the period is extracted; based on the signal data at the corresponding time point, signal channel decomposition analysis is performed, wavelet transform decomposition is used to determine the decomposition layer number, channel feature component analysis and extraction are performed on the signal data at the corresponding time point, and a channel feature component set of the signal data at the corresponding time point is obtained; based on the channel feature component data at the corresponding time point, energy distribution data of the corresponding channel feature component is analyzed, and an energy entropy data set in the corresponding channel component is obtained; wherein, the wavelet transform decomposition analysis calculation of the signal data is

[0013]

[0014] Wherein, X(i, t) is the signal data corresponding to the time point i in the period with the corresponding number t; c j,n (i, t) is the energy distribution of the signal data corresponding to the time point i in the period with the corresponding number t in the channel feature component n under j-layer decomposition; is the wavelet basis function of the signal data corresponding to the time point i in the period with the corresponding number t in the channel feature component n under j-layer decomposition; J is a constant;

[0015] Based on the energy distribution of the channel feature component in the signal data at each time point, the corresponding energy entropy in each channel component is analyzed, and the calculation formula is

[0016]

[0017] Wherein, E(j, n, i) is the energy entropy data of the channel feature component n under j-layer decomposition in the signal data at the corresponding time point i; p j,n,i is the energy proportion of the channel feature component n under j-layer decomposition in the signal data at the corresponding time point i; wherein, the calculation formula of the energy proportion p j,n,i of the corresponding channel feature component is

[0018]

[0019] Based on the empirical mode decomposition of each channel characteristic component of the corresponding time point signal, the instantaneous characteristic frequency in each channel characteristic component is determined; the characteristic correlation group of the corresponding channel characteristic component is constructed in combination with the energy entropy data and the instantaneous characteristic frequency data corresponding to each channel characteristic component of the corresponding time point signal; wherein the analysis based on the empirical mode decomposition of the channel characteristic component is

[0020]

[0021] Wherein, imfn(i, t) is the IMF component of the signal data channel characteristic component n corresponding to the time point i in the corresponding number t period; r(i, t) is the remaining residual;

[0022] The instantaneous characteristic frequency analysis is performed on the IMF component in the corresponding channel characteristic component obtained by the empirical mode decomposition, and the calculation is

[0023]

[0024] Wherein, F(n, i, t) is the instantaneous characteristic frequency of the signal data channel characteristic component n corresponding to the time point i in the corresponding number t period; H(·) is the Hilbert transform.

[0025] Based on the characteristic correlation group of the signal data corresponding channel characteristic component at each time point of the target period signal, the characteristic correlation group of the signal data at each time point is analyzed respectively, and the representation correlation group corresponding to the signal data at each time point is determined; the analysis is

[0026] K(n,i,t)=E(j,n,i)*F(n,i,t);

[0027] Wherein, K(n, i, t) is the characteristic correlation evaluation value of the signal data channel characteristic component n corresponding to the time point i in the number t period; based on the characteristic correlation evaluation value of the signal data corresponding channel characteristic component at each time point of the corresponding period signal, the maximum value of the characteristic correlation group of the corresponding channel characteristic component is taken as the representation correlation group of the signal data at the corresponding time point;

[0028] Based on the characteristic vectorization of the representation correlation group of the signal data at each time point in the corresponding period, the representation vector group of the signal data at each time point in the period is determined, and the characteristic state data of the signal data at each time point in the period is analyzed to obtain the characteristic state evaluation value of the signal data at each time point in the corresponding period; the analysis is as follows

[0029] Z(i,t)=K(n,i,t) z *cosα(n,i,t) z ;

[0030] Wherein, Z(i, t) is the characteristic state evaluation value of the signal data corresponding to the time point i in the period t; K(n, i, t)z is the characteristic correlation evaluation value of the characteristic correlation group of the signal data channel characteristic component n corresponding to the time point i in the period t; a(n, i, t)z is the horizontal angle of the signal data curve corresponding to the time point i in the period t; the angle is taken as the angle between the line connecting the signal curve point corresponding to the current time point and the signal curve point corresponding to the adjacent next time point and the horizontal line;

