Satellite test data analysis method and system

Abstract

The embodiment of the invention provides a satellite test data analysis method and system, and the method comprises the steps: firstly obtaining an original data set which is collected in a satellite test process and comprises multiple types of observation data units, carrying out the feature conversion processing of the original data set, and obtaining an analysis feature set comprising time sequence features and space features, then calling a pre-trained satellite test state evaluation model to carry out multi-dimensional state evaluation, generating a test state evaluation result, determining exception type and influence range information according to the evaluation result, and finally generating a test optimization instruction containing an exception positioning identifier based on the information and feeding back the test optimization instruction to the satellite control terminal. Satellite test problems can be solved in time, test success rate and efficiency are improved, and risk cost is reduced.

Description

Technical Field

[0001] The present application relates to the field of aerospace computer technology, and in particular to a satellite test data analysis method and system. Background Art

[0002] In the aerospace industry, satellite testing is a crucial step in ensuring stable performance and proper functionality during launch and operation. Satellite testing generates a large amount of observational data from various test modules, covering every subsystem and functional module of the satellite. Existing methods for analyzing satellite test data primarily rely on manual inspection and analysis of each type of data. This approach is not only inefficient but also prone to missing critical information due to human negligence. Furthermore, existing analysis methods often only analyze data from a single test module in isolation, making it difficult to comprehensively and deeply explore the relationships between different test modules and accurately assess the overall status of satellite testing. When an anomaly is discovered, it is difficult to quickly and accurately determine the anomaly type and its impact range, resulting in untimely and inaccurate adjustments to test parameters. This increases the risk and cost of satellite testing and may even affect the satellite's ultimate performance and on-orbit operations. Summary of the Invention

[0003] In view of this, the purpose of this application is to provide a satellite test data analysis method and system.

[0004] In conjunction with the first aspect of the present application, a satellite test data analysis method is provided, which is applied to a satellite test data analysis system. The method includes:

[0005] Acquire a raw data set collected during a satellite test, wherein the raw data set includes multiple types of observation data units generated by different test modules in a continuous period of time;

[0006] Performing data feature conversion on the original data set to obtain an analysis feature set for test state analysis, wherein the analysis feature set includes a temporal feature reflecting the operating state of the test modules and a spatial feature reflecting the association relationship between the modules;

[0007] Calling a pre-trained satellite test state assessment model to perform multi-dimensional state assessment processing on the analysis feature set to generate a test state assessment result of the satellite test process, wherein the test state assessment result includes a description of the operation stability of each test module and a description of the parameter correlation between modules;

[0008] Determining, based on the test status assessment result, the type of anomaly present during the satellite test and information about the impact range corresponding to the anomaly type, wherein the impact range information includes the action boundary of the anomaly module and the affected area of ​​the associated modules;

[0009] A test optimization instruction including an abnormality location identifier is generated based on the abnormality type and the impact range information, and the test optimization instruction is fed back to the satellite control terminal to trigger a test parameter adjustment operation.

[0010] In combination with the second aspect of the present application, a satellite test data analysis system is provided, which includes a machine-readable storage medium and a processor, wherein the machine-readable storage medium stores machine-executable instructions. When the processor executes the machine-executable instructions, the satellite test data analysis system implements the aforementioned satellite test data analysis method.

[0011] In conjunction with the third aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer-executable instructions. When the computer-executable instructions are executed, the aforementioned satellite test data analysis method is implemented.

[0012] In combination with any of the above aspects, by acquiring a raw data set containing multiple types of observation data units collected during a satellite test, data feature conversion processing is performed on the raw data set to obtain an analysis feature set containing temporal and spatial features. This feature set can reflect the operating status of the test modules and the inter-module relationships from both temporal and spatial dimensions. A pre-trained satellite test state assessment model is then used to perform a multi-dimensional state assessment on the analysis feature set, generating a test state assessment result that includes a description of the operational stability of each test module and the parameter relationships between modules, thereby achieving a precise assessment of the satellite test process. Based on the test state assessment result, the anomaly type and impact range are determined, enabling the anomaly to be quickly and accurately located and its impact range to be clearly defined. Based on this information, a test optimization instruction containing an anomaly location identifier is generated and fed back to the satellite control terminal, triggering test parameter adjustment operations. This enables timely and effective resolution of problems encountered during satellite testing, improving the success rate and efficiency of satellite testing, and reducing testing risks and costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained by combining these drawings without paying any creative work.

[0014] Figure 1 A flowchart of a satellite test data analysis method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0015] To more clearly illustrate the technical solutions of the embodiments of this specification, the following briefly describes the drawings required for describing the embodiments. Obviously, the drawings described below are merely examples or embodiments of this specification. Those skilled in the art can apply this specification to other similar scenarios based on these drawings without inventive effort. Unless otherwise apparent from the context or otherwise noted, the same reference numerals in the figures represent the same structure or operation.

[0016] It should be understood that the "system", "unit" and / or "module" used herein are a method for distinguishing different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the words can be replaced by other expressions.

[0017] As used in this specification and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not refer to the singular but also include the plural. Generally speaking, the terms "comprises" and "include" only indicate the inclusion of the steps and elements specifically identified, and these steps and elements do not constitute an exclusive list. A method or apparatus may also include other steps or elements.

[0018] Flowcharts are used throughout this specification to illustrate the operations performed by systems according to embodiments of this specification. It should be understood that preceding or following operations do not necessarily need to be performed in exact order. Instead, the steps may be processed in reverse order or simultaneously. Furthermore, other operations may be added to these processes, or one or more operations may be removed from these processes.

[0019] Figure 1 The following is a flow chart of a satellite test data analysis method provided by an embodiment of the present application. It should be understood that in other embodiments, the order of some steps in the satellite test data analysis method of this embodiment can be shared based on actual needs, or some steps can be omitted or maintained. The satellite test data analysis method includes the following details:

[0020] Step S110: obtaining an original data set collected during the satellite test, wherein the original data set includes multiple types of observation data units generated by different test modules in a continuous period.

