Method, device and equipment for detecting credibility of state parameters of cockpit display and control system of airplane and medium
By training and generating a model to reconstruct the standard distribution characteristics of the state parameters of the display and control system, calculating the deviation and dynamically updating it, the problem of decreased detection accuracy in existing technologies is solved, and accurate and reliable detection is achieved under different flight stages.
Patent Information
- Application Number
- CN202611593160.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-15
- Publication Date
- 2026-11-10
AI Technical Summary
In the existing technology, the detection method of the status parameters of the aircraft cockpit display and control system uses fixed threshold judgment, which cannot adapt to the dynamic changes in different flight stages, resulting in decreased detection accuracy and easy misjudgment or omission of the reliability of status parameters.
By training and generating a model, real-time collection of state parameter time series is performed, reconstructing standard distribution characteristics that conform to the current flight phase, calculating deviation and dynamically updating the model to adapt to changes in flight phase and improve detection accuracy.
It enables the reliability detection of state parameters under different flight phases, reduces false judgments, improves the accuracy and reliability of detection, and adapts to the dynamic changes in flight phases.
Smart Images

Figure CN123067431A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of civil aircraft avionics systems and flight safety technology, and in particular to a method, apparatus, equipment and medium for detecting the reliability of status parameters of an aircraft cockpit display and control system. Background Technology
[0002] With the development of integrated avionics systems in modern civil aircraft, cockpit display and control systems need to collect, process, and display large amounts of flight data in real time to provide data support for status monitoring and decision support during flight. In this process, the reliability of the display and control system itself directly affects the quality of its data acquisition, transmission, processing, and display.
[0003] To assess the operational status of a display and control system, it is typically necessary to monitor parameters reflecting its health, such as data transmission capability parameters, data stability parameters, and computing resource status parameters. Data transmission capability parameters include data transmission bandwidth, data stability parameters include latency jitter and packet loss rate, and computing resource status parameters include processor (CPU) utilization and memory utilization. These status parameters describe the operational status of the display and control system itself, which differs from the flight data it processes.
[0004] In actual operation, the authenticity, accuracy, and reliability of the aforementioned status parameters directly affect the reliability of the judgment on the operating status of the display and control system. For example, if there is a deviation in the data transmission bandwidth data, it may lead to a distortion in the assessment of the display and control system's transmission capability. Since these status parameters are the basis for monitoring the status and diagnosing faults in the display and control system, if the parameters themselves are unreliable, the health assessment and auxiliary decisions made based on these status parameters will lose their reliability, thereby affecting flight safety. Therefore, it is necessary to conduct reliability testing on these status parameters to ensure that they can truly reflect the actual operating status of the display and control system.
[0005] However, in actual flight, the numerical distribution of the aforementioned state parameters exhibits significant dynamic changes. During different flight phases, such as taxiing, takeoff, climb, cruise, descent, and landing, the workload and resource consumption of the display and control system vary significantly, leading to different normal numerical distributions of the state parameters. For example, during takeoff and climb, the display and control system has a heavy data processing workload, with high processor utilization and data transmission bandwidth usage; while during cruise, the system load is relatively low, and the normal numerical range of the aforementioned state parameters decreases significantly. Therefore, the normal data distribution of the same state parameter is not the same in different flight phases, and its reasonable value range changes dynamically with the flight phase.
[0006] Existing data reliability detection methods typically use a fixed threshold judgment method. Since the normal data distribution of the above-mentioned state parameters is not the same in different flight stages, it is easy to misjudge normal states as abnormalities or miss real abnormalities as normal, resulting in a decrease in detection accuracy. Summary of the Invention
[0007] In view of this, the purpose of this application is to provide a method, apparatus, equipment and medium for detecting the reliability of status parameters of an aircraft cockpit display and control system, so as to improve the detection accuracy.
[0008] In a first aspect, embodiments of this application provide a method for detecting the reliability of status parameters of an aircraft cockpit display and control system, including: During flight, the status parameters of the aircraft cockpit display and control system are collected in real time to obtain a time series of status parameters; the status parameters include parameters reflecting data transmission capability, data stability, and computing resource status. Extract the state parameters collected within the current time window from the state parameter time series to form the current state parameter time series; The current state parameter time series is input into the trained generative model. The generative model reconstructs a standard state parameter time series that conforms to the standard distribution characteristics of the current flight phase based on the standard distribution characteristics of each state parameter learned during the training phase. Calculate the deviation between the current state parameter time series and the standard state parameter time series; Based on the deviation, the reliability of the state parameters collected within the current time window is determined.
[0009] In conjunction with the first aspect, this application provides a first possible implementation of the first aspect, wherein calculating the deviation between the current state parameter time series and the standard state parameter time series includes: The deviation is calculated using the following formula:
[0010] in, The deviation is represented by N; N is the total number of state parameters collected within the current time window. This represents the value of the i-th state parameter in the time series of the current state parameters; The value of the i-th state parameter in the time series of the standard state parameters.
