An aircraft mission sensor ground test data alignment method

Abstract

The application relates to the technical field of airplane test data analysis, and discloses an airplane task sensor ground test data alignment method. The method firstly performs nearby alignment on data starting time, then realizes unification of sampling intervals based on resampling, performs upsampling operation on data with large sampling intervals, and realizes data filling based on interpolation. The application can effectively improve the overall quality of data, provide more accurate and reliable basis for subsequent data analysis and modeling, and promote airplane performance analysis and optimization.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data analysis, in particular to aircraft test data preprocessing, more particularly to an aircraft mission sensor ground test data alignment method. BACKGROUND

[0002] The aircraft related performance analysis by the aircraft mission sensor ground test data is an important basis for inferring whether the accuracy of the mission sensor during flight meets the factory index requirements. However, due to the large and complex flight system, each subsystem contains multiple sensors, the sampling rates of each sensor are inconsistent, and the sampling start time is deviated, so the test data from multiple sensors has the problems of different sampling intervals and time misalignment, which leads to that the mission sensor ground test data cannot be directly used for data analysis and modeling. Therefore, it is necessary to align the mission sensor ground test data, unify the sampling rate and sampling start time of each group of data, and improve the data utilization. For the time synchronization problem of time series data, currently, the sampling start time of data is unified by means of clock based on hardware assistance, and then the resampling method is used to unify the sampling interval. Due to the large number of aircraft sensors, and the inconsistent data start time is not only related to the sensor sampling start time, but also the transmission process may cause time deviation, therefore, the inconsistent problem of sampling start time cannot be completely solved based on hardware.

[0003] The invention patent with publication number CN111125632A discloses a variable sequence calculation method containing multiple spacecraft telemetry parameters. Although the scheme of the patent also involves the concept of data alignment, there are still the following problems: (1) In the data alignment, the time scale is first reserved according to the minimum value of the time scale and the corresponding content, the empty value is filled for the content without corresponding time scale, and then the interpolation is performed on the empty value; (2) The interpolated data is reserved for the maximum sampling interval, and the method defaults that the maximum sampling interval is an integer multiple of other sampling intervals. This leads to that in the actual application process, if the start time of part of the files is too early, it is difficult to ensure the quality of the early interpolation data by using the above-mentioned prior art method, and the process of filling the empty value first and then interpolating is very cumbersome; further, the method reserves the maximum sampling interval, filters the data with small sampling interval, which causes the missing of data characteristics and affects the results of subsequent data mining. SUMMARY

[0004] This application addresses the issue of time synchronization in aircraft test data, which involves multiple processes such as acquisition and transmission. It is difficult to directly solve the problem of inconsistent start times by adding a clock in hardware. Therefore, a method for aligning ground test data from aircraft mission sensors is proposed. First, the start times of the data are aligned to the nearest possible values. Then, resampling is used to unify the sampling intervals. Data with large sampling intervals are upsampled, and interpolation is used to fill in the gaps. This application effectively improves the overall quality of the data, providing a more accurate and reliable foundation for subsequent data analysis and modeling, and promoting aircraft performance analysis and optimization.

[0005] To achieve the aforementioned objectives, the technical solution of this application is as follows:

[0006] A method for aligning ground test data of aircraft mission sensors includes the following steps:

[0007] The sampling start time of the input data is statistically analyzed, and the start time of the data is unified by truncation and nearest alignment.

[0008] Determine the resampling interval, determine the timescale of the aligned data based on the sampling start time and the resampling interval, and then resample the data;

[0009] The upsampled data is filled using the pre-interpolation method.

[0010] Preferably, the step of statistically analyzing the sampling start time of the input data and unifying the start time of the data through truncation and nearest alignment includes:

[0011] The sampling start time of all test data is statistically analyzed, and the latest start time is selected as the target time for time alignment. Other data are then extracted, and the test data from the nearest time to the target time and thereafter are retained to obtain the common time period of all data.

[0012] The start times of all the extracted test data were statistically analyzed again, and the median was selected and rounded down as the alignment target start time. The start times of other data sampling were directly modified to the alignment target start time.