[0021] During satellite testing, continuous data collection is required from multiple test modules to fully understand the operational status of each satellite component. These modules cover various satellite functional systems, such as attitude control, energy supply, and communication transmission. Each test module is equipped with corresponding sensors and monitoring equipment to collect various relevant physical quantities and operational parameters in real time.

[0022] The attitude control module is primarily responsible for adjusting the satellite's attitude to maintain its predetermined orbit and orientation. To monitor its operational status, it collects data such as angular velocity measured by the gyroscope and acceleration measured by the accelerometer. The energy supply module provides power to the satellite's various systems. Observed data includes the output voltage and current of the solar panels, as well as the charge and discharge status of the battery. The communication transmission module is responsible for transmitting information between the satellite and ground stations or other satellites. Data collected includes signal strength, communication bit error rate, and data transmission rate.

[0023] These different types of observation data units are collected continuously over a continuous period of time. For example, the angular velocity data of the attitude control module may be recorded every certain period of time (such as a few seconds), forming a continuous time series. Similarly, data from the energy supply module and the communication transmission module are also collected at certain time intervals, thus forming a raw data set. The data in this raw data set is diverse and time-series, including multiple types of data generated by different test modules and reflecting the changes in this data over a continuous period of time.

[0024] Step S120: performing data feature conversion processing on the original data set to obtain an analysis feature set for test status analysis, wherein the analysis feature set includes temporal features reflecting the operating status of the test modules and spatial features reflecting the association relationship between modules.

[0025] Since the data in the original data set may have problems such as noise and missing values, and its original form may not be directly used to analyze the status of satellite tests, data feature conversion processing is required.

[0026] Step S121: performing data cleaning on the original data set to remove noise interference information in the data units and fill in missing data segments to generate a cleaned data set.

[0027] In the actual data collection process, due to factors such as sensor accuracy limitations, external environmental interference, and equipment failures, the data units in the raw data set may be contaminated by noise and there may also be data missing. In order to improve the quality of the data, the raw data needs to be cleaned.

[0028] Step S1211: Identify the noise distribution pattern of the data unit in the original data set, and use an adaptive filtering algorithm to perform noise suppression processing on the data unit according to the noise distribution pattern to obtain a denoised data unit.

[0029] Different types of noise have different distribution patterns. Common noise types include Gaussian noise and impulse noise. Gaussian noise is characterized by data errors following a Gaussian distribution, meaning the probability density function of the noise follows a bell-shaped curve. Impulse noise, on the other hand, manifests as sudden, large noise spikes, typically caused by sudden external interference or momentary equipment failure.

[0030] To identify noise distribution patterns, statistical analysis can be used. For example, statistical characteristics such as the mean, variance, skewness, and kurtosis can be calculated to determine the type of noise. If the skewness of the data is close to 0 and the kurtosis is close to 3, Gaussian noise may be present. If the data contains significant outliers with large amplitudes, impulse noise may be present.

[0031] Based on the identified noise distribution pattern, an appropriate adaptive filtering algorithm is selected for noise suppression. For Gaussian noise, a mean filter can be used. Mean filtering works by locally averaging the data, taking the average of all data within a certain range around the data point as the new value for that data point. This effectively smooths the data and reduces the impact of Gaussian noise. For impulse noise, a median filter is more suitable. Median filtering sorts the data within a certain range around the data point and takes the median as the new value for that data point, effectively removing spikes from impulse noise.

[0032] In practical applications, filtering algorithm parameters can be dynamically adjusted based on the specific data. For example, for mean filtering, the local averaging range can be adjusted; for median filtering, the sorted data range can be adjusted. This adaptive approach can better adapt to data with different noise distribution patterns, improve noise suppression, and ultimately produce denoised data units.

[0033] Step S1212: Detect the missing data position in the denoised data unit, and extract the complete data units in the consecutive time periods before and after the missing position as reference data units.

[0034] Even after denoising, missing data may still exist. To detect missing data, we can traverse the entire data sequence and search for data points with empty values ​​or values ​​outside the normal range. For example, if a sensor's data should fall within a specific range, but a value outside that range or no data is recorded, we can consider that data point to be missing.

[0035] Once the positions of the missing data are determined, it is necessary to extract the complete data units in consecutive time periods before and after the missing positions as reference data units. These reference data units contain the normal data information before and after the missing data and can be used for subsequent interpolation filling processing. For example, if a certain data point is missing at time t, then several consecutive data points before and after time t can be extracted as reference data units, and the number of these data points can be determined according to specific circumstances.

[0036] Step S1213: Perform interpolation filling processing on the missing data positions based on the data change trend of the reference data units to generate a cleaned data set containing a complete data sequence.

[0037] After obtaining the reference data units, interpolation filling can be performed on the missing data positions according to their data change trends. Common interpolation methods include linear interpolation, polynomial interpolation, etc.

[0038] Linear interpolation is a simple and commonly used interpolation method, which assumes that the data changes linearly between two adjacent data points. If the missing data point is between two known data points, then the value of the missing data point can be calculated through a linear relationship based on the values and time intervals of these two known data points. For example, if the values corresponding to time t1 and t2 are y1 and y2 respectively, the time of the missing data point is t, and t1 < t < t2, then the value y of the missing data point can be calculated through the linear interpolation formula:

[0039] y = y1 + (y2 - y1) * (t - t1) / (t2 - t1)

[0040] Polynomial interpolation can handle more complex data change trends. It approximates the data change of the reference data units by fitting a polynomial function and then uses this polynomial function to calculate the values of the missing data points. The order of the polynomial can be selected according to the number of reference data units and the complexity of the data.

[0041] Through the interpolation filling process, the missing data points are supplemented to generate a cleaned data set containing a complete data sequence. The data in this cleaned data set has undergone noise suppression and missing value filling, and the quality has been significantly improved.