[0011] In conjunction with the first aspect, this application provides a second possible implementation of the first aspect, wherein determining the reliability of the state parameters collected within the current time window based on the deviation includes: When the deviation is greater than a preset deviation threshold, query the deviation corresponding to the previous M consecutive historical time windows of the current time window; If the deviation of multiple historical time windows in the first M consecutive historical time windows is greater than the preset deviation threshold, then the next flight stage of the current flight stage is taken as the new current flight stage, and the generative model is dynamically updated based on the training samples of the new current flight stage, so that the generative model relearns the standard distribution characteristics of each state parameter of the new current flight stage, and obtains the dynamically updated generative model. The dynamically updated generative model is used as the new trained generative model, and the execution step is returned to input the current state parameter time series into the trained generative model.
[0012] In conjunction with the second possible implementation of the first aspect, this application provides a third possible implementation of the first aspect, wherein, after querying the deviations corresponding to the previous M consecutive historical time windows when the deviation is less than a preset deviation threshold, the method further includes: If the deviation corresponding to the first M consecutive historical time windows is less than or equal to the preset deviation threshold, then the state parameters in the current time window are determined to be unreliable.
[0013] In conjunction with the second possible implementation of the first aspect, this application provides a fourth possible implementation of the first aspect, wherein determining the reliability of the state parameters collected within the current time window based on the deviation further includes: When the deviation is less than or equal to the preset deviation threshold, the state parameters within the current time window are determined to be reliable.
[0014] In conjunction with the second possible implementation of the first aspect, this application provides a fifth possible implementation of the first aspect, wherein the step of dynamically updating the generative model based on training samples of the new current flight phase, so that the generative model relearns the standard distribution characteristics of each state parameter of the new current flight phase, to obtain a dynamically updated generative model, includes: Obtain new training samples for the current flight phase; the training samples include historical state parameters and newly added state parameters that conform to the standard distribution characteristics of each state parameter in the current flight phase; the historical state parameters are the training samples used when the generative model was first trained; the newly added state parameters are the state parameters accumulated by the generative model during its use and determined to be reliable by the generative model. For each historical state parameter that conforms to the standard distribution characteristics of the state parameters in the current flight phase, the sampling probability of each historical state parameter is calculated using the following formula:
[0015] in, This represents the sampling probability of the j-th historical state parameter that conforms to the standard distribution characteristics of the state parameters in the current flight phase. The j-th historical state parameter; These are the standard state parameters reconstructed based on the j-th historical state parameter; This is a historical sample set of state parameters that conforms to the standard distribution characteristics of each state parameter in the current flight phase; Based on the sampling probability of each historical state parameter, multiple historical state parameters are collected from the historical sample set as target historical state parameters. The target historical state parameters and the newly added state parameters are used to train the generative model in this phase, so that the generative model can relearn the standard distribution characteristics of each state parameter in the new current flight phase, and obtain a dynamically updated generative model.
[0016] In conjunction with the fifth possible implementation of the first aspect, this application provides a sixth possible implementation of the first aspect, wherein the generative model includes an encoder and a generator, and the generator includes a bottom-level module and a top-level module; the encoder is used to perform dimensionality reduction processing on the input current state parameter time series to extract latent representation features; the bottom-level module is used to convert the latent representation features into intermediate features; the top-level module is used to convert the intermediate features into a standard state parameter time series that conforms to the standard distribution features of each state parameter in the current flight stage based on the standard distribution features of each state parameter in the current flight stage learned during the training phase; The process of training the generative model using the target historical state parameters and the newly added state parameters, so that the generative model can relearn the standard distribution characteristics of each state parameter in the new current flight phase, and obtain a dynamically updated generative model, includes: The learnable parameters in the top-level module are trained using the target historical state parameters and the newly added state parameters, so that the top-level module can relearn the standard distribution characteristics of each state parameter in the new current flight phase, and obtain a dynamically updated generative model.
[0017] Secondly, embodiments of this application also provide a reliability detection device for the status parameters of an aircraft cockpit display and control system, comprising: The data acquisition module is used to acquire the status parameters of the aircraft cockpit display and control system in real time during the flight of the aircraft, and obtain the status parameter time series; the status parameters include parameters reflecting data transmission capability, data stability and computing resource status; The extraction module is used to extract the state parameters collected within the current time window from the state parameter time series to form the current state parameter time series; The input module is used to input the current state parameter time series into the trained generative model. The generative model reconstructs a standard state parameter time series that conforms to the standard distribution characteristics of the current flight phase based on the standard distribution characteristics of each state parameter learned during the training phase. The calculation module is used to calculate the deviation between the current state parameter time series and the standard state parameter time series; The judgment module is used to determine the reliability of the state parameters collected within the current time window based on the deviation.