[0013] For example, if the sampling start times are t 1,0 ,t 2,0 ,t 3,0 ,......,t n,0 , t i,0 If the value is the maximum among them, then the remaining test data will be truncated, and the truncated t will be retained. i,0 The data from the nearest time point and thereafter is truncated to a sampling start time of t. 1,0 ',t 2,0 ',t 3,0 ',......,tn,0 wherein t j,0 is the median, based on t j,0 corrects the test data start time, and the corrected sample start time is t j,0 , and the synchronization of the complete data start time is completed.

[0014] As preferably, the determining of the resampling interval, the determination of the time scale of the aligned data based on the sample start time and the resampling interval, and the resampling of the data, comprise:

[0015] The time scales of all data are counted, and the minimum value of the sample interval is selected as the target sample interval T, and the time scale is reset based on the sample start time t j,0 ' and the target sample interval T, and the reset time scale is t j,0 ', t j,0 +T, t j,0 +2*T,......, t j,0 +m*T, and the reset time scale is arranged in order to obtain the aligned reset time scale.

[0016] As preferably, the filling of the upsampling data by the forward interpolation method comprises:

[0017] The nearest original time scale to each reset time scale is taken as the corresponding time scale, and the entire row of data corresponding to the corresponding time scale is taken as the data of the reset time scale; for the upsampling data, the data corresponding to the newly added time scale is filled by the forward interpolation method, that is, the data corresponding to the nearest previous original time scale corresponding to the newly added time scale is filled in.

[0018] As preferably, the nearest time point refers to the time point closest to the target time of interception.

[0019] As preferably, the nearest original time scale refers to the time scale closest to each reset time scale.

[0020] As preferably, the time scale is the time stamp.

[0021] The beneficial effects of the present application are:

[0022] (1) The method of the present application first aligns the data start time in the vicinity, and through the way of truncation and alignment of the start time in the vicinity, the information related to the change of the aircraft state can be retained, the interference of irrelevant information is reduced, the overall quality of the data is effectively improved, and a more accurate and reliable basis is provided for subsequent data analysis and modeling.

[0023] (2) The method of the present application synchronizes the time intervals of aircraft test data using resampling, eliminating errors and noise that may be introduced by different collection times, ensuring data consistency and comparability. In addition, by adjusting the time interval, the number of data points can be reduced while preserving the trend and characteristics of the data, simplifying the data analysis process and improving the efficiency of data processing. Resampling can also perform data smoothing, reducing noise interference and improving data quality and reliability. Therefore, it can provide more accurate and reliable basis for aircraft test data analysis, modeling and prediction, etc., and promote aircraft performance analysis and optimization. BRIEF DESCRIPTION OF DRAWINGS

[0024] The foregoing and the following detailed description of the present application will become more apparent when read in conjunction with the following drawings, in which:

[0025] Figure 1 is a flowchart of the method of the present application. DETAILED DESCRIPTION

[0026] In order for those skilled in the art to better understand the technical solutions in the present application, the following will further illustrate the technical solutions for achieving the purposes of the present application through several specific embodiments. It should be noted that the technical solutions claimed in the present application include but are not limited to the following embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor shall fall within the scope of protection of the present application.

[0027] The aircraft-related performance analysis by aircraft mission sensor ground test data is an important basis for inferring whether the accuracy of the mission sensor during flight meets the factory index requirements. However, due to the large and complex flight system, each subsystem contains multiple sensors, and the sampling rates of each sensor are inconsistent, with deviations in the sampling start time, so the test data from multiple sensors have different sampling intervals and time misalignment, which makes the mission sensor ground test data cannot be directly used for data analysis and modeling. Therefore, it is necessary to align the mission sensor ground test data, unify the sampling rate and sampling start time of each group of data, and improve data utilization. For the time synchronization problem of time series data, currently, the sampling start time of data is unified based on hardware assistance, and then resampling is used to unify the sampling interval. Due to the large number of aircraft sensors, and the fact that the data start time is not only related to the sensor sampling start time, but also may have time deviations during transmission, therefore, it is not possible to completely solve the problem of inconsistent sampling start time based on hardware.