[0042] Step S122: Perform feature extraction processing on the cleaned data set, extract the change rule features of each test module in consecutive time periods as time series features, and extract the association pattern features between the data units of different test modules as spatial features.

[0043] Although the cleaned data set has been noise-removed and missing values ​​have been supplemented, it is still raw observational data and cannot directly reflect the status of the satellite test. Therefore, it is necessary to extract useful features from this data, including temporal features that reflect the operating status of the test modules and spatial features that reflect the correlation between modules.

[0044] Step S1221: for each cleaned data unit of the test module, calculate the variation of the data values ​​in adjacent time periods and count the consistency of the variation direction in the continuous time periods, and generate variation regularity characteristics reflecting the module operation stability as the time series characteristics.

[0045] For each cleaned data unit of the test module, we first calculate the variation between adjacent time periods. Taking the angular velocity data of the attitude control module as an example, assuming the angular velocities collected at time t and time t+1 are ω(t) and ω(t+1), respectively, the variation between adjacent time periods is Δω(t) = ω(t+1) - ω(t). By calculating the variation between each adjacent time period, we can obtain a variation sequence.

[0046] Next, we count the consistency of the direction of change within consecutive time periods. This consistency reflects whether the data is continuously increasing, decreasing, or fluctuating over a period of time. This can be done by defining a change direction indicator. If Δω(t) > 0, the direction of change is positive; if Δω(t) < 0, the direction of change is negative; and if Δω(t) = 0, the data has not changed.

[0047] By counting the number of times the change direction remains the same within consecutive time periods, we can obtain a statistic that measures the consistency of the change direction. For example, if the change direction is consistently positive within a continuous time period, this indicates that the data for the test module during this period has shown a continuous upward trend, possibly indicating that the module's operating status is relatively stable. Conversely, if the change direction changes frequently, this indicates that the module's operating status may be fluctuating.

[0048] By comprehensively considering the consistency of the change amount and change direction of data values ​​in adjacent time periods, we can generate change regularity characteristics that reflect the operational stability of the module. These change regularity characteristics constitute time series characteristics. Time series characteristics can reflect the changes in the operating status of the test module in continuous time periods.

[0049] Step S1222: for the cleaned data units of any two test modules, analyze the synchronous change degree and asynchronous change frequency of the data values ​​in the same time period, and generate correlation pattern features reflecting the collaborative relationship between the modules as the spatial features.

[0050] To analyze the correlation between different test modules, we need to study the cleaned data units of any two test modules. The degree of synchronous change of data values ​​within the same time period reflects whether the data changes of the two test modules at the same time point are consistent. The asynchronous change frequency indicates the degree of temporal difference in the data changes of the two test modules.

[0051] The correlation coefficient method can be used to calculate the degree of synchronous change in data values ​​within the same time period. The correlation coefficient is a value between -1 and 1 that measures the degree of linear correlation between two variables. For the cleaned data units of two experimental modules, assuming they are x(t) and y(t), where t represents time, the degree of synchronous change can be assessed by calculating their correlation coefficient ρ(x, y). The formula for calculating the correlation coefficient is:

[0052] ρ(x,y)=cov(x,y) / (σ(x)*σ(y))

[0053] Where cov(x, y) is the covariance of x and y, and σ(x) and σ(y) are the standard deviations of x and y, respectively. If the correlation coefficient is close to 1, it indicates that the data of the two experimental modules have a strong synchronous trend in the same period; if the correlation coefficient is close to -1, it indicates that their changing trends are opposite; if the correlation coefficient is close to 0, it indicates that there is no obvious synchronization relationship between the data of the two experimental modules.

[0054] The frequency of asynchronous changes can be calculated by counting the time differences between data changes in two experimental modules. For example, record the time points at which x(t) and y(t) change significantly and calculate the difference between these time points. If the time difference is small, it indicates that the data changes in the two experimental modules are relatively synchronized; if the time difference is large, it indicates that their changes are asynchronous.

[0055] By analyzing the degree of synchronous changes and the frequency of asynchronous changes in data values ​​within the same time period, we can generate correlation pattern features that reflect the collaborative relationships between modules. These features constitute spatial features. Spatial features can reveal the interrelationships between different test modules and are important for understanding the overall operational status of the satellite system.

[0056] Step S123: performing feature association processing on the temporal features and the spatial features, establishing a mapping relationship between features of different test modules in the same time period, and generating an associated feature set with temporal-spatial coupling characteristics as the analysis feature set.

[0057] Temporal and spatial features reflect the operational status of satellite test modules and the inter-module relationships from different perspectives, but they are independent of each other. To more comprehensively analyze the status of satellite tests, it is necessary to correlate these features and establish a mapping relationship between them.

[0058] Step S1231: Mark a time identifier for each time period, and perform time dimension alignment processing on the temporal features of each test module in the time period and the corresponding spatial features to obtain a feature alignment set with consistent time identifiers.

[0059] Before performing feature association processing, each time period must be marked with a timestamp. This timestamp can be a specific timestamp or a relative time sequence number. By assigning a unique timestamp to each time period, the corresponding time position of each data point and feature can be clearly identified.

[0060] Then, the temporal features of each experimental module within each time period are aligned with the corresponding spatial features. For example, at a specific time point t, the temporal features of the attitude control module (such as the change in angular velocity and the consistency of the change direction) are matched with the spatial features (such as the correlation coefficient and the frequency of asynchronous changes) of it and other experimental modules (such as the energy supply module and the communication transmission module). This ensures that all features correspond to the same time point, thus obtaining a feature alignment set with consistent time signatures.

[0061] Temporal dimension alignment is the basis of feature association processing, which enables different types of features to be compared and analyzed on the same time scale.

[0062] Step S1232: performing a cross-correlation analysis on the temporal features and spatial features in the feature alignment set, extracting the influence weight of the temporal feature change on the spatial feature and the feedback strength of the spatial feature change on the temporal feature.