[0018] Thirdly, embodiments of this application also provide an electronic device, including: a processor, a memory, and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the memory via the bus, and when the machine-readable instructions are executed by the processor, the steps in any of the possible implementations of the first aspect described above are performed.
[0019] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps in any of the possible implementations of the first aspect described above.
[0020] This application provides a method, apparatus, device, and medium for reliable detection of state parameters in an aircraft cockpit display and control system. Specifically, this embodiment reconstructs a standard state parameter time series that conforms to the standard distribution characteristics of each state parameter in the current flight phase, based on the standard distribution characteristics learned by a trained generative model. This allows the detection benchmark to dynamically match with the flight phase, avoiding the problem of decreased detection accuracy caused by a fixed threshold failing to adapt to changes in data distribution, thus improving the accuracy of reliable detection of cockpit display and control system state parameters. Furthermore, it calculates the deviation between the current state parameter time series and the standard state parameter time series, and judges the reliability of the state parameters collected within the current time window based on this deviation, rather than directly comparing a single state parameter with a fixed threshold. This deviation comprehensively reflects the degree to which multiple state parameters deviate from the standard distribution characteristics of the current flight phase within the current time window, reducing misjudgments caused by instantaneous fluctuations of a single state parameter and improving the reliability of reliable detection of cockpit display and control system state parameters.
[0021] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description
[0022] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0023] Figure 1 A flowchart is shown below illustrating a method for detecting the reliability of status parameters of an aircraft cockpit display and control system, as provided in an embodiment of this application. Figure 2 A flowchart is shown below illustrating a method for determining the reliability of state parameters collected within the current time window, as provided in an embodiment of this application. Figure 3 A schematic diagram of the structure of a reliability detection device for the status parameters of an aircraft cockpit display and control system provided in an embodiment of this application is shown. Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation
[0024] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.
[0025] Considering that existing data reliability detection methods typically employ fixed threshold judgments, and given that the normal data distribution of aircraft cockpit display and control system status parameters varies across different flight phases, it is easy to misjudge normal states as abnormal ones, or to miss true abnormalities as normal ones, leading to decreased detection accuracy. Therefore, this application provides a method, apparatus, device, and medium for reliable detection of aircraft cockpit display and control system status parameters, which are described below through embodiments.
[0026] To facilitate understanding of this embodiment, a method for detecting the reliability of status parameters of an aircraft cockpit display and control system, as disclosed in this application, will first be described in detail. For example... Figure 1 As shown, the process includes the following steps S101-S105: S101: During aircraft flight, the status parameters of the aircraft cockpit display and control system are collected in real time to obtain a time series of status parameters; the status parameters include parameters reflecting data transmission capability, data stability and computing resource status.
[0027] In this embodiment, the status parameters generated by the display and control system during operation can be acquired in real time through the existing data acquisition interface of the aircraft cockpit display and control system. These status parameters are used to reflect whether the display and control system itself is working normally and reliably, and are not the flight data (such as flight speed, altitude, attitude, etc.) processed and displayed by the display and control system.
[0028] In this embodiment, parameters reflecting data transmission capability may include data transmission bandwidth; parameters reflecting data stability may include latency jitter and packet loss rate; and parameters reflecting computing resource status may include processor utilization (i.e., CPU utilization) and memory utilization. These various status parameters are arranged sequentially according to the acquisition time, forming a status parameter time series. Each time point in this status parameter time series corresponds to a set of multi-dimensional status parameters, that is, it simultaneously contains multiple status parameter values from different dimensions.
[0029] For example, the state parameters collected at a certain moment may include the data transmission bandwidth, latency jitter, packet loss rate, processor utilization, and memory utilization at that moment. The real-time acquisition process continues throughout the entire flight of the aircraft, thereby obtaining a continuous time series of state parameters.
[0030] S102: Extract the state parameters collected within the current time window from the state parameter time series to form the current state parameter time series.
[0031] In this embodiment, since the state parameter time series is collected continuously, while the generation model processes data of a fixed length each time it performs a credibility test, it is necessary to extract a segment from the continuous state parameter time series as the input for this test.
[0032] Specifically, a time window length is preset, which represents the number of consecutive time points included in each detection. The time window length can be set according to the rate of change of the display and control system's state parameters and the real-time requirements of the detection. For example, the time window length can be set to 10 seconds, 30 seconds, etc., and this embodiment does not impose a specific limitation on it.
[0033] At the current detection moment, all state parameters closest to the current moment and with a time span equal to the aforementioned time window length are extracted from the state parameter time series to form the current state parameter time series. This current state parameter time series is the set of state parameters collected within the current time window, containing multidimensional state parameter values at each time point within the current time window.
[0034] For example, if the time window length is set to include N consecutive time points, then the current state parameter time series contains all state parameters between the (t-N+1)th time point and the tth time point, where t represents the current time point.