[0031] The state stability of the target period signal data is analyzed in combination with the characteristic state evaluation value of the signal data at each time point in the period; the change difference value of the signal curve data of the adjacent time points is analyzed by analyzing the characteristic state evaluation value of the signal data of the continuous time points on the target period signal curve; the running state value of the target period data is analyzed in combination with the change difference value analysis data of the signal curve data of the continuous time points on the target period signal curve; based on the running state value analysis data of the target period signal data, the stable threshold is determined, and the target period signal of the current chip is judged abnormally; wherein, the state value analysis of the target period signal is

[0032]

[0033] Wherein, W is the running state value of the target period signal; Z(i+1, t) is the characteristic state evaluation value of the signal data of the adjacent next time point of the time point i on the target period signal curve; Q(i, i+1) is the curvature of the adjacent time point signal curve on the target period curve; Q(i, i+1)max and Q(i, i+1)min are the maximum and minimum values of the curvature of the adjacent time point signal curve on the target period curve; Z(i, t)max and Z(i, t)min are the maximum and minimum values of the characteristic state evaluation value of the signal data on the target period signal curve;

[0034] By introducing the determination of the stable threshold, if the running state value of the target period signal is less than or equal to the threshold, it is judged that the running state of the current chip is good; otherwise, it is judged that the running state of the current chip is abnormal.

[0035] The corresponding signal curve data and the corresponding running state value analysis data of each target period data of the chip test signal are output respectively;

[0036] The running state judgment results of each target period chip are output respectively, and the abnormal state chip and the abnormal target period signal data are marked.

[0037] A chip test signal analysis processing system based on big data, the system comprises a target signal acquisition module, a signal feature analysis module, a signal cycle steady analysis module and a chip data feedback module;

[0038] The target signal acquisition module acquires chip test signals, obtains target test signals through signal preprocessing, and sets a cycle window to retrieve target cycle signals from the target test signals; the signal feature analysis module performs signal channel decomposition processing on the corresponding cycle time point signals based on the target cycle signals, and obtains channel feature data of the corresponding time point signal data; the channel feature data includes channel feature components and instantaneous feature frequencies of the corresponding channels; the corresponding time point signal data is constructed into a feature correlation group in combination with the channel feature data; the signal cycle steady analysis module analyzes feature state data of the signals at the corresponding time points based on the feature correlation groups corresponding to the signal data at the time points, and performs steady state analysis of the signals at the continuous time points in a cycle based on the feature state data of the signals at the time points, determines the chip signal state based on the analysis data; the chip data feedback module outputs the cycle signal steady state value analysis data of the chip test signals and marks abnormal cycle signal data and abnormal state chips.

[0039] The target signal acquisition module comprises a test signal acquisition unit and a target signal preprocessing unit;

[0040] The test signal acquisition unit is used to acquire test signal data generated by the operation of the chip;

[0041] The target signal preprocessing unit comprises signal frequency band positioning processing and signal noise reduction processing;

[0042] The signal frequency band positioning processing is to adjust the sampling frequency according to the highest frequency of the chip test signal by using signal sampling debugging of the signal acquisition device, and to obtain initial signal data of the target test signal in the target frequency interval;

[0043] The signal noise reduction processing is to convert the initial signal data of the collected target test signal through time domain and frequency domain conversion, remove the interference signal data in the target test signal through a filtering algorithm, and output the target test signal data after noise reduction.

[0044] The signal feature analysis module comprises a signal channel component analysis unit and a signal channel feature correlation group construction unit;

[0045] The signal channel component analysis unit sets a cycle window to intercept cycle signals from the target test signal data after preprocessing, and performs label processing on the intercepted cycle signal data to obtain target cycle signal data;

[0046] The target period signal data corresponding to the label is called, and signal data at each time point in the period is extracted respectively; based on the signal data at the corresponding time point, signal channel decomposition analysis is performed, the decomposition layer is determined by wavelet transform decomposition, the channel characteristic component analysis and extraction of the signal data at the corresponding time point are performed, and the channel characteristic component set of the signal data at the corresponding time point is obtained; based on the channel characteristic component data at the corresponding time point, the energy distribution data of the corresponding channel characteristic component is analyzed, and the energy entropy data set in the corresponding channel component is obtained;

[0047] The signal channel feature association group construction unit determines the instantaneous feature frequency in each channel feature component based on the empirical mode decomposition of the corresponding time point signal of each channel feature component; the feature association group of the corresponding channel feature component is constructed by combining the energy entropy data and the instantaneous feature frequency data of each channel feature component at the corresponding time point signal.