[0063] After obtaining a feature alignment set with consistent time signatures, we need to perform cross-correlation analysis on the temporal and spatial features. The purpose of cross-correlation analysis is to study the mutual influence between temporal and spatial features, that is, how changes in temporal features affect spatial features, and how changes in spatial features feed back into temporal features.

[0064] Regression analysis can be used to extract the weight of the impact of changes in temporal features on spatial features. Taking the temporal features of the attitude control module (such as the change in angular velocity) and the spatial features of the communication transmission module (such as the correlation coefficient of signal strength) as an example, a regression model is established, using the temporal features as the independent variable and the spatial features as the dependent variable. By fitting a large amount of data, a regression coefficient can be obtained, which can be used as the weight of the impact of changes in temporal features on spatial features.

[0065] Similarly, to extract the feedback strength of spatial feature changes on temporal features, a similar regression analysis method can be used. Using spatial features as independent variables and temporal features as dependent variables, a regression model is established and the regression coefficient is calculated. This regression coefficient reflects the feedback strength of spatial feature changes on temporal features.

[0066] Through cross-correlation analysis, the mutual influence relationship between temporal features and spatial features can be obtained, and these relationships are quantified by influence weights and feedback strengths.

[0067] Step S1233: constructing a feature association matrix based on the influence weight and feedback strength, and taking the feature combination with significant association relationship in the feature association matrix as the association feature set.

[0068] Based on the weight of the influence of the extracted temporal feature changes on the spatial features and the feedback strength of the spatial feature changes on the temporal features, a feature correlation matrix can be constructed. The feature correlation matrix is ​​a two-dimensional matrix in which the elements represent the degree of correlation between different features.

[0069] The rows and columns of the feature correlation matrix correspond to different features. For example, rows represent temporal features, and columns represent spatial features. The value of each element in the feature correlation matrix represents the corresponding influence weight or feedback strength. By analyzing the elements in the feature correlation matrix, we can determine whether the correlation between different features is significant.

[0070] A threshold can be set. When the value of an element in the feature correlation matrix exceeds this threshold, the corresponding features are considered to have a significant correlation. Significantly correlated features are then combined and extracted to form a correlation feature set. This correlation feature set exhibits temporal-spatial coupling, integrating information from both temporal and spatial features. This allows for a more comprehensive reflection of the satellite test status and can therefore be used as an analytical feature set for test status analysis.

[0071] Step S130: Calling a pre-trained satellite test state assessment model to perform multi-dimensional state assessment processing on the analysis feature set to generate a test state assessment result of the satellite test process, wherein the test state assessment result includes a description of the operation stability of each test module and a description of the parameter correlation between modules.

[0072] The pre-trained satellite test status assessment model is an artificial intelligence model trained with a large amount of data. It can perform multi-dimensional status assessment on the analysis feature set, thereby generating detailed assessment results about the satellite test process.

[0073] Step S131: inputting the analysis feature set into the state stability assessment module of the satellite test state assessment model, analyzing the fluctuation range and fluctuation frequency of the timing characteristics of each test module, and generating a state stability description reflecting the module operation reliability.

[0074] The state stability assessment module mainly focuses on the timing characteristics of each test module, and evaluates the operational reliability of the module by analyzing the fluctuation range and frequency of these characteristics.

[0075] Step S1311: extracting the time series features of each test module in the analysis feature set, and counting the maximum value, minimum value and average value of each time series feature within a preset time window.

[0076] Extract time series features for each test module from the analysis feature set. These time series features reflect the changes in the module's operating status over a continuous period. To analyze fluctuations in these features, a preset time window is required. This preset time window can be determined based on specific analysis requirements and data characteristics. For example, a shorter time window can be selected to capture short-term fluctuations, while a longer time window can be selected to observe long-term trends.

[0077] Within a preset time window, the maximum, minimum, and average values ​​of each time series feature are calculated. For example, for the attitude control module's angular velocity variation, within a specific time window, the maximum and minimum values ​​of all angular velocity variations are found, and their average is calculated. These statistics can reflect the approximate range and average level of the test module's operating status within that time window.

[0078] Step S1312: Calculate the difference between the maximum value and the minimum value as a fluctuation range parameter, and count the number of times the fluctuation range parameter exceeds a preset threshold as a fluctuation frequency parameter.

[0079] Based on the statistically obtained maximum and minimum values, the difference between them is calculated. This difference is the fluctuation range parameter. The fluctuation range parameter reflects the magnitude of the variation of the timing characteristics within a preset time window. For example, if the maximum angular velocity variation of the attitude control module within a time window is 10 and the minimum is 2, the fluctuation range parameter is 10-2 = 8.

[0080] The preset threshold is a pre-set standard value used to determine whether the fluctuation range is excessive. By counting the number of times the fluctuation range parameter exceeds the preset threshold, we can obtain the fluctuation frequency parameter. The fluctuation frequency parameter reflects the frequency of fluctuations in the time series characteristics over a period of time. If the fluctuation frequency parameter is high, it indicates that the operating status of the test module is unstable and there may be a high fluctuation risk.

[0081] Step S1313: constructing a module operation reliability evaluation function based on the fluctuation range parameter and the fluctuation frequency parameter, and generating a state stability description including a reliability level and a fluctuation risk point through the module operation reliability evaluation function.

[0082] To comprehensively evaluate the operational reliability of the test module, a module operational reliability evaluation function needs to be constructed based on the fluctuation range parameter and the fluctuation frequency parameter. The evaluation function can be a complex mathematical model that takes the fluctuation range parameter and the fluctuation frequency parameter as input and outputs a numerical value representing the module operational reliability.

[0083] For example, the evaluation function can use a weighted summation approach, multiplying the fluctuation range parameter and the fluctuation frequency parameter by different weights and then adding them together to obtain a comprehensive evaluation value. The choice of weights can be adjusted according to actual conditions to reflect the different degrees of influence of fluctuation range and fluctuation frequency on module operational reliability.

[0084] Based on the output of the evaluation function, the module's operational reliability can be classified into different reliability levels, such as high reliability, medium reliability, and low reliability. Furthermore, by analyzing the specific time points when the fluctuation range parameters exceed the preset threshold, fluctuation risk points can be identified. Fluctuation risk points indicate the time points or data features with the greatest fluctuation risk, which is crucial for timely identification and resolution of issues.

[0085] The state stability description generated by the evaluation function includes the reliability level and fluctuation risk points. This information can help operators quickly understand the operational reliability of each test module.

[0086] Step S132: Input the analysis feature set into the parameter correlation evaluation module of the satellite test state evaluation model, analyze the correlation degree and correlation change rate of the spatial features of different test modules, and generate a parameter correlation description reflecting the module collaborative effectiveness.

[0087] After the analysis feature set is input into the parameter correlation assessment module of the satellite test state assessment model, the module will conduct an in-depth analysis of the spatial characteristics of different test modules, focusing on the degree of correlation and the rate of correlation change, in order to generate a parameter correlation description reflecting the effectiveness of module collaboration.

[0088] Spatial characteristics reflect the correlation between different test modules, and the closeness of this correlation is an important indicator for measuring the strength of this relationship. The spatial characteristics of any two test modules require analysis from multiple perspectives. For example, the spatial characteristics of the communication module and the power module may involve the correlation between communication signal strength and power system thrust. The closeness of this correlation can be assessed by calculating a correlation index between the two. Correlation indices can be calculated using various methods, such as the Pearson correlation coefficient. When calculating the Pearson correlation coefficient, the covariance of the two variables and their respective standard deviations can be considered. Covariance reflects whether the changing trends of the two variables are consistent. A positive covariance indicates that the two variables have the same changing trends; a negative covariance indicates that the changing trends are opposite. The standard deviation measures the degree of dispersion of the variables. By comprehensively considering the covariance and standard deviation, a correlation coefficient between -1 and 1 is obtained. The closer the coefficient is to 1 or -1, the closer the correlation between the spatial characteristics of the two test modules is; the closer it is to 0, the weaker the correlation is.

[0089] In addition to correlation metrics, other factors can be considered to assess the closeness of association. For example, we can observe the synchronization of the spatial characteristics of two experimental modules over time. If the spatial characteristics of the two experimental modules change simultaneously at most time points, then we can assume that their closeness of association is high. We can also analyze the causal relationship between spatial characteristics, that is, whether changes in the spatial characteristics of one experimental module will lead to corresponding changes in the spatial characteristics of another experimental module. By comprehensively analyzing these factors, we can obtain a comprehensive assessment of the closeness of association.

[0090] The correlation change rate reflects how the correlation between the spatial features of different test modules changes over time. To analyze the correlation change rate, first record the correlation strength at different time points. You can calculate and record the correlation strength between the spatial features of every two test modules at regular intervals. For example, calculate the correlation strength between the spatial features of the communication module and the power module every hour.

[0091] Then, by comparing the degree of correlation between adjacent time points, the change in the degree of correlation can be obtained. For example, at time t1 and t2, the degree of correlation of the spatial features of the communication module and the power module is calculated as C1 and C2, respectively, and the change in the degree of correlation is C2-C1. Dividing this change by the time interval (t2-t1) yields the correlation change rate. The correlation change rate can reflect whether the collaborative relationship between different test modules is stable. If the correlation change rate is large, it means that the correlation relationship between the test modules has changed significantly in a short period of time, which may indicate that the collaborative work between the modules is unstable; if the correlation change rate is small, it means that the correlation relationship is relatively stable.

[0092] When analyzing the correlation change rate, you can also consider the trend of the change. For example, is the correlation change rate continuously increasing, continuously decreasing, or fluctuating? If the correlation change rate is continuously increasing, it may indicate that the coordination relationship between modules is gradually deteriorating; if it is continuously decreasing, it may indicate that the coordination relationship is gradually improving. By analyzing the correlation change rate and its trend, you can gain a deeper understanding of the collaborative work between different test modules.

[0093] Based on the analysis of the correlation strength and correlation change rate of the spatial features of different experimental modules, a parameter correlation description is generated to reflect the effectiveness of module collaboration. This parameter correlation description takes both correlation strength and correlation change rate into consideration. For example, if the spatial features of two experimental modules have a high correlation strength and a low correlation change rate, this indicates that the collaboration between the two experimental modules is effective, which can be reflected in the parameter correlation description as "close and stable correlation, high collaborative effectiveness."

[0094] If the degree of correlation is low and the rate of change is high, there may be problems with the collaborative work between modules. This can be reflected in the parameter correlation description as "weak and unstable correlation, low collaborative effectiveness." The parameter correlation description can also specifically indicate the correlation between the test modules and the time points or time periods when significant correlation changes occurred.

[0095] Step S133: Perform comprehensive evaluation processing on the state stability description and the parameter correlation description to generate a test state evaluation result including a module-level evaluation conclusion and a system-level evaluation conclusion.

[0096] After obtaining the state stability description and parameter correlation description, these two descriptions need to be comprehensively evaluated to generate a more comprehensive test state evaluation result. The test state evaluation result includes module-level evaluation conclusions and system-level evaluation conclusions.

[0097] The state stability description reflects the operational reliability of each test module, while the parameter correlation description reflects the collaborative effectiveness between different test modules. These two aspects of information need to be combined in a comprehensive assessment. For example, if the state stability description of a test module indicates high operational reliability, but the parameter correlation descriptions with other test modules indicate low collaborative effectiveness, the comprehensive assessment needs to consider the impact of this situation on the overall satellite test status.

[0098] The state stability description and parameter correlation description can be combined using a weighted approach. Different weights are assigned to the state stability description and the parameter correlation description, each weighted based on their importance to the satellite test status. For example, if module operational reliability is considered more important to the overall satellite test status, a higher weight can be assigned to the state stability description; if inter-module collaboration is considered to have a greater impact on the satellite test status, a higher weight can be assigned to the parameter correlation description. A comprehensive evaluation value is then calculated based on the respective evaluation results and weights to obtain the final result.

[0099] Based on the results of the comprehensive evaluation, a module-level evaluation conclusion is generated. This module-level evaluation conclusion specifically assesses each test module. For each test module, a comprehensive description of its state stability and parameter correlations with other modules can be considered. For example, for the communications module, the evaluation conclusion may indicate its operational reliability, its effectiveness in synergy with other modules such as the power module and the detection module, and whether there are any potential issues that could affect the overall satellite test status. The module-level evaluation conclusion provides a basis for adjusting and optimizing individual test modules.

[0100] When generating module-level evaluation conclusions, test modules can also be categorized. For example, test modules with high operational reliability and high collaborative effectiveness can be grouped into one category, while those with low operational reliability or low collaborative effectiveness can be grouped into another. This categorization provides a clearer picture of the status of each test module, facilitating subsequent management and processing.

[0101] In addition to module-level assessment conclusions, system-level assessment conclusions are also required. These conclusions are based on the perspective of the entire satellite test system, comprehensively considering the stability of all test modules and the inter-module parameter correlations. The system-level assessment provides an overall assessment of the operational status of the entire satellite test system, including system stability and potential risks.

[0102] When generating system-level evaluation conclusions, the interplay between test modules can be considered. If the collaborative effectiveness of multiple test modules is low, this could significantly impact the performance of the entire satellite test system, and this fact should be clearly noted in the system-level evaluation conclusions. Furthermore, the system-level evaluation conclusions can also provide general improvement recommendations, such as whether certain test modules need to be adjusted or whether the collaborative relationships between modules need to be optimized. By combining module-level and system-level evaluation conclusions, a comprehensive and in-depth understanding of the status of satellite testing can be achieved.

[0103] Step S140: determining the abnormality type existing in the satellite test process and the impact range information corresponding to the abnormality type according to the test status evaluation result, wherein the impact range information includes the action boundary of the abnormal module and the affected area of ​​the associated module.

[0104] Step S141: parsing the module-level evaluation conclusion and the system-level evaluation conclusion in the test status evaluation result, and identifying abnormal modules that deviate from the normal state in the evaluation conclusion.

[0105] The module-level and system-level evaluation conclusions in the test state evaluation results contain a wealth of information. When analyzing these conclusions, it is necessary to focus on identifying abnormal modules that deviate from the normal state. For the module-level evaluation conclusions, the operational reliability of each test module and its collaborative effectiveness with other modules can be described in detail. If the operational reliability of a test module is low, such as if the state stability description shows a large number of fluctuation risk points, or if the collaborative effectiveness with other modules is poor, such as if the parameter correlation description shows a weak and unstable correlation, then it can be considered that the test module may be an abnormal module.

[0106] The system-level evaluation conclusions allow for a holistic evaluation of the satellite test system. If the system-level evaluation concludes that problems with certain test modules significantly impact the performance of the entire system, these modules may also be considered abnormal. A comprehensive analysis of the module-level and system-level evaluation conclusions allows for accurate identification of abnormal modules.

[0107] Step S142: According to the state stability description and parameter correlation description of the abnormal module, determine the abnormal manifestation and match the preset abnormal type classification rules to generate the abnormal type.

[0108] After identifying an abnormal module, it is necessary to determine the abnormal manifestation based on its state stability description and parameter correlation description. Abnormal manifestations may include unstable module operation and poor coordination with other modules. For example, if the state stability description of the abnormal module shows that both the fluctuation range parameter and the fluctuation frequency parameter are high, it means that the module is unstable and the abnormal manifestation may be frequent fluctuations. If the parameter correlation description shows that the correlation with other modules is low and the correlation change rate is high, it means that the module is poorly coordinated with other modules and the abnormal manifestation may be dyssynergy.

[0109] The preset exception classification rules are derived from extensive historical data and experience. These rules map different exception manifestations to different exception types. For example, unstable module operation might be assigned an "operational failure" exception type, while poor module coordination might be assigned a "coordination failure" exception type. By matching the exception manifestation with the preset exception classification rules, an exception type is generated.

[0110] Step S143: extracting information of other modules associated with the spatial features of the abnormal module, analyzing the parameter correlation description change trend of the associated modules, determining the impact degree and impact range boundary of the abnormal module on the associated modules, and generating the impact range information.

[0111] The spatial characteristics of an abnormal module contain information about its associations with other modules. By extracting this association information, we can identify the other modules associated with the abnormal module. We can then analyze the parameter correlations of these associated modules to describe their changing trends. For example, we can observe how the closeness of the association between the associated module and the abnormal module, as well as the rate of change in the association, changes before and after the abnormal module experiences a problem.

[0112] If the closeness of the association between the associated module and the abnormal module decreases significantly after the abnormal module encounters a problem, and the rate of change in the association increases, it indicates that the abnormal module has a significant impact on the associated module. By analyzing the trend of changes in the parameter correlation description of the associated module, the degree of influence of the abnormal module on the associated module can be determined.

[0113] The impact range boundary refers to the scope of the abnormal module's impact. The impact range boundary can be determined based on changes in the parameter association descriptions of associated modules. For example, if the parameter association description of a certain associated module changes significantly after a problem with the abnormal module occurs, while the parameter association description of another associated module remains unaffected, the scope of the affected associated module can be determined as the impact range boundary. By determining the impact level and impact range boundary, impact range information is generated. The impact range information includes the action boundary of the abnormal module and the affected area of ​​the associated modules.

[0114] Step S150: generating a test optimization instruction including an abnormality location identifier based on the abnormality type and the impact range information, and feeding back the test optimization instruction to the satellite control terminal to trigger a test parameter adjustment operation.

[0115] After determining the anomaly type and impact range, test optimization instructions need to be generated to resolve the problem during the satellite test. The test optimization instructions contain an anomaly location identifier to accurately indicate the location and scope of the problem.

[0116] Step S151: matching a preset optimization strategy library according to the abnormality type, and extracting parameter adjustment direction and adjustment range suggestions corresponding to the abnormality type.

[0117] The preset optimization strategy library is built based on different exception types and extensive historical experience. When generating test optimization instructions, you need to match the corresponding optimization strategy from the optimization strategy library based on the exception type. For example, if the exception type is "operational failure," the optimization strategy library may provide parameter adjustment directions and adjustment ranges to address the operational failure, such as adjusting the operating parameters of the test module to improve operational stability.

[0118] The optimization strategy library recommends different parameter adjustment directions and adjustment ranges for different anomaly types. Parameter adjustment directions can include increasing or decreasing certain parameter values, while the recommended adjustment range specifies the degree of adjustment. By matching the optimization strategy library, you can accurately obtain the parameter adjustment direction and adjustment range recommendations corresponding to the anomaly type.

[0119] Step S152: extracting identification information of the abnormal module and its associated modules as an abnormality location identifier according to the impact range boundary in the impact range information.

[0120] The impact range boundary in the impact range information defines the scope of the abnormal module and its associated modules. Based on this scope, identification information of the abnormal module and its associated modules, such as module numbers and names, is extracted as the abnormality location identifier. The abnormality location identifier accurately points out the problem location, facilitating subsequent test parameter adjustments.

[0121] Step S153: performing information fusion processing on the parameter adjustment direction, adjustment range suggestion and the abnormal location identifier to generate a test optimization instruction including the module identifier, adjustment direction and adjustment range.

[0122] The parameter adjustment direction and adjustment range recommendations obtained from the optimization strategy library are combined with the anomaly location indicators for information fusion processing. This information fusion process integrates this information into a complete test optimization instruction. The test optimization instruction clearly indicates which test modules require parameter adjustment, the direction of adjustment, and the adjustment range.

[0123] For example, a test optimization instruction could be expressed as, "For the communication module (number 123), increase operating parameter A by a certain percentage; for the power module (number 456), decrease operating parameter B by a certain percentage." These test optimization instructions accurately guide the satellite control terminal to perform test parameter adjustments.

[0124] Step S154: Feedback the test optimization instruction to the satellite control terminal to trigger the test parameter adjustment operation.

[0125] The generated test optimization instructions are fed back to the satellite control terminal. Upon receiving the instructions, the satellite control terminal can adjust the parameters of the corresponding test modules based on the information contained in the instructions. By adjusting the test parameters, problems encountered during the satellite test can be resolved, improving the performance and reliability of the satellite test. Furthermore, after adjusting the test parameters, the satellite test data can be retrieved again and the aforementioned data analysis and evaluation process can be repeated to verify the effectiveness of the adjustments and ensure the smooth progress of the satellite test.

[0126] One or more embodiments of this specification provide a computer program product, which includes a computer program. When the computer program is executed by a processor, the satellite test data analysis method of the above embodiment is implemented.

[0127] While the basic concepts have been described above, it will be apparent to those skilled in the art that the detailed disclosure is merely illustrative and does not limit this specification. Although not explicitly stated herein, various modifications, improvements, and revisions to this specification may be made by those skilled in the art. Such modifications, improvements, and revisions are suggested in this specification and remain within the spirit and scope of the exemplary embodiments of this specification.

[0128] This specification also uses specific terms to describe the embodiments of this specification. For example, "one embodiment," "an embodiment," and / or "some embodiments" refer to a feature, structure, or characteristic associated with at least one embodiment of this specification. Therefore, it should be emphasized and noted that references to "one embodiment," "an embodiment," or "an alternative embodiment" two or more times in different locations in this specification do not necessarily refer to the same embodiment. Furthermore, certain features, structures, or characteristics of one or more embodiments of this specification may be appropriately combined.

[0129] In addition, unless expressly stated in the claims, the order of the processing elements and sequences, the use of alphanumeric characters, or the use of other names described in this specification are not intended to limit the order of the processes and layer flows of this specification. Although the above disclosure discusses some of the invention embodiments that are currently considered useful through various examples, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that are consistent with the spirit and scope of the embodiments of this specification. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only by software solutions, such as installing the described system on an existing server or mobile device.

[0130] Similarly, it should be noted that, in order to simplify the presentation of this specification and thus facilitate understanding of one or more embodiments of the invention, the foregoing descriptions of the embodiments of this specification sometimes combine multiple features into a single embodiment, figure, or description thereof. However, this disclosure method does not imply that the subject matter of this specification requires more features than those recited in the claims. In fact, an embodiment may have fewer features than all of the features of a single disclosed embodiment.

[0131] Finally, it should be understood that the embodiments described in this specification are intended only to illustrate the principles of the embodiments of this specification. Other variations are also possible and fall within the scope of this specification. Therefore, by way of example and not limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly described and illustrated in this specification.

Claims

1. A satellite test data analysis method, characterized in that: The method comprises: Acquire a raw data set collected during a satellite test, wherein the raw data set includes multiple types of observation data units generated by different test modules in a continuous period of time; Performing data feature conversion on the original data set to obtain an analysis feature set for test state analysis, wherein the analysis feature set includes a temporal feature reflecting the operating state of the test modules and a spatial feature reflecting the association relationship between the modules; Calling a pre-trained satellite test state assessment model to perform multi-dimensional state assessment processing on the analysis feature set to generate a test state assessment result of the satellite test process, wherein the test state assessment result includes a description of the operation stability of each test module and a description of the parameter correlation between modules; Determining, based on the test status assessment result, the type of anomaly present during the satellite test and information about the impact range corresponding to the anomaly type, wherein the impact range information includes the action boundary of the anomaly module and the affected area of ​​the associated modules; A test optimization instruction including an abnormality location identifier is generated based on the abnormality type and the impact range information, and the test optimization instruction is fed back to the satellite control terminal to trigger a test parameter adjustment operation.

2. The satellite test data analysis method according to claim 1, characterized in that: The performing of data feature conversion processing on the original data set to obtain an analysis feature set for test state analysis includes: Performing data cleaning on the original data set to remove noise interference information in the data units and fill in missing data segments to generate a cleaned data set; Performing feature extraction processing on the cleaned data set, extracting the change regularity characteristics of each test module in a continuous period as a temporal feature, and extracting the correlation pattern characteristics between data units of different test modules as a spatial feature; Feature association processing is performed on the temporal features and the spatial features, a mapping relationship between features of different test modules in the same time period is established, and a correlation feature set with temporal-spatial coupling characteristics is generated as the analysis feature set.

3. The satellite test data analysis method according to claim 2, characterized in that: The performing data cleaning on the original data set to remove noise interference information in the data units and to complete missing data segments to generate a cleaned data set includes: Identifying a noise distribution pattern of a data unit in the original data set, and performing noise suppression processing on the data unit using an adaptive filtering algorithm according to the noise distribution pattern to obtain a denoised data unit; Detecting missing data positions in the denoised data units, and extracting complete data units in consecutive time periods before and after the missing positions as reference data units; The missing data positions are interpolated and filled based on the data change trend of the reference data units to generate a cleaned data set containing a complete data sequence.

4. The satellite test data analysis method according to claim 2, characterized in that: The feature extraction process is performed on the cleaned data set to extract the change regularity characteristics of each test module in a continuous period as a temporal feature, and to extract the correlation pattern characteristics between data units of different test modules as a spatial feature, including: For each cleaned data unit of the test module, the change amount of the data value in adjacent time periods is calculated and the consistency of the change direction in the continuous time periods is counted to generate the change regularity characteristics reflecting the module operation stability as the said time series characteristics; For the cleaned data units of any two test modules, the synchronous change degree and asynchronous change frequency of the data values ​​in the same period are analyzed, and the correlation pattern features reflecting the collaborative relationship between the modules are generated as the spatial features.

5. The satellite test data analysis method according to claim 2, characterized in that: The performing feature association processing on the time series features and the spatial features, establishing a mapping relationship between features of different test modules in the same time period, and generating a set of associated features with time series-space coupling characteristics as the analysis feature set includes: Mark each time period with a time identifier, and perform time dimension alignment processing on the temporal features of each test module in the time period with the corresponding spatial features to obtain a feature alignment set with consistent time identifiers; Performing cross-correlation analysis on the temporal features and spatial features in the feature alignment set to extract the influence weight of the temporal feature change on the spatial feature and the feedback strength of the spatial feature change on the temporal feature; A feature association matrix is ​​constructed based on the influence weight and feedback strength, and a feature combination with a significant association relationship in the feature association matrix is ​​used as the association feature set.

6. The satellite test data analysis method according to claim 1, characterized in that: The calling of a pre-trained satellite test state assessment model to perform multi-dimensional state assessment processing on the analysis feature set to generate a test state assessment result of the satellite test process includes: Inputting the analysis feature set into the state stability assessment module of the satellite test state assessment model, analyzing the fluctuation range and fluctuation frequency of the timing characteristics of each test module, and generating a state stability description reflecting the module operation reliability; Inputting the analysis feature set into the parameter correlation evaluation module of the satellite test state evaluation model, analyzing the correlation tightness and correlation change rate of the spatial features of different test modules, and generating a parameter correlation description reflecting the module collaborative effectiveness; A comprehensive evaluation process is performed on the state stability description and the parameter correlation description to generate a test state evaluation result including a module-level evaluation conclusion and a system-level evaluation conclusion.

7. The satellite test data analysis method according to claim 6, characterized in that: Inputting the analysis feature set into the state stability assessment module of the satellite test state assessment model, analyzing the fluctuation range and fluctuation frequency of the timing characteristics of each test module, and generating a state stability description reflecting the operational reliability of the module, includes: Extracting the time series features of each test module in the analysis feature set, and counting the maximum value, minimum value and average value of each time series feature within a preset time window; Calculating the difference between the maximum value and the minimum value as a fluctuation range parameter, and counting the number of times the fluctuation range parameter exceeds a preset threshold as a fluctuation frequency parameter; A module operation reliability evaluation function is constructed based on the fluctuation range parameter and the fluctuation frequency parameter, and a state stability description including a reliability level and a fluctuation risk point is generated by running the module reliability evaluation function.

8. The satellite test data analysis method according to claim 1, characterized in that: The determining, according to the test status evaluation result, of the abnormality type existing in the satellite test process and the impact range information corresponding to the abnormality type, includes: Analyze the module-level evaluation conclusions and system-level evaluation conclusions in the test status evaluation results, and identify abnormal modules that deviate from the normal state in the evaluation conclusions; According to the state stability description and parameter correlation description of the abnormal module, determine the abnormal manifestation and match the preset abnormal type classification rules to generate the abnormal type; Extract the information of other modules associated with the spatial features of the abnormal module, analyze the parameter correlation description change trend of the associated modules, determine the impact degree and impact range boundary of the abnormal module on the associated modules, and generate the impact range information.

9. The satellite test data analysis method according to claim 1, characterized in that: The generating of the test optimization instruction including the abnormality location identifier based on the abnormality type and the impact range information includes: According to the abnormality type, a preset optimization strategy library is matched, and the parameter adjustment direction and adjustment range suggestions corresponding to the abnormality type are extracted; Extracting identification information of the abnormal module and its associated modules as an abnormal location identifier according to the impact range boundary in the impact range information; The parameter adjustment direction, adjustment range suggestion and the abnormal location identifier are subjected to information fusion processing to generate a test optimization instruction including a module identifier, an adjustment direction and an adjustment range.

10. A satellite test data analysis system, characterized in that: The invention comprises a processor and a computer-readable storage medium, wherein the computer-readable storage medium stores machine-executable instructions, and when the machine-executable instructions are executed by a computer, the satellite test data analysis method according to any one of claims 1 to 9 is implemented.

Images