[0028] Based on this, the embodiment of the present application proposes an airplane task sensor ground test data alignment method, which firstly aligns the data starting time in the vicinity, then realizes the uniformity of the sampling interval based on resampling, performs upsampling operation on the data with larger sampling interval, and realizes the filling of the data based on interpolation.

[0029] The embodiment discloses an airplane task sensor ground test data alignment method, and the accompanying drawings Figure 1 illustrate the embodiment of the present application.

[0030] Step S1. Statistics of the sampling starting time of the input data, and the starting time of the data is unified by truncation and nearest alignment.

[0031] In the embodiment, the sampling starting time of the input data is counted, and the starting time of the data is unified by truncation and nearest alignment, which is as follows:

[0032] First, the sampling starting time of all test data is counted, the latest starting time is selected as the cutting target time of time alignment, and other data is cut off, and the test data from the nearest time point of the cutting target time and after the cutting target time is retained, and the common time period of each data is obtained.

[0033] Secondly, the starting time of all the test data after cutting is counted, the median is selected and rounded as the alignment target starting time, and the sampling starting time of other data is directly modified to the alignment target starting time.

[0034] In the embodiment, the data in other files is cut off according to the latest starting time, and only the time period covered by all files is retained, for example, the latest is 4 minutes 30 seconds 80 milliseconds, and a certain data starts from 1 minute 2 seconds, so the data before it is cut off to the nearest time point of 4 minutes 30 seconds 80 milliseconds.

[0035] In the embodiment, it should be noted that the nearest time point of the cutting target time is the time point closest to the cutting target time, for example: the cutting target time is 80 ms, the last time point in the file is 70 ms, and the next time point is 110 ms, then the cutting starts from 70 ms.

[0036] In the embodiment, it should be noted that the common time period is the time period of data overlap.

[0037] The embodiment takes the test data of two modules and As the original data, the sampling intervals of the two test data are 40 ms and 160 ms respectively, and the starting times are 00:00:00_477 and 00:00:00_558 respectively. After splicing, the original time scale and the corresponding data of the first test data are shown in Table 1, and the original time scale and the corresponding data of the second test data are shown in Table 1.

[0038] Table 1 Original time scale and data of two groups of test data

[0039] Time scale 1 Time scale 2 Sequence 1 Sequence 2 00:00:00_477 00:00:00_558 2.66 0.2 00:00:00_478 00:00:00_559 2.66 0.2 00:00:00_517 00:00:00_717 2.66 0.2 00:00:00_557 00:00:00_718 2.65 0.2 00:00:00_558 00:00:00_882 2.65 0.3 00:00:00_596 00:00:01_043 2.65 0.3 00:00:00_637 00:00:01_198 2.65 0.3 00:00:00_680 00:00:01_199 2.66 0.3 00:00:00_716 00:00:01_363 2.66 0.4 00:00:00_717 00:00:01_520 2.66 0.4

[0040] Since the starting times are 00:00:00_477 and 00:00:00_558 respectively, the data from the first 558 ms to the end of the first data and the data from the beginning to 717 ms of the second data are taken as the common time period of the two ends. After the interception, the starting times of the two ends are 558 ms, and the median is 558 ms. After rounding, the new sampling starting time is set to 560 ms.

[0041] The minimum sampling interval 40 ms is selected as the new sampling interval, and the target time scale is obtained and arranged in time sequence as 00:00:00_560, 00:00:00_600, 00:00:00_640, 00:00:00_680 and 00:00:00_720.

[0042] The corresponding data of the previous original data closest to each target time scale is selected for interpolation, that is, the data corresponding to 558 ms of the two ends is filled into 560 ms, the data corresponding to 596 ms of the first data and 559 ms of the second data is filled into 600 ms, and so on. The final aligned data is shown in Table 2.

[0043] Table 2 Time scale and data of aligned data

[0044] Time scale Sequence 1 Sequence 2 00:00:00_560 2.65 0.2 00:00:00_600 2.65 0.2 00:00:00_640 2.65 0.2 00:00:00_680 2.65 0.2 00:00:00_720 2.66 0.2 .