[0048] The signal period steady-state analysis module includes a signal feature state evaluation unit and a signal curve steady-state analysis unit.

[0049] The signal feature state evaluation unit analyzes the feature association group of the signal data at each time point based on the feature association group of the corresponding channel feature component of the signal data at each time point of the target period signal, determines the representation association group of the signal data at each time point, and analyzes the feature state data of the signal data at each time point in the period to obtain the feature state evaluation value of the signal data at each time point in the corresponding period.

[0050] Based on the representation association group of the signal data at each time point in the corresponding period, the representation vector group of the signal data at each time point in the period is determined, and the feature state data of the signal data at each time point in the period is analyzed to obtain the feature state evaluation value of the signal data at each time point in the corresponding period.

[0051] The signal curve steady-state analysis unit analyzes the state steady-state of the target period signal data by combining the feature state evaluation value of the signal data at each time point in the period; the change difference value analysis of the signal curve data of adjacent time points is performed by analyzing the feature state evaluation value of the signal data of continuous time points on the target period signal curve; the period signal running steady-state value of the target period data is analyzed by combining the change difference value analysis data of the signal curve data of continuous time points on the target period signal curve; based on the period signal running steady-state value analysis data of the target period signal data, a stable threshold is determined, and the target period signal of the current chip is abnormally judged.

[0052] The chip data feedback module includes a data feedback unit and an abnormality marking unit.

[0053] The data feedback unit outputs the corresponding signal curve data and the corresponding period signal running steady-state value analysis data of each target period data divided by the chip test signal respectively.

[0054] The abnormality marking unit respectively outputs based on the running state judgment result of each target period chip, and marks the abnormal state chip and the abnormal target period signal data.

[0055] Compared with the prior art, the present application has the following beneficial effects:

[0056] The present application acquires target period signal by collecting chip test signal and using target frequency band positioning and noise reduction processing; on this basis, the energy distribution and instantaneous characteristic frequency characteristics in each channel characteristic component are determined by wavelet decomposition of signal data in the target period, so as to establish signal data characteristic association group corresponding to the time point; the characteristic association group is determined by determining the signal data at each time point, so as to analyze the characteristic state value of the signal data at the corresponding time point, and based on this, the running steady state analysis of the target period signal curve is carried out to determine the running state of the chip in the corresponding period; the present application effectively analyzes the real-time running state characteristics of the chip and determines the running steady state of the chip signal, and accurately measures and analyzes. BRIEF DESCRIPTION OF DRAWINGS

[0057] Figure 1 It is a structural schematic diagram of a chip test signal analysis processing system based on big data according to the present application;

[0058] Figure 2 It is a flowchart of a chip test signal analysis processing method based on big data according to the present application. DETAILED DESCRIPTION

[0059] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0060] Embodiment: as shown in the figure, the present application provides a technical solution: Figure 1

[0061] A chip test signal analysis processing system based on big data, the system comprises a target signal acquisition module, a signal characteristic analysis module, a signal period steady state analysis module and a chip data feedback module;

[0062] ​The target signal acquisition module acquires chip test signals, obtains target test signals through signal preprocessing, and sets a period window to retrieve target period signals from the target test signals; the signal feature analysis module performs corresponding period time point signal channel decomposition processing based on the target period signals, and obtains channel feature data of the corresponding time point signal data; the channel feature data includes channel feature components and instantaneous feature frequencies of the corresponding channels; a feature correlation group of the corresponding time point signal data is constructed in combination with the channel feature data; the signal period steady state analysis module analyzes feature state data of signals at corresponding time points based on the feature correlation groups corresponding to the signal data at the time points, and performs continuous time point signal state running steady state analysis based on the feature state data of the signals at the time points, determines the chip signal state based on the analysis data; and the chip data feedback module outputs period signal running steady state value analysis data of the chip test signals and marks abnormal period signal data and abnormal state chips.

[0063] The target signal acquisition module includes a test signal acquisition unit and a target signal preprocessing unit.

[0064] The test signal acquisition unit is configured to acquire test signal data generated by chip running.

[0065] The target signal preprocessing unit includes signal frequency band positioning processing and signal noise reduction processing.

[0066] The signal frequency band positioning processing is performed by using a signal acquisition device to perform signal sampling debugging, adjusting the sampling frequency according to the highest frequency of the chip test signals, and obtaining initial signal data of the target test signals in a target frequency interval.

[0067] The signal noise reduction processing is performed by converting the acquired initial signal data of the target test signals through time domain and frequency domain conversion, removing interference signal data in the target test signals through a filtering algorithm, and outputting the target test signal data after noise reduction.

[0068] The signal feature analysis module includes a signal channel component analysis unit and a signal channel feature correlation group construction unit.

[0069] The signal channel component analysis unit sets a period window to intercept period signals from the preprocessed target test signal data, and performs label processing on the intercepted period signal data to obtain target period signal data.

[0070] The target period signal data corresponding to the label is called, and signal data at each time point in the period is extracted respectively; based on the signal data at the corresponding time point, signal channel decomposition analysis is performed, the decomposition layer is determined by wavelet transform decomposition, the channel characteristic component analysis and extraction of the signal data at the corresponding time point are performed, and the channel characteristic component set of the signal data at the corresponding time point is obtained; based on the channel characteristic component data at the corresponding time point, the energy distribution data of the corresponding channel characteristic component is analyzed, and the energy entropy data set in the corresponding channel component is obtained;

[0071] The signal channel feature correlation group construction unit determines the instantaneous feature frequency in each channel characteristic component based on the empirical mode decomposition of each channel characteristic component of the corresponding time point signal; the feature correlation group of the corresponding channel characteristic component is constructed by combining the energy entropy data and the instantaneous feature frequency data corresponding to each channel characteristic component at the corresponding time point signal.

[0072] The signal period steady-state analysis module includes a signal feature state evaluation unit and a signal curve steady-state analysis unit.

[0073] The signal feature state evaluation unit analyzes the feature correlation group of the signal data at each time point based on the feature correlation group of the corresponding channel characteristic component of the signal data at each time point of the target period signal, determines the representation correlation group of the signal data at each time point, and analyzes the feature state data of the signal data at each time point in the period to obtain the feature state evaluation value of the signal data at each time point in the corresponding period.

[0074] Based on the representation correlation group of the signal data at each time point in the corresponding period, the representation vector group of the signal data at each time point in the period is determined, and the feature state data of the signal data at each time point in the period is analyzed to obtain the feature state evaluation value of the signal data at each time point in the corresponding period.

[0075] The signal curve steady-state analysis unit analyzes the state steady-state of the target period signal data by combining the feature state evaluation value of the signal data at each time point in the period; the change difference value analysis of the signal curve data of adjacent time points is performed by analyzing the feature state evaluation value of the signal data of continuous time points on the target period signal curve; the period signal running steady-state value of the target period data is analyzed by combining the change difference value analysis data of the signal curve data of continuous time points on the target period signal curve; based on the period signal running steady-state value analysis data of the target period signal data, a stable threshold is determined, and the target period signal of the current chip is abnormally judged.

[0076] The chip data feedback module includes a data feedback unit and an abnormality marking unit.

[0077] The data feedback unit outputs the corresponding signal curve data and the corresponding period signal running steady-state value analysis data of each target period data divided by the chip test signal respectively.

[0078] The abnormality marking unit outputs based on the running state judgment result of each target period chip, and marks the abnormal state chip and the abnormal target period signal data;

[0079] As shown in Figure 2 The present application provides another technical solution:

[0080] A chip test signal analysis processing method based on big data, the method comprising the following steps:

[0081] Collecting chip test signals, obtaining target test signals through signal preprocessing, and setting a period window to retrieve target period signals from the target test signals; the signal preprocessing includes signal frequency band positioning processing and signal noise reduction processing;

[0082] Based on the target period signal, corresponding period time point signal channel decomposition processing is performed to obtain channel feature data of the corresponding time point signal data; the channel feature data includes channel feature components and instantaneous feature frequencies of the corresponding channel; a feature correlation group of the corresponding time point signal data is constructed based on the channel feature data, the feature state data of the signal at the corresponding time point is analyzed based on the feature correlation group corresponding to the signal data at each time point in the period, and the signal state running steady state analysis of the continuous time points in the period is performed based on the signal feature state data at each time point, and the chip signal state is determined based on the analysis data.

[0083] The signal frequency band positioning processing is to adjust the sampling frequency according to the highest frequency of the chip test signal by using the signal acquisition device for signal sampling debugging, and to obtain the initial signal data of the target test signal in the target frequency interval;

[0084] The signal noise reduction processing is to convert the initial signal data of the collected target test signal through time domain frequency domain conversion, remove the interference signal data in the target test signal through a filtering algorithm, and output the denoised target test signal data.

[0085] The target test signal data after preprocessing is set with a period window to intercept the period signal, and the intercepted period signal data is labeled to obtain the target period signal data;

[0086] The target periodic signal data corresponding to the label is called, and the signal data at each time point in the period is extracted respectively; based on the signal data at the corresponding time point, signal channel decomposition analysis is carried out, the decomposition layer is determined by wavelet transform decomposition, the channel characteristic component analysis and extraction of the signal data at the corresponding time point are carried out, and the channel characteristic component set of the signal data at the corresponding time point is obtained; based on the channel characteristic component data at the corresponding time point, the energy distribution data of the corresponding channel characteristic component is analyzed, and the energy entropy data set in the corresponding channel component is obtained; wherein the wavelet transform decomposition analysis calculation of signal data is

[0087]

[0088] Wherein, X(i, t) is the signal data corresponding to the time point i in the period corresponding to the number t; c j,n (i, t) is the energy distribution of the signal data corresponding to the time point i in the period corresponding to the number t in the channel characteristic component n under j layer decomposition; is the wavelet basis function of the channel characteristic component n of the signal data corresponding to the time point i in the period corresponding to the number t under j layer decomposition; J is a constant;

[0089] Based on the energy distribution of the channel characteristic component in the signal data at each time point, the corresponding energy entropy in each channel component is analyzed, and the calculation formula is

[0090]

[0091] Wherein, E(j, n, i) is the energy entropy data of the channel characteristic component n under j layer decomposition in the signal data at the time point i; p j,n,i is the energy proportion of the channel characteristic component n under j layer decomposition in the signal data at the time point i; wherein the calculation formula of the energy proportion p j,n,i of the corresponding channel characteristic component is

[0092]

[0093] Based on the empirical mode decomposition of the signal channel characteristic component at the corresponding time point, the instantaneous characteristic frequency in each channel characteristic component is determined; the feature association group of the corresponding channel characteristic component is constructed by combining the energy entropy data and the instantaneous characteristic frequency data of each channel characteristic component at the corresponding time point; wherein the analysis of empirical mode decomposition based on channel characteristic component is

[0094]

[0095] Wherein, imfn(i, t) is the IMF component of the channel characteristic component n of the signal data corresponding to the time point i in the period corresponding to the number t; r(i, t) is the remaining residual;

[0096] The instantaneous characteristic frequency analysis is performed on the imf component in the corresponding channel characteristic component obtained by the empirical mode decomposition, and the calculation is

[0097]

[0098] Wherein, F(n, i, t) is the instantaneous characteristic frequency of the signal data channel characteristic component n corresponding to the time point i in the period numbered t; H(·) is the Hilbert transform.