[0035] S103: Input the current state parameter time series into the trained generative model. Based on the standard distribution characteristics of each state parameter in the current flight phase learned during the training phase, the generative model reconstructs a standard state parameter time series that conforms to the standard distribution characteristics of each state parameter in the current flight phase.
[0036] In this embodiment, after obtaining the time series of the current state parameters, it is input into the trained generative model. This generative model has already learned the standard distribution characteristics of each state parameter under different flight phases during the training phase. The standard distribution characteristics refer to the statistical distribution patterns exhibited by each state parameter under normal and reliable conditions, including but not limited to the normal value range, mean, variance, correlation between parameters, and normal fluctuation patterns over time. Due to differences in the workload and resource consumption of the display and control system, the standard distribution characteristics of each state parameter also differ under different flight phases. For example, processor utilization and data transmission bandwidth utilization are typically higher during takeoff and climb, and relatively lower during cruise.
[0037] During the training phase, the generative model learns from a large number of historical state parameter time series under normal conditions, thereby mastering the standard distribution characteristics of each state parameter in the current flight phase. When the current state parameter time series is input into the trained generative model, the model reconstructs the input data based on the learned standard distribution characteristics of each state parameter in the current flight phase, outputting a new time series, which is the standard state parameter time series. The standard state parameter time series represents the numerical sequence that each state parameter should theoretically exhibit under normal and reliable conditions in the current flight phase, that is, a state parameter time series that conforms to the standard distribution characteristics.
[0038] It's important to note that the reconstruction process of the generative model is not a simple copy of the input data, but rather a generative reconstruction based on the learned standard distribution characteristics. If the input data itself is normal and reliable, the difference between the reconstructed standard state parameter time series and the input data will be small; if the input data contains anomalies, the difference between the reconstructed standard state parameter time series and the input data will be large. Therefore, by calculating the deviation between the current state parameter time series and the standard state parameter time series in subsequent steps, the reliability of the state parameters collected within the current time window can be determined.
[0039] S104: Calculate the deviation between the current state parameter time series and the standard state parameter time series.
[0040] In one possible implementation, step S104 can be performed according to the following steps: The deviation is calculated using the following formula:
[0041] in, Indicates the deviation; N is the total number of state parameters collected within the current time window; This represents the value of the i-th state parameter in the current state parameter time series; The value of the i-th state parameter in the standard state parameter time series.
[0042] This formula compares the differences between the current state parameter time series and the corresponding state parameters in the standard state parameter time series, takes the logarithm, and then averages the results to obtain the overall degree of deviation. The closer the two are, the better. and The closer the ratio is to 1, the closer its logarithm is to 0, indicating a higher degree of deviation. The smaller the value, the greater the difference between the two, and the greater the deviation. The larger the deviation, the better. Therefore, the deviation can reflect the extent to which the state parameters collected within the current time window deviate from the standard distribution characteristics of the current flight phase.
[0043] S105: Based on the deviation, determine the reliability of the state parameters collected within the current time window.
[0044] In one possible implementation, such as Figure 2 As shown, when executing step S105, the specific steps S1051-S1053 can be performed as follows: S1051: When the deviation is greater than the preset deviation threshold, query the deviation corresponding to the previous M consecutive historical time windows of the current time window.
[0045] The preset deviation threshold is a pre-determined judgment boundary based on the deviation distribution between the time series of state parameters under normal conditions and the time series of standard state parameters, used to distinguish between credible and unreliable states. M is a preset positive integer representing the number of consecutive historical time windows. By querying the deviation of the first M consecutive historical time windows, it can be determined whether the current increase in deviation is an occasional instantaneous phenomenon or a persistent trend.
[0046] S1052: If the deviation of multiple historical time windows in the first M consecutive historical time windows is greater than the preset deviation threshold, then the next flight stage of the current flight stage is taken as the new current flight stage, and the generative model is dynamically updated based on the training samples of the new current flight stage, so that the generative model can relearn the standard distribution characteristics of each state parameter of the new current flight stage, and obtain the dynamically updated generative model.
[0047] In this embodiment, the existence of multiple historical time windows with deviations exceeding a preset deviation threshold means that within the M historical time windows queried, more than one time window's deviation exceeds the preset deviation threshold. This situation indicates that the increased deviation is not caused by instantaneous fluctuations in a single time window, but is persistent, usually signifying a change in the aircraft's flight phase, leading to an overall change in the normal data distribution of state parameters. Therefore, it is necessary to use the next flight phase as the new current flight phase and dynamically update the generative model using the training samples of this new current flight phase. This allows the generative model to relearn the standard distribution characteristics of each state parameter in the new current flight phase to adapt to the new data distribution.
[0048] S1053: Use the dynamically updated generative model as the new trained generative model, and return to the execution step to input the current state parameter time series into the trained generative model.
[0049] In this embodiment, the updated generative model can reconstruct the standard state parameter time series according to the standard distribution characteristics of the new current flight phase, thereby ensuring the accuracy of subsequent deviation calculations and the effectiveness of reliability judgments.