[0045] Step S2. Determine the resampling interval, determine the time scale of the aligned data based on the sampling starting time and the resampling interval, and resample the data.

[0046] In this embodiment, the determination of the resampling interval, the determination of the time scale of the aligned data based on the sampling starting time and the resampling interval, and the resampling of the data are as follows:

[0047] The time scales of all data are counted, the minimum sampling interval is selected as the target sampling interval, the time scale is reset based on the sampling starting time and the target sampling interval, the reset time scale is obtained, and the aligned reset time scale is obtained by arranging the reset time scale in order.

[0048] In the embodiment, the sampling start time at this time is the alignment target start time in the last step.

[0049] In the embodiment, it is to be noted that resampling refers to converting original time series data into data of another time frequency, that is, adjusting the time interval of original data into a longer or shorter time interval.

[0050] In the embodiment, it is to be further noted that the time scale is the time stamp, which refers to the group of data recording time.

[0051] Step S3. The upsampling data is filled by using the forward interpolation method.

[0052] In the embodiment, the upsampling data is filled by using the forward interpolation method, and the specific process is as follows.

[0053] The original time stamp most adjacent to each reset time stamp is taken as the corresponding time stamp, and the entire row of data corresponding to the time stamp is taken as the data of the reset time stamp; for the upsampling data, the data corresponding to the most adjacent previous original time stamp corresponding to the new time stamp is filled in.

[0054] In the embodiment, it is to be noted that the upsampling is to make the interval more dense by difference value for the data with large sampling interval. For example, the interval is 80 ms, and the difference value becomes 40, which is more dense and is upsampling.

[0055] The above is only the preferred embodiment of the present application, and does not hinder the present application in any form, and any simple modification and equivalent change of the above embodiment according to the technical essence of the present application fall within the protection scope of the present application.

Claims

1. An aircraft mission sensor ground test data alignment method, characterized by, The method comprises the following steps: statistically determining the sampling start time of input data, unifying the start time of data by truncation and nearest alignment; determining a resampling interval, determining the time scale of aligned data based on the sampling start time and the resampling interval, and resampling the data; using the forward interpolation method to fill in the upsampling data; wherein The statistical determination of the sampling start time of input data, the unification of the start time of data by truncation and nearest alignment comprises: statistically determining the sampling start time of all test data, selecting the latest start time as the truncation target time of time alignment, and truncating other data, retaining the test data from the nearest time point of the truncation target time and thereafter, and obtaining the common time period of each data; statistically determining the start time of all test data after truncation again, selecting the median and rounding it as the alignment target start time, and directly modifying the sampling start time of other data to the alignment target start time; The determination of the resampling interval, the determination of the time scale of aligned data based on the sampling start time and the resampling interval, and the resampling of the data comprises: statistically determining the time scale of all data, selecting the minimum sampling interval as the target sampling interval, resetting the time scale based on the sampling start time and the target sampling interval to obtain the reset time scale, arranging the reset time scale in order to obtain the aligned reset time scale; The filling of the upsampling data by using the forward interpolation method comprises: taking the nearest original time scale of each reset time scale as the corresponding time scale, taking the entire row of data corresponding to the corresponding time scale as the data of the reset time scale, and for the upsampling data, filling the entire data corresponding to the nearest previous original time scale corresponding to the new time scale.

2. An aircraft mission sensor ground test data alignment method as in claim 1 wherein, The nearest time point refers to the time point closest to the truncation target time.

3. A method for aligning ground test data for an aircraft mission sensor according to claim 1, wherein, The nearest original time scale refers to the time scale closest to each reset time scale.

4. The method of claim 1, wherein, The time scale is the timestamp.

5. The method of claim 1, wherein, The resampling refers to converting the original time series data into data of another time frequency.

6. A method for aligning ground test data for an aircraft mission sensor according to claim 1, wherein, The upsampling refers to making the interval more dense by difference for data with a large sampling interval.

Images