[0099] Based on the characteristic correlation group of the signal data corresponding channel characteristic component at each time point of the target period signal, the characteristic correlation group of the signal data at each time point is analyzed respectively, and the characteristic correlation group corresponding to the signal data at each time point is determined; The analysis is

[0100] K(n, i, t) = E(j, n, i) * F(n, i, t);

[0101] Wherein, K(n, i, t) is the characteristic correlation evaluation value of the signal data channel characteristic component n corresponding to the time point i in the period numbered t; Based on the characteristic correlation evaluation value of the signal data corresponding channel characteristic component at each time point of the corresponding period signal, the maximum value of the characteristic correlation group of the corresponding channel characteristic component is taken as the characteristic correlation group of the signal data at the corresponding time point;

[0102] Based on the characteristic correlation group of the signal data at each time point in the corresponding period, the characteristic vectorization is performed, the characteristic vector group of the signal data at each time point in the period is determined, and the characteristic state data of the signal data at each time point in the period is analyzed, to obtain the characteristic state evaluation value of the signal data at each time point in the corresponding period; The analysis is as follows

[0103] Z(i, t) = K(n, i, t) z *cosα(n, i, t) z ;

[0104] Wherein, Z(i, t) is the characteristic state evaluation value of the signal data corresponding to the time point i in the period numbered t; K(n, i, t)z is the characteristic correlation evaluation value of the characteristic correlation group of the signal data channel characteristic component n corresponding to the time point i in the period numbered t; α(n, i, t)z is the horizontal angle of the signal data curve corresponding to the time point i in the period numbered t; The angle is taken as the angle between the line connecting the current time point corresponding signal curve point and the adjacent next time point signal curve point and the horizontal line;

[0105] The characteristic state evaluation value of the signal data at each time point in a cycle is combined to perform state steady analysis on the target cycle signal data; the characteristic state evaluation value of the signal data at the continuous time points on the target cycle signal curve is analyzed to obtain the change difference value of the signal curve data at adjacent time points; the change difference value analysis data of the signal curve data at the continuous time points on the target cycle signal curve is combined to analyze the steady state value of the cycle signal of the target cycle data; based on the steady state value analysis data of the cycle signal of the target cycle signal data, a stable threshold is determined to perform abnormality judgment on the target cycle signal of the current chip; wherein the steady state value analysis of the target cycle signal is

[0106]

[0107] Wherein, W is the steady state value of the target cycle signal; Z(i+1, t) is the characteristic state evaluation value of the signal data at the next time point adjacent to the time point i on the target cycle signal curve; Q(i, i+1) is the curvature of the signal curve at adjacent time points on the target cycle curve; Q(i, i+1)max and Q(i, i+1)min are the maximum and minimum values of the curvature of the signal curve at adjacent time points on the target cycle curve; Z(i, t)max and Z(i, t)min are the maximum and minimum values of the characteristic state evaluation value of the signal data on the target cycle signal curve;

[0108] By introducing the steady state threshold, if the steady state value of the target cycle signal is less than or equal to the threshold, it is judged that the current chip target cycle running state is good; otherwise, it is judged that the current chip target cycle running state is abnormal.

[0109] The corresponding signal curve data and the corresponding steady state value analysis data of the target cycle data of the chip test signal are output respectively;

[0110] The running state judgment results of each target cycle chip are output respectively, and the abnormal state chip and the abnormal target cycle signal data are marked.

[0111] It is apparent to those skilled in the art that the present application is not limited to the details of the foregoing exemplary embodiments, and that the present application can be implemented in other concrete forms without departing from the spirit or essential characteristics of the present application. Therefore, the embodiments should be considered in all respects as illustrative and not restrictive, the scope of the present application being indicated by the appended claims rather than by the foregoing description, and it is intended that all changes and modifications which come within the meaning and range of equivalency of the claims are resolvable by the following claims. Any reference signs in the claims should not be construed as limiting the claims to the figures in which the reference signs are used.

Claims

1. A chip test signal analysis and processing method based on big data, characterized in that: The method comprises the following steps: Collecting chip test signals, obtaining target test signals through signal preprocessing, and setting a period window to obtain target period signals from the target test signals; the signal preprocessing comprises signal frequency band positioning processing and signal noise reduction processing; Based on the target period signals, corresponding period time point signal channel decomposition processing is performed to obtain channel feature data of the corresponding time point signal data; the channel feature data comprises channel feature components and instantaneous feature frequencies of the corresponding channels; a feature correlation group of the corresponding time point signal data is constructed based on the channel feature data, the feature state data of the signal at the corresponding time point is analyzed based on the feature correlation group of the signal data at each time point in the period, and the signal state running steady state analysis of the continuous time points in the period is performed based on the feature state data of the signals at each time point, and the chip signal state is determined based on the analysis data; The signal noise reduction processing is to convert the initial signal data of the collected target test signals through time domain and frequency domain conversion, remove the interference signal data in the target test signals through a filtering algorithm, and output the target test signal data after noise reduction; The target test signal data after preprocessing is set to a period window to obtain period signals, and the obtained period signals are labeled to obtain target period signal data; The target period signal data corresponding to the labels is called to extract the signal data at each time point in the period; the signal channel decomposition analysis is performed based on the signal data at the corresponding time points, the decomposition layer is determined by wavelet transform decomposition, the channel feature component analysis and extraction are performed on the signal data at the corresponding time points, and the channel feature component set of the signal data at the corresponding time points is obtained; the energy distribution data of the corresponding channel feature components is analyzed based on the channel feature component data at the corresponding time points, and the energy entropy data set of the corresponding channel components is obtained; The instantaneous feature frequencies of the channel feature components at the corresponding time points are determined by empirical mode decomposition based on the channel feature components at the corresponding time points; the feature correlation group of the corresponding channel feature components is constructed by multiplying the energy entropy data and the instantaneous feature frequency data corresponding to the channel feature components at the corresponding time points.

2. The chip test signal analysis processing method based on big data according to claim 1, characterized in that: The signal frequency band positioning processing is to perform signal sampling debugging by using a signal collection device, adjust the sampling frequency according to the highest frequency of the chip test signals, and obtain the initial signal data of the target test signals in the target frequency range.

3. The chip test signal analysis processing method based on big data according to claim 2, characterized in that: Based on the feature correlation groups of the channel feature components corresponding to the signal data at each time point of the target period signals, the feature correlation groups of the signal data at each time point are analyzed respectively to determine the representation correlation groups of the signal data at each time point. The feature vectorization is performed on the characteristic correlation groups of the signal data at each time point in the corresponding period, the feature vector groups of the signal data at each time point in the corresponding period are determined, and the feature state data of the signal data at each time point in the period is analyzed to obtain the feature state evaluation values of the signal data at each time point in the corresponding period; The state stability of the target period signal data is analyzed in combination with the feature state evaluation values of the signal data at each time point in the period; the change difference value analysis of the signal curve data of adjacent time points is performed on the feature state evaluation values of the signal data of continuous time points on the target period signal curve; the analysis of the period signal running stability value of the target period data is performed in combination with the change difference value analysis data of the signal curve data of continuous time points on the target period signal curve; and the stable threshold is determined based on the analysis data of the period signal running stability value of the target period signal data to judge the abnormality of the target period signal of the current chip.

4. The chip test signal analysis processing method based on big data according to claim 3, characterized in that: The corresponding signal curve data and the corresponding period signal running stability value analysis data of each target period data divided by the chip test signal are output respectively; The running state judgment results of each target period chip are output respectively, and the abnormal state chip and the abnormal target period signal data are marked.

5. A big data based chip test signal analysis processing system for implementing a big data based chip test signal analysis processing method according to claim 1, characterized in that: The system comprises a target signal acquisition module, a signal feature analysis module, a signal period stability analysis module, and a chip data feedback module; The target signal acquisition module acquires the chip test signal, obtains the target test signal through signal preprocessing, sets a period window to retrieve the target period signal, the signal feature analysis module performs corresponding period time point signal channel decomposition processing based on the target period signal to obtain the channel feature data of the corresponding time point signal data; the channel feature data comprises channel feature components and instantaneous feature frequencies of corresponding channels; the feature correlation groups of the corresponding time point signal data are constructed in combination with the channel feature data; the signal period stability analysis module analyzes the feature state data of the corresponding time point signal based on the corresponding feature correlation groups of the signal data at each time point in the period, performs the running stability analysis of the signal state of continuous time points in the period based on the feature state data of each time point, and determines the signal state of the chip based on the analysis data; the chip data feedback module outputs the period signal running stability value analysis data of the chip test signal and marks the abnormal period signal data and the abnormal state chip.

6. The big data based chip test signal analysis processing system of claim 5, wherein: The target signal acquisition module comprises a test signal acquisition unit and a target signal preprocessing unit; The test signal acquisition unit is used to acquire the test signal data generated by the chip running; The target signal preprocessing unit comprises signal frequency band positioning processing and signal noise reduction processing; The signal frequency band positioning processing is performed by using the signal acquisition device to perform signal sampling debugging, adjusting the sampling frequency according to the highest frequency of the chip test signal, and obtaining the initial signal data of the target test signal in the target frequency interval; The signal noise reduction processing is to convert the initial signal data of the collected target test signal through time domain and frequency domain, remove the interference signal data in the target test signal through a filtering algorithm, and output the target test signal data after noise reduction.

7. The big data based chip test signal analysis processing system of claim 6, wherein: The signal feature analysis module includes a signal channel component analysis unit and a signal channel feature correlation group construction unit. The signal channel component analysis unit sets a period window to intercept a period signal after preprocessing the target test signal data, and performs label processing on the intercepted period signal data to obtain target period signal data. The target period signal data corresponding to the label is called, and the signal data at each time point in the period is extracted respectively. Based on the signal data at the corresponding time point, signal channel decomposition analysis is performed, the decomposition layer is determined by wavelet transform decomposition, the channel feature component analysis is performed on the signal data at the corresponding time point, and the channel feature component set of the signal data at the corresponding time point is obtained. Based on the channel feature component data analysis of the corresponding channel feature component, the energy entropy data set in the corresponding channel component is obtained. The signal channel feature correlation group construction unit performs empirical mode decomposition on each channel feature component of the corresponding time point signal to determine the instantaneous characteristic frequency of each channel feature component; and the feature correlation group of the corresponding channel feature component is constructed by combining the energy entropy data and the instantaneous characteristic frequency data of each channel feature component at the corresponding time point signal.

8. The big data based chip test signal analysis processing system of claim 7, wherein: The signal cycle steady state analysis module includes a signal feature state evaluation unit and a signal curve steady state analysis unit. The signal feature state evaluation unit analyzes the feature correlation group of the signal data at each time point based on the feature correlation group of the corresponding channel feature component of the target period signal at each time point, determines the representation correlation group of the signal data at each time point, and determines the representation correlation group of the signal data at each time point. Based on the feature vectorization of the representation correlation group of the signal data at each time point in the corresponding period, the representation vector group of the signal data at each time point in the period is determined, and the feature state data of the signal data at each time point in the period is analyzed to obtain the feature state evaluation value of the signal data at each time point in the corresponding period. The signal curve steady state analysis unit combines the feature state evaluation value of the signal data at each time point in the period to perform state steady state analysis on the target period signal data; analyzes the change difference value of the signal curve data of adjacent time points by analyzing the feature state evaluation value of the signal data of consecutive time points on the target period signal curve; analyzes the period signal running steady state value of the target period data by combining the change difference value analysis data of the signal curve data of consecutive time points on the target period signal curve; and determines the stable threshold based on the period signal running steady state value analysis data of the target period signal data to judge the abnormality of the target period signal of the current chip.

9. The big data based chip test signal analysis processing system of claim 8, wherein: The chip data feedback module includes a data feedback unit and an abnormality marking unit. The data feedback unit outputs the corresponding signal curve data and the corresponding period signal running steady state value analysis data of each target period data divided by the chip test signal respectively. The abnormality marking unit respectively outputs based on the running state judgment result of each target period chip, and marks the abnormal state chip and the abnormal target period signal data.

Images