[0050] In one possible implementation, such as Figure 2 As shown, after executing step S1051, the following step S1054 can also be executed: S1054: If the deviation corresponding to the first M consecutive historical time windows is less than or equal to the preset deviation threshold, then the state parameters in the current time window are determined to be unreliable.
[0051] In this embodiment, although the current deviation is greater than the preset deviation threshold, the deviations of the previous M historical time windows have not exceeded the preset deviation threshold, indicating that the increase in deviation is instantaneous and not due to continuous distribution changes caused by flight phase switching. Therefore, the state parameters within the current time window are determined to be unreliable.
[0052] In one possible implementation, when performing step S105, the following steps S1055 can also be performed: S1055: When the deviation is less than or equal to the preset deviation threshold, the state parameters within the current time window are determined to be reliable.
[0053] In this embodiment, the difference between the current state parameter time series and the standard state parameter time series is within the allowable range, indicating that the state parameters collected within the current time window conform to the standard distribution characteristics of each state parameter in the current flight phase, and can be considered reliable.
[0054] In this embodiment, the reliability of the state parameters collected within the current time window can be judged based on the deviation between the standard state parameter time series reconstructed by the generative model and the current state parameter time series. When the deviation is continuously abnormal, the flight phase switching and model update are triggered. When the deviation is momentarily abnormal, it is determined to be unreliable. This adapts to the dynamic changes in the normal data distribution of state parameters under different flight phases and improves the detection accuracy.
[0055] In one possible implementation, when performing step S1052 to dynamically update the generative model based on the training samples of the new current flight phase, so that the generative model relearns the standard distribution characteristics of each state parameter of the new current flight phase and obtains the dynamically updated generative model, the following steps S10521-S10524 can be specifically performed: S10521: Obtain new training samples for the current flight phase; the training samples include historical state parameters and newly added state parameters that conform to the standard distribution characteristics of each state parameter in the current flight phase; the historical state parameters are the training samples used when the generative model was first trained; the newly added state parameters are the state parameters accumulated by the generative model during its use and determined to be reliable by the generative model.
[0056] In this embodiment, after the new current flight phase is determined, training samples need to be collected to update the generative model. These training samples consist of two parts: The first part consists of historical state parameters that conform to the standard distribution characteristics of state parameters in the current flight phase. These historical state parameters refer to the training samples used during the initial training of the generative model; they originate from historical data collected by the aircraft under normal conditions in each flight phase. Since the standard distribution characteristics of state parameters differ across flight phases, it is necessary to select samples from all historical training samples that conform to the standard distribution characteristics of state parameters in the new current flight phase as the historical state parameters available for this update.
[0057] The second part consists of newly added state parameters. These parameters are those collected in real-time during flight after the generative model is deployed and deemed reliable by the model. Specifically, when step S1055 determines that the state parameters collected within the current time window are reliable, these parameters can be accumulated and saved as new training samples for subsequent dynamic updates of the generative model. These samples reflect the normal state data newly collected during actual use of the generative model, providing effective information under the current operating conditions for model updates.
[0058] S10522: For each historical state parameter that conforms to the standard distribution characteristics of each state parameter in the current flight phase, the sampling probability of each historical state parameter is calculated using the following formula:
[0059] in, This represents the sampling probability of the j-th historical state parameter that conforms to the standard distribution characteristics of the state parameters in the current flight phase. The j-th historical state parameter; These are the standard state parameters reconstructed based on the j-th historical state parameter; This is a historical sample set of state parameters that conforms to the standard distribution characteristics of each state parameter in the current flight phase.
[0060] The numerator of the formula Let L1 norm represent the difference between the j-th historical state parameter and its corresponding standard state parameter, reflecting the reconstruction error when this historical state parameter is reconstructed by the current generative model. The denominator is the set of historical samples. The sum of the L1 norms of the reconstruction errors of all historical state parameters serves as a normalization function. Therefore, the sampled probability... This reflects the proportion of the reconstruction error of the j-th historical state parameter in the total reconstruction error of all historical state parameters. Historical state parameters with larger reconstruction errors have a higher sampling probability, indicating they are more likely to be prioritized for training in this model update. The purpose of this design is to make the model pay more attention to historical state parameters with poor current model reconstruction performance and insufficient feature learning during updates, thereby more effectively learning the standard distribution characteristics of each state parameter under the new current flight phase and improving the model's adaptability to new data distributions.
[0061] S10523: Based on the sampling probability of each historical state parameter, collect multiple historical state parameters from the historical sample set as target historical state parameters.
[0062] In this embodiment, after calculating the sampling probability of each historical state parameter, the parameters are selected from the historical sample set according to that probability. Sampling is performed to obtain several historical state parameters, which are then used as target historical state parameters for the current model update training. The sampling process is without replacement. Since the sampling probability is related to the magnitude of the reconstruction error, historical state parameters with larger reconstruction errors are more likely to be collected as target historical state parameters, thus making the training samples more targeted.
[0063] S10524: Use the target's historical state parameters and newly added state parameters to train the generative model in this phase, so that the generative model can relearn the standard distribution characteristics of each state parameter in the new current flight phase and obtain a dynamically updated generative model.
[0064] In this embodiment, after obtaining the target historical state parameters and the newly added state parameters, both are used as training samples for this model training. The target historical state parameters provide the model with historical distribution information of normal state parameters under the new current flight phase, while the newly added state parameters reflect reliable state parameters collected by the generative model during actual use. Together, they help the generative model learn the standard distribution characteristics of each state parameter in the new current flight phase. Through this training, the learnable parameters in the generative model are adjusted, resulting in a dynamically updated generative model that can accurately reconstruct the standard state parameter time series that conforms to the standard distribution characteristics of the new flight phase.
[0065] In one possible implementation, the generative model includes an encoder and a generator, the generator including a bottom-level module and a top-level module; the encoder is used to perform dimensionality reduction processing on the input current state parameter time series to extract latent representation features; the bottom-level module is used to convert the latent representation features into intermediate features; the top-level module is used to convert the intermediate features into a standard state parameter time series that conforms to the standard distribution features of the current flight phase state parameters based on the standard distribution features of the current flight phase state parameters learned in the training phase. Based on the structure of the above generative model, step S10524 can be executed as follows: The learnable parameters in the top-level module are trained using the target's historical state parameters and newly added state parameters, so that the top-level module can relearn the standard distribution characteristics of each state parameter in the new current flight phase, and obtain a dynamically updated generative model.
[0066] In other words, this model training only updates the learnable parameters of the top-level module in the generator, while the parameters of the encoder and the bottom-level modules of the generator remain unchanged during training. The reason for updating only the top-level module is that it is directly responsible for converting intermediate features into the final standard state parameter time series, and its output is directly related to the specific numerical distribution of each state parameter. When the flight phase changes, causing a change in the standard distribution of state parameters, only the parameters of the top-level module need to be adjusted to adapt the generative model to the new data distribution. The latent representation features and intermediate features extracted by the bottom-level modules and encoder are relatively general feature expressions, exhibiting relative stability across different flight phases. Keeping their parameters unchanged preserves the general feature extraction capabilities learned by the generative model in the initial training phase, avoiding the model forgetting existing knowledge due to a comprehensive update. It also reduces the computational cost of model updates and improves the efficiency of dynamic updates.
[0067] Through the steps S10521 to S10524 above, the generated model can be locally updated using the training samples of the new current flight phase after the flight phase switch, thereby quickly adapting to the new standard distribution characteristics of state parameters and ensuring the accuracy of subsequent credibility detection.
[0068] Based on the same technical concept, embodiments of this application also provide a reliability detection device for the status parameters of an aircraft cockpit display and control system, such as... Figure 3 As shown, it includes: The acquisition module 301 is used to acquire the status parameters of the aircraft cockpit display and control system in real time during the flight of the aircraft, and obtain the status parameter time series; the status parameters include parameters reflecting data transmission capability, data stability and computing resource status; Extraction module 302 is used to extract the state parameters collected within the current time window from the state parameter time series to form the current state parameter time series; Input module 303 is used to input the current state parameter time series into the trained generative model. The generative model reconstructs a standard state parameter time series that conforms to the standard distribution characteristics of the current flight phase based on the standard distribution characteristics of each state parameter learned during the training phase. Calculation module 304 is used to calculate the deviation between the current state parameter time series and the standard state parameter time series; The judgment module 305 is used to judge the reliability of the state parameters collected within the current time window based on the deviation.
[0069] Optionally, when calculating the deviation between the current state parameter time series and the standard state parameter time series, the calculation module 304 is specifically used for: The deviation is calculated using the following formula:
[0070] in, The deviation is represented by N; N is the total number of state parameters collected within the current time window. This represents the value of the i-th state parameter in the time series of the current state parameters; The value of the i-th state parameter in the time series of the standard state parameters.
[0071] Optionally, when the judgment module 305 is used to judge the reliability of the state parameters collected within the current time window based on the deviation, it is specifically used for: When the deviation is greater than a preset deviation threshold, query the deviation corresponding to the previous M consecutive historical time windows of the current time window; If the deviation of multiple historical time windows in the first M consecutive historical time windows is greater than the preset deviation threshold, then the next flight stage of the current flight stage is taken as the new current flight stage, and the generative model is dynamically updated based on the training samples of the new current flight stage, so that the generative model relearns the standard distribution characteristics of each state parameter of the new current flight stage, and obtains the dynamically updated generative model. The dynamically updated generative model is used as the new trained generative model, and the execution step is returned to input the current state parameter time series into the trained generative model.
[0072] Optionally, the device further includes: The determination module is used to, when the deviation is greater than a preset deviation threshold in the determination module 305, query the deviation corresponding to the previous M consecutive historical time windows of the current time window, and if the deviation corresponding to the previous M consecutive historical time windows is less than or equal to the preset deviation threshold, then determine that the state parameters in the current time window are unreliable.
[0073] Optionally, when the judgment module 305 is used to judge the reliability of the state parameters collected within the current time window based on the deviation, it is further used to: When the deviation is less than or equal to the preset deviation threshold, the state parameters within the current time window are determined to be reliable.
[0074] Optionally, when the judgment module 305 dynamically updates the generative model based on the training samples of the new current flight phase, so that the generative model relearns the standard distribution characteristics of each state parameter of the new current flight phase and obtains the dynamically updated generative model, it is specifically used for: Obtain new training samples for the current flight phase; the training samples include historical state parameters and newly added state parameters that conform to the standard distribution characteristics of each state parameter in the current flight phase; the historical state parameters are the training samples used when the generative model was first trained; the newly added state parameters are the state parameters accumulated by the generative model during its use and determined to be reliable by the generative model. For each historical state parameter that conforms to the standard distribution characteristics of the state parameters in the current flight phase, the sampling probability of each historical state parameter is calculated using the following formula:
[0075] in, This represents the sampling probability of the j-th historical state parameter that conforms to the standard distribution characteristics of the state parameters in the current flight phase. The j-th historical state parameter; These are the standard state parameters reconstructed based on the j-th historical state parameter; This is a historical sample set of state parameters that conforms to the standard distribution characteristics of each state parameter in the current flight phase; Based on the sampling probability of each historical state parameter, multiple historical state parameters are collected from the historical sample set as target historical state parameters. The target historical state parameters and the newly added state parameters are used to train the generative model in this phase, so that the generative model can relearn the standard distribution characteristics of each state parameter in the new current flight phase, and obtain a dynamically updated generative model.
[0076] Optionally, the generative model includes an encoder and a generator, the generator including a bottom-level module and a top-level module; the encoder is used to perform dimensionality reduction processing on the input current state parameter time series to extract latent representation features; the bottom-level module is used to convert the latent representation features into intermediate features; the top-level module is used to convert the intermediate features into a standard state parameter time series that conforms to the standard distribution features of each state parameter in the current flight stage based on the standard distribution features of each state parameter in the current flight stage learned during the training phase; When the judgment module 305 is used to train the generator model using the target historical state parameters and the newly added state parameters, so that the generator model can relearn the standard distribution characteristics of each state parameter in the new current flight phase and obtain a dynamically updated generator model, it is specifically used for: The learnable parameters in the top-level module are trained using the target historical state parameters and the newly added state parameters, so that the top-level module can relearn the standard distribution characteristics of each state parameter in the new current flight phase, and obtain a dynamically updated generative model.
[0077] Figure 4 A schematic diagram of an electronic device provided in this application embodiment includes: a processor 401, a memory 402, and a bus 403. The memory 402 stores machine-readable instructions executable by the processor 401. When the electronic device runs the above-described information processing method, the processor 401 and the memory 402 communicate through the bus 403. The processor 401 executes the machine-readable instructions to perform the steps described in the embodiment of the reliability detection method for the status parameters of the aircraft cockpit display and control system.
[0078] This application also provides a computer-readable storage medium storing a computer program, which, when run by a processor, executes the steps described in the embodiment of the method for detecting the reliability of status parameters of an aircraft cockpit display and control system.
[0079] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the above-described device, electronic equipment, and computer-readable storage medium can be referred to the corresponding process in the aforementioned embodiment of the reliability detection method for the status parameters of the aircraft cockpit display and control system, and will not be repeated here.
[0080] In the several embodiments provided in this application, it should be understood that the disclosed methods, apparatuses, electronic devices, and computer-readable storage media can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or modules may be electrical, mechanical, or other forms.
[0081] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0082] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0083] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0084] Finally, it should be noted that the above-described embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The scope of protection of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this application. Such modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of protection of the claims.
Claims
1. A method for detecting the reliability of status parameters of an aircraft cockpit display and control system, characterized in that, include: During aircraft flight, the status parameters of the aircraft cockpit display and control system are collected in real time to obtain the time series of status parameters. The status parameters include parameters that reflect data transmission capability, data stability, and computing resource status; Extract the state parameters collected within the current time window from the state parameter time series to form the current state parameter time series; The current state parameter time series is input into the trained generative model. The generative model reconstructs a standard state parameter time series that conforms to the standard distribution characteristics of the current flight phase based on the standard distribution characteristics of each state parameter learned during the training phase. Calculate the deviation between the current state parameter time series and the standard state parameter time series; Based on the deviation, the reliability of the state parameters collected within the current time window is determined.
2. The method according to claim 1, characterized in that, The calculation of the deviation between the current state parameter time series and the standard state parameter time series includes: The deviation is calculated using the following formula: in, The deviation is represented by N; N is the total number of state parameters collected within the current time window. This represents the value of the i-th state parameter in the time series of the current state parameters; The value of the i-th state parameter in the standard state parameter time series.
3. The method according to claim 1, characterized in that, The determination of the reliability of the state parameters collected within the current time window based on the deviation includes: When the deviation is greater than a preset deviation threshold, query the deviation corresponding to the previous M consecutive historical time windows of the current time window; If the deviation of multiple historical time windows in the first M consecutive historical time windows is greater than the preset deviation threshold, then the next flight stage of the current flight stage is taken as the new current flight stage, and the generative model is dynamically updated based on the training samples of the new current flight stage, so that the generative model relearns the standard distribution characteristics of each state parameter of the new current flight stage, and obtains the dynamically updated generative model. The dynamically updated generative model is used as the new trained generative model, and the execution step is returned to input the current state parameter time series into the trained generative model.
4. The method according to claim 3, characterized in that, When the deviation is less than a preset deviation threshold, after querying the deviations corresponding to the M consecutive historical time windows of the current time window, the method further includes: If the deviation corresponding to the first M consecutive historical time windows is less than or equal to the preset deviation threshold, then the state parameters in the current time window are determined to be unreliable.
5. The method according to claim 3, characterized in that, The step of determining the reliability of the state parameters collected within the current time window based on the deviation also includes: When the deviation is less than or equal to the preset deviation threshold, the state parameters within the current time window are determined to be reliable.
6. The method according to claim 3, characterized in that, The process of dynamically updating the generative model based on training samples from the new current flight phase, so that the generative model relearns the standard distribution characteristics of each state parameter in the new current flight phase, yields a dynamically updated generative model, including: Obtain new training samples for the current flight phase; the training samples include historical state parameters and newly added state parameters that conform to the standard distribution characteristics of each state parameter in the current flight phase; the historical state parameters are the training samples used when the generative model was first trained; the newly added state parameters are the state parameters accumulated by the generative model during its use and determined to be reliable by the generative model. For each historical state parameter that conforms to the standard distribution characteristics of the state parameters in the current flight phase, the sampling probability of each historical state parameter is calculated using the following formula: in, This represents the sampling probability of the j-th historical state parameter that conforms to the standard distribution characteristics of each state parameter in the current flight phase; The j-th historical state parameter; These are the standard state parameters reconstructed based on the j-th historical state parameter; This is a historical sample set of state parameters that conforms to the standard distribution characteristics of each state parameter in the current flight phase; Based on the sampling probability of each historical state parameter, multiple historical state parameters are collected from the historical sample set as target historical state parameters. The target historical state parameters and the newly added state parameters are used to train the generative model in this phase, so that the generative model can relearn the standard distribution characteristics of each state parameter in the new current flight phase, and obtain a dynamically updated generative model.
7. The method according to claim 6, characterized in that, The generative model includes an encoder and a generator. The generator includes a bottom-level module and a top-level module. The encoder is used to perform dimensionality reduction processing on the input current state parameter time series to extract latent representation features. The bottom-level module is used to convert the latent representation features into intermediate features. The top-level module is used to convert the intermediate features into a standard state parameter time series that conforms to the standard distribution features of the state parameters of the current flight phase, based on the standard distribution features of the state parameters of the current flight phase learned during the training phase. The process of training the generative model using the target historical state parameters and the newly added state parameters, so that the generative model can relearn the standard distribution characteristics of each state parameter in the new current flight phase, and obtain a dynamically updated generative model, includes: The learnable parameters in the top-level module are trained using the target historical state parameters and the newly added state parameters, so that the top-level module can relearn the standard distribution characteristics of each state parameter in the new current flight phase, and obtain a dynamically updated generative model.
8. A reliability testing device for status parameters of an aircraft cockpit display and control system, characterized in that, include: The data acquisition module is used to acquire the status parameters of the aircraft cockpit display and control system in real time during the flight of the aircraft, and obtain the time series of status parameters. The status parameters include parameters that reflect data transmission capability, data stability, and computing resource status; The extraction module is used to extract the state parameters collected within the current time window from the state parameter time series to form the current state parameter time series; The input module is used to input the current state parameter time series into the trained generative model. The generative model reconstructs a standard state parameter time series that conforms to the standard distribution characteristics of the current flight phase based on the standard distribution characteristics of each state parameter learned during the training phase. The calculation module is used to calculate the deviation between the current state parameter time series and the standard state parameter time series; The judgment module is used to determine the reliability of the state parameters collected within the current time window based on the deviation.
9. An electronic device, characterized in that, include: The device includes a processor, a memory, and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is in operation, the processor communicates with the memory via the bus, and the machine-readable instructions, when executed by the processor, perform the steps of the method as described in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the method as described in any one of claims 1 to 7.
