An intelligent flight monitoring method for large UAV ground stations

By intelligently processing remote control instructions and telemetry data of the drone ground station, an intelligent flight monitoring system was established, which solved the problem of insufficient traditional "display control" capabilities, and realized intelligent auxiliary decision-making and fault management of large drones. Improves the efficiency and safety of flight missions.

CN115798269BActive Publication Date: 2024-11-19THE 54TH RESEARCH INSTITUTE OF CHINA ELECTRONICS TECHNOLOGY GROUP CORPORATION
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202211700472.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-12-29
Publication Date
2024-11-19
Estimated Expiration
2042-12-29

AI Technical Summary

Technical Problem

The basic "display control" capabilities of traditional drone ground command and measurement and control stations cannot meet the needs of ground inspection and support and flight missions of large drones. The operators face high mental pressure and risk of decision-making errors, and lack intelligent flight monitoring and fault management capabilities.

Method used

An intelligent flight monitoring method for large drone ground stations was designed. By reading remote control instructions and telemetry parameter attribute configuration tables, data collection, preprocessing and analysis are carried out to generate remote control operations and telemetry status feedback during flight missions. The process, combined with experts to formulate alarm rules, carry out fault alarms and command pushes, realize data transparency and visual presentation, and establish aircraft control feature models and fault prediction alarm models.

Benefits of technology

It has enhanced the intelligent decision-making capabilities of the drone ground station, reduced the mental pressure and decision-making mistakes of the operator, improved the efficiency and safety of the flight mission, and realized intelligent auxiliary control and fault management of the drone.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115798269B_ABST
    Figure CN115798269B_ABST
Patent Text Reader

Abstract

The present invention belongs to the field of UAV ground measurement and control software technology, and discloses an intelligent flight monitoring method for large UAV ground stations. In order to solve the problem that the basic "display and control" capabilities of traditional UAV ground command and control stations are insufficient and cannot meet the requirements of UAV ground inspection and flight missions, it is necessary to provide a corresponding intelligent decision-making auxiliary flight monitoring method to help operators perform fault management, platform control and flight control management, so that the UAV can maximize its combat effectiveness. The present invention specifically includes data collection and analysis processing, pre-flight automatic processing, message intelligent push and flight mission evaluation. The present invention rationally integrates and distributes multiple functions, and expands to intelligent flight monitoring based on the data law "features" and visualization curves obtained by data analysis and processing.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of unmanned aerial vehicle ground measurement and control software, and in particular relates to an intelligent flight monitoring method for a large unmanned aerial vehicle ground station. Background Art

[0002] Due to the advantages of small size, low cost, good maneuverability, good concealment and strong adaptability, the application of drones is becoming more and more extensive and their development prospects are broad, especially large drones, which are of great significance to my country's military and civilian fields. They are mainly used for military reconnaissance and strike, emergency communication relay, artificial rainmaking and meteorological detection, and material transportation.

[0003] A complete UAV system consists of the UAV platform itself and the ground command and control station. The UAV ground command and control station, as the "nerve center" of the entire UAV system, plays a vital role in the entire UAV system. Various domestic universities, research institutes and military units have conducted extensive research on the platform development and intelligent autonomous control technology of large UAVs, but the development is limited, and the research on the intelligentization of UAV ground command and control stations has become particularly prominent.

[0004] With the rapid development of UAV technology, its applications are becoming more and more extensive, and its performance is becoming increasingly powerful. Ground flight monitoring involves more complex logic and control. The basic "display and control" capabilities of traditional UAV ground command and control stations can no longer meet the needs of UAV ground inspection and flight missions. Corresponding intelligent decision-making assistance flight monitoring methods are needed to help operators reduce mental stress and load, reduce decision-making and operational errors, and provide operators with intelligent flight decision-making assistance solutions to help them manage faults, platform control and flight control management, reduce their operating burden, and enable UAVs to achieve maximum combat effectiveness. Summary of the invention

[0005] In order to solve the gap in the intelligent processing of flight monitoring in the ground command and control station, the "nerve center" of the UAV system, the present invention implements the theory, improves the various shortcomings of the traditional ground station flight "display and control", provides the operator with an intelligent flight auxiliary decision-making solution, assists the operator to make fast and accurate control decisions, and constructs an intelligent flight monitoring system for large UAV ground stations.

[0006] The present invention adopts the following technical scheme:

[0007] An intelligent flight monitoring method for a large UAV ground station specifically includes the following steps:

[0008] (1) Read the remote control command and telemetry parameter attribute configuration table, and set the remote control command and telemetry parameter object format standard unification rules;

[0009] (2) collecting, preprocessing, and analyzing the input data according to the rules set in step (1); obtaining the remote control operation during the entire flight mission through data processing and analysis of the remote control data; and obtaining the ideal fitting curve and trend perception prediction results of the data through data processing and analysis of the telemetry data;

[0010] (3) For the pre-takeoff ground waiting stage, the UAV flight priori law is combined with the remote control operation characteristics and telemetry ideal fitting curve obtained in step (2) to generate the various command operations and telemetry status feedback confirmation process required for the pre-flight ground waiting stage, and the processing and analysis results are integrated into the structure frame object attributes and saved in the attribute configuration table in step (1);

[0011] (4) For the air mission phase, the telemetry parameter structure frame object storing the warning rules formulated by the experts is integrated with the remote control operation characteristics and telemetry trend perception prediction results obtained through data analysis and processing in step (2). When a specific parameter is abnormal or exceeds the limit, a fault alarm and a designated remote control command push service are performed;

[0012] (5) When the flight mission is completed, the data of the entire process is saved, and the remote control command and telemetry parameter data are compared with the telemetry ideal fitting curve obtained in step (2), and combined with the alarm rules in step (4), illegal data points are marked and counted to evaluate the completion of the flight mission.

[0013] Furthermore, step (2) specifically includes the following steps:

[0014] (2.1) Judge and verify the input source code data, then decode it and encapsulate it into universal remote control command and universal telemetry parameter object classes;

[0015] (2.2) preprocessing the encapsulated universal remote control command and universal telemetry parameter object class according to the rules set in step (1), including data standardization, smoothing filter processing, resampling, feature extraction, anomaly detection and data clipping;

[0016] (2.3) Extract features and correlation coefficients from the preprocessed data, perform multi-parameter visualization analysis, and perform regression and fitting to generate analysis results; among them, after eliminating empty instructions and invalid data from the remote control data, the remote control operation during the entire flight mission is obtained, and the ideal fitting curve and trend perception prediction of the telemetry data are obtained through data processing and analysis.

[0017] Furthermