A system and method for automatic interpretation of test data

Abstract

The present invention discloses an automatic test data interpretation system, comprising: a data acquisition module, a data preprocessing module, a data type judgment module, and a data interpretation module; the data acquisition module is used to acquire pre-recorded test data; the data preprocessing module is used to preprocess the acquired test data; the data type judgment module is used to judge the type of the test data according to the number of jumps in the data; the data interpretation module is used to select corresponding feature extraction methods according to different types of the test data, and based on the extracted feature values and a preset threshold, judge whether the data is abnormal. The present invention first judges the type of the input data, then extracts the features of different types of data, and then compares the feature values with the preset threshold or threshold to judge whether the data is abnormal data. By using the method of automatic computer judgment, the accuracy and efficiency of judgment are improved, and the labor cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of data interpretation, and in particular to a system and method for automatically interpreting test data. Background Art

[0002] For complex equipment with multiple systems, the testing workload is very large after the whole machine is integrated. After a test project is completed, the relevant personnel of the subsystem need to interpret the data. Only after the data is judged to be correct can the next test be carried out. However, the equipment test interpretation is done manually, and the interpretation quality is affected by the designer's lack of experience, resulting in many problems such as test omissions and misjudgments. Summary of the invention

[0003] In order to improve the accuracy of the interpretation of the measured data, the present invention provides a test data automatic interpretation system and method, which first judges the input data type, then extracts the features of different types of data, and then compares the feature value with the preset threshold or threshold to determine whether the data is abnormal data. The computer automatic judgment method is used to improve the accuracy and efficiency of the judgment and reduce the labor cost. The specific technical solution is as follows:

[0004] The present invention provides a test data automatic interpretation system, a data acquisition module, a data preprocessing module, a data type judgment module and a data interpretation module; wherein:

[0005] The data acquisition module is used to acquire pre-recorded test data;

[0006] The data preprocessing module is used to preprocess the acquired test data;

[0007] A data type determination module, used to determine the type of the test data according to the number of jumps in the data;

[0008] The data interpretation module is used to select a corresponding feature extraction method according to different types of the test data, and judge whether the data is abnormal based on the extracted feature value and a preset threshold.

[0009] Furthermore, the data preprocessing module is used to specifically adopt a threshold comparison algorithm to judge the test data with a preset threshold, and eliminate abnormal data whose test data is greater than the threshold.

[0010] Furthermore, the data type determination module is used to specifically adopt a left-side average detection algorithm to detect the amplitude average of several continuous signals, count the number of step points appearing in the several continuous signals, and determine the type of test data according to the number of step points.

[0011] Further, the types of the test data include: step signal, pulse signal, and level signal; the data type determination module is specifically used to determine the type of the test data by the following method:

[0012] If the test data has no transition and is constantly a high signal or a low signal, it is determined that the test signal is a level signal;

[0013] If the test data has one transition and is constantly a high signal or a low signal after the transition, it is determined that the test signal is a step signal;

[0014] If the test data has two transitions and the directions of the two transitions are opposite, it is determined that the test data is pulse data.

[0015] Further, the data interpretation module is specifically used to extract different characteristic data from the test data by using different feature extraction algorithms according to different types of the test data; different preset test indexes are adopted for different types of the test data, and the characteristic data is compared with the test indexes to determine whether the test data is abnormal.

[0016] Further, the characteristic data includes: type, time, transition direction, number of transitions, average amplitude before transition, maximum value before transition, minimum value before transition, average amplitude after transition, maximum value after transition, minimum value after transition, and if it is a pulse signal, pulse width.

[0017] The second aspect of the present invention provides a method for automatically interpreting test data, including the steps of:

[0018] Obtain the pre-recorded test data;

[0019] Preprocess the obtained test data;

[0020] Judge the type of the test data according to the number of transitions of the data;

[0021] According to different types of the test data, select the corresponding feature extraction method, and based on the extracted feature values and the preset threshold, judge whether the data is abnormal.

[0022] Further, the preprocessing of the obtained test data specifically includes the steps of: adopting a threshold comparison algorithm to judge the test data with a preset threshold, and removing the abnormal data where the test data is greater than the threshold.

[0023] Further, judging the type of the test data according to the number of jumps of the data specifically includes the steps of: using a left - hand average detection algorithm to detect the amplitude average value of several consecutive signals, counting the number of step points in the several consecutive signals, and judging the type of the test data according to the number of step points.

[0024] The third aspect of the present invention provides an electronic device, which includes a computer - readable storage medium and a processor; wherein, the computer - readable storage medium is used to store the above - mentioned test data automatic interpretation method;

[0025] When the processor runs, it executes the test data automatic interpretation method.

[0026] The beneficial effects of the present invention are as follows:

[0027] The present invention discloses a test data automatic interpretation system, including: a data acquisition module, a data pre - processing module, a data type judgment module, and a data interpretation module; wherein: the data acquisition module is used to acquire pre - recorded test data; the data pre - processing module is used to pre - process the acquired test data; the data type judgment module is used to judge the type of the test data according to the number of jumps of the data; the data interpretation module is used to select a corresponding feature extraction method according to different types of the test data, and judge whether the data is abnormal based on the extracted feature values and a preset threshold. The present invention first judges the type of the input data, then extracts the features of different types of data, and then compares the feature values with a pre - set threshold or limit to judge whether the data is abnormal data. By using the method of automatic computer judgment, the accuracy and efficiency of judgment are improved, and the labor cost is reduced. Description of the Drawings

[0028] Figure 1 is a schematic diagram of the test data automatic interpretation system provided by an embodiment of the present invention;

[0029] Figure 2 is a schematic diagram of the test data type judgment process in an embodiment of the present invention;

[0030] Figure 3 is the original data of the step signal of the present invention;

[0031] Figure 4 is the original data of the pulse signal of the present invention;

[0032] Figure 5 is a flowchart of the specific implementation of the frame interpretation module provided by the present invention. Detailed Embodiments

[0033] See Figure 1, an embodiment of the present invention provides an automatic test data interpretation system, including: a data acquisition module, a data preprocessing module, a data type judgment module, and a data interpretation module; where:

[0034] The data acquisition module is used to acquire pre-recorded test data;

[0035] The data preprocessing module is used to preprocess the acquired test data;

[0036] The data type judgment module is used to judge the type of the test data according to the number of jumps of the data;

[0037] The data interpretation module is used to select a corresponding feature extraction method according to different types of the test data, and judge whether the data is abnormal based on the extracted feature values and a preset threshold.

[0038] The above test data is directly acquired from a test device and then imported into the system for data interpretation.

[0039] The type of the above test data can be judged by the data type judgment module. The data type judgment module is specifically used to adopt a left average detection algorithm to detect the amplitude average value of several consecutive signals, count the number of step points in several consecutive signals, and judge the type of the test data according to the number of step points. The specific calculation process is as follows: Adopt the left average detection method to detect the background and select the average of the signal amplitudes of the left P points (P can be set according to the signal characteristics, such as set to 5 in this implementation), the detection threshold is set to THR (the detection threshold can be set according to the characteristics of the test signal, such as for a TTL signal with an amplitude of 3V, the threshold can be set to 0.8), and the current detection point is recorded as x n , and the detected result is recorded as jc n , then

[0040]

[0041] If two consecutive frames of jc n are non-zero, the first non-zero moment is recorded as 1 step point, and the step point information (step index value, step time, step direction) is stored; otherwise, it is considered a glitch, and the index value of the glitch point is recorded.

[0042] See Figure 2 , which is a schematic diagram of the test data type judgment process in the embodiment of the present invention. See Figure 3 is the original data of the step signal in the present invention, Figure 4 is the original data of the pulse signal in the present invention. The types of the test data include: step signals, pulse signals, and level signals; the data type judgment module is specifically used to judge the type of the test data by the following method:

[0043] If there is no transition in the test data and it is constantly a high signal or a low signal, it is determined that the test signal is a level signal;

[0044] If there is one transition in the test data and it is constantly a high signal or a low signal after the transition, it is determined that the test signal is a step signal;

[0045] If there are two transitions in the test data and the directions of the two transitions are opposite, it is determined that the test data is pulse data. The number of steps of the pulse signal appears in pairs. Denote the time of the step signal corresponding to the i-th pulse as T 2i-1 , T 2i , then the pulse width is pw = T 2i -T 2i-1 .

[0046] In an alternative embodiment of the embodiment of the present invention, it further includes: judging and removing outliers in the acquired test data. The outliers include wild values, glitches, etc. Specifically: set the wild value point threshold, perform wild value removal on the original data, and the removal method is: if the absolute value of the signal value is greater than the threshold, it is considered a wild value point and filled with the previous normal amplitude value. For the detected glitch points, fill them with the amplitude value of the previous normal point.

[0047] The characteristic data of the above-mentioned level signal includes: the maximum value, minimum value, average value, and accuracy of the signal within the judgment time period.

[0048] The calculation process of the average value is:

[0049]

[0050] Among them, is the signal average value, L is the number of sampling points within the judgment time period, and x i is the sampling value of the i-th point signal.

[0051] Accuracy (1σ) calculation:

[0052] The characteristic data of both the step signal and the pulse signal includes: type, time, transition direction, number of transitions, average amplitude before transition, maximum value before transition, minimum value before transition, average amplitude after transition, maximum value after transition, minimum value after transition, and if it is a pulse signal, the pulse width.

[0053] According to the comparison between the extracted signal characteristics and the standard solution signal, if they match, the output conclusion is qualified, and the signal matching principle is as follows:

[0054] If it is a level signal: judge whether the signal amplitude meets the standard requirements.

[0055] If it is a pulse signal: Determine whether the number of pulses, width, and pulse amplitude meet the requirements of the standard solution. Table 1 is the pulse signal detection table.

[0056] Table 1

[0057]

[0058] If it is a step signal: Determine whether the amplitude before the jump and the amplitude after the jump meet the requirements of the standard solution. Table 2 is the step signal interpretation result.

[0059] Table 2

[0060] Test Parameters Technical Requirements Measurement Moment s Measured Value Conclusion 0-2s DA -0.2～0.2 64.200000 -0.022786 Qualified 2-4s DA 9.8～10.2 66.200000 10.000980 Qualified 4-6s DA 9.8～10.2 68.200000 10.000980 Qualified 6-8s DA 9.8～10.2 70.200000 10.000980 Qualified 8-10s DA -0.2～0.2 72.200000 0.001000 Qualified 10-12s DA -1.2～-0.8 74.200000 -1.022620 Qualified 12-14s DA -5.2～-4.8 76.200000 -5.038360 Qualified 14-16s DA -10.2～-9.8 78.200000 -10.077720 Qualified 16-18s DA -0.2～0.2 80.200000 0.001000 Qualified 18-20s DA 0.8～1.2 82.200000 0.945880 Qualified 20-22s DA 4.8～5.2 84.200000 5.040360 Qualified 22-24s DA 9.8～10.2 86.200000 10.000980 Qualified 24-26s DA -0.2～0.2 88.200000 0.001000 Qualified

[0061] In an alternative embodiment of the embodiment of the present invention, a frame data interpretation module is further included. Refer to Figure 5 , which is the flowchart of the specific implementation of the frame interpretation module provided by the present invention. The frame data interpretation includes the following processes:

[0062] (1) Interpretation of the starting condition: According to the protocol regulations, the moment when the frame count is 1 is used as the starting point. To prevent misjudgment of the zero point caused by abnormal values, search for three consecutive frames in the data string that are [1, 2, 3] respectively. Then, record the frame with a frame count of 1 as the starting frame. There may be multiple starting frames in one data.

[0063] (2) Interpretation of the termination condition: Subtract the previous frame count Fn from the next frame count Fn+1, denoted as Dn, Dn = Fn+1 – Fn. If Dn is not 1, it means the counting is not continuous and is marked as abnormal. When one of the following two situations occurs, it is judged as the end point:

[0064] a) There is no abnormal point. If there is still a starting point after this starting point, then use the previous frame of the next starting point as the termination point. If there is no starting point afterwards, then use the last frame as the termination point.

[0065] b) There are abnormal points, and the abnormal points appear continuously for 100 frames. Then, use the frame before the appearance of the abnormality as the termination point.

[0066] (3) Missing frame statistics: Each starting point to the termination point is used as a statistical unit. According to the standard solution value of the termination frame - the starting count value as the theoretical frame count quantity, denoted as N1, and the actual stored number of frame counts as the actual count quantity, denoted as N2. N1 - N2 is the number of missing frames.

[0067] (4) Wrong frame statistics: Each starting point to the termination point is used as a statistical unit. According to the value of the next frame minus the previous frame Dn, starting from D2, if Dn is not 1 and satisfies (Dn-1 + Dn = 2) with the previous difference, then the number of wrong frames is incremented by 1 (indicating that the next frame is normal). If the difference does not satisfy (Dn-1 + Dn = 2), then the wrong frames are accumulated upwards (indicating that the next frame is also abnormal).

[0068] For the input data shown in Figure 4 , the frame count reading results are shown in Table 3, and Table 3 is the frame count reading result.

[0069] Table 3

[0070]

[0071] The present invention discloses an automatic test data reading system, including: a data acquisition module, a data preprocessing module, a data type judgment module, and a data reading module; wherein: the data acquisition module is used to acquire pre-recorded test data; the data preprocessing module is used to preprocess the acquired test data; the data type judgment module is used to judge the type of the test data according to the number of jumps of the data; the data reading module is used to select a corresponding feature extraction method according to different types of the test data, and judge whether the data is abnormal based on the extracted feature values and a preset threshold. The present invention first judges the type of the input data, then extracts the features of different types of data, and then compares the feature values with the preset threshold or threshold to judge whether the data is abnormal data. By using the method of automatic computer judgment, the accuracy and efficiency of judgment are improved, and the labor cost is reduced.

[0072] The second aspect of the present invention provides an automatic test data reading method, including the steps of:

[0073] Acquire pre-recorded test data;

[0074] Preprocess the acquired test data;

[0075] Judge the type of the test data according to the number of jumps of the data;

[0076] Select a corresponding feature extraction method according to different types of the test data, and judge whether the data is abnormal based on the extracted feature values and a preset threshold.

[0077] Further, the preprocessing of the acquired test data specifically includes the steps of: using a threshold comparison algorithm to judge the test data with a preset threshold, and eliminating abnormal data where the test data is greater than the threshold.

[0078] Further, the step of judging the type of the test data according to the number of jumps of the data specifically includes the steps of: using a left-side average detection algorithm to detect the amplitude average value of several consecutive signals, counting the number of step points in several consecutive signals, and judging the type of the test data according to the number of step points.

[0079] A third aspect of the present invention provides an electronic device, which includes a computer-readable storage medium and a processor; wherein, the computer-readable storage medium is used to store the automatic test data interpretation method described in any one of the above;

[0080] When the processor runs, it executes the automatic test data interpretation method.

Claims

1. A test data automatic interpretation system, It is characterized in that Data acquisition module, data preprocessing module, data type judgment module and data interpretation module; wherein: The data acquisition module is used to acquire pre-recorded test data; The data preprocessing module is used to preprocess the acquired test data; A data type determination module, used to determine the type of the test data according to the number of jumps in the data; A data interpretation module is used to select a corresponding feature extraction method according to different types of the test data, and judge whether the data is abnormal based on the extracted feature value and a preset threshold; The data type determination module is used to specifically adopt the left average detection algorithm to detect the amplitude average of several continuous signals, count the number of step points appearing in several continuous signals, and determine the type of test data according to the number of step points; the specific process is: The left average detection method is used to detect the background. The average signal amplitude of P points on the left is selected. The detection threshold is set to THR. The current detection point is recorded as x n , the result after detection is recorded as jc n ,but If jc is non-zero for two consecutive frames n and non-zero for the first time, the moment is recorded as a step point, and the information of the step point is stored; otherwise, it is considered a glitch, and the index value of the glitch point is recorded. For the detected glitch points, they are filled with the amplitude value of the previous normal point. It also includes a frame data reading module for: Determine the start condition: According to the protocol, the moment when the frame count is 1 is taken as the start point. To prevent abnormal values ​​from misjudging the zero point, find three consecutive frames in the data string that are [1, 2, 3] respectively. The frame count of 1 is recorded as the start frame. There may be multiple start frames in one data. Determine the termination condition: Use the count value Fn+1 of the next frame to subtract the count value Fn of the previous frame, recorded as Dn, Dn=Fn+1-Fn, if Dn is not 1, it means that the count is discontinuous, marked as abnormal, when one of the following two situations occurs, it is judged as the end point: a) There is no abnormal point, if there is another starting point after the starting point, then the frame before the next starting point is used as the end point, if there is no starting point after that, then the last frame is used as the end point; b) There is an abnormal point, and the abnormal point appears for 100 frames continuously, then the point before the abnormality appears is used as the end point; Missed frame statistics: Each starting and ending point is taken as a statistical unit. The theoretical frame count is calculated based on the end frame standard solution value minus the start count value, recorded as N1. The actual number of frame counts stored is the actual count, recorded as N2. N1-N2 is the number of lost frames. Error frame statistics: Each starting and ending point is taken as a statistical unit. Based on the value Dn of the next frame minus the previous frame, starting from D2, if Dn is not 1 and the difference with the previous frame satisfies Dn-1+Dn=2, the number of error frames is increased by 1. If the difference does not satisfy Dn-1+Dn=2, the error frames are accumulated upward.

2. The automatic test data interpretation system according to claim 1, It is characterized in that The data preprocessing module is used to specifically use a threshold comparison algorithm to judge the test data with a preset threshold, and eliminate abnormal data whose test data is greater than the threshold.

3. The automatic test data interpretation system according to claim 1, It is characterized in that The types of the test data include: step signal, pulse signal and level signal; the data type determination module is used to determine the type of the test data specifically by the following method: If the test data has no jump and is always a high signal or always a low signal, then the test signal is determined to be a level signal; If there is a jump in the test data and it is constantly a high signal or a low signal after the jump, it is determined that the test signal is a step signal; If there are two jumps in the test data and the directions of the two jumps are opposite, it is determined that the test data is pulse data.

4. The test data automatic interpretation system according to claim 1, characterized in that the data interpretation module is specifically used to extract different characteristic data from the test data by using different feature extraction algorithms according to different types of test data; different preset test indexes are used for different types of test data, and the characteristic data is compared with the test indexes to determine whether there is an abnormality in the test data.

5. The test data automatic interpretation system according to claim 4, characterized in that the characteristic data of the level signal includes: the maximum value, minimum value, average value, and accuracy of the signal within the interpretation time period; The characteristic data of both the step signal and the pulse signal includes: type, time, jump direction, number of jumps, average amplitude before jump, maximum value before jump, minimum value before jump, average amplitude after jump, maximum value after jump, minimum value after jump, and if it is a pulse signal, the pulse width.

6. A test data automatic interpretation method, characterized in that it includes the steps of: obtaining pre-recorded test data; preprocessing the obtained test data; judging the type of the test data according to the number of jumps of the data; selecting a corresponding feature extraction method according to different types of the test data, and judging whether the data is abnormal based on the extracted feature values and preset thresholds; The step of judging the type of the test data according to the number of jumps of the data includes using a left-side average detection algorithm to detect the average amplitude of several consecutive signals, counting the number of step points in several consecutive signals, and judging the type of the test data according to the number of step points; The specific process is: Adopt the left - hand average detection method. For background detection, select the average of the signal amplitudes of the left - hand P points. Set the detection threshold as THR, and denote the current detection point as x n , and denote the detection result as jc n , then If jc is non-zero for two consecutive frames n and non-zero for the first time, the moment is recorded as a step point and the information of this step point is stored; otherwise, it is considered a glitch, and the index value of the glitch point is recorded. For the detected glitch points, they are filled with the amplitude value of the previous normal point. Interpretation start condition: According to the protocol regulations, the moment when the frame count is 1 is used as the starting point. To prevent misjudging the zero point due to abnormal values, find three consecutive frames in the data string that are [1, 2, 3] respectively, then record the frame with frame count 1 as the starting frame, and there may be multiple starting frames in one data; Interpretation end condition: Subtract the previous frame count Fn from the next frame count Fn+1, denoted as Dn, Dn = Fn+1 - Fn. If Dn is not 1, it means the counting is not continuous and is marked as abnormal. When one of the following two situations occurs, it is judged as the end point: a) There is no abnormal point. If there is still a starting point after this starting point, then use the frame before the next starting point as the end point. If there is no starting point after that, then use the last frame as the end point; b) There are abnormal points, and the abnormal points continuously appear for 100 frames, then use the frame before the appearance of the abnormality as the end point; Missing frame statistics: Each starting point to the end point is used as a statistical unit. According to the theoretical frame count quantity which is the standard solution value of the end frame - the starting count value, denoted as N1, and the actual storage number of the frame count is the actual count quantity, denoted as N2, N1 - N2 is the number of missing frames; Error frame statistics: Each starting point to the ending point is used as a statistical unit. According to the value Dn obtained by subtracting the previous frame from the next frame, starting from D2, if Dn is not 1 and satisfies Dn-1 + Dn = 2 with the previous difference, the number of error frames is incremented by 1. If the difference does not satisfy Dn-1 + Dn = 2, the error frames are accumulated upward.

7. The automatic test data interpretation method according to claim 6, characterized in that the preprocessing of the obtained test data specifically includes the steps of: using a threshold comparison algorithm to judge the test data against a pre-set threshold, and eliminating abnormal data where the test data is greater than the threshold.

8. The automatic test data interpretation method according to claim 6, characterized in that judging the type of the test data according to the number of jumps of the data specifically includes the steps of: using a left-side average detection algorithm to detect the amplitude average value of several consecutive signals, counting the number of step points in several consecutive signals, and judging the type of the test data according to the number of step points.

9. An electronic device, characterized in that the electronic device includes a computer-readable storage medium and a processor; wherein, the computer-readable storage medium is used to store the automatic test data interpretation method according to any one of claims 6-8; when the processor runs, it executes the automatic test data interpretation method.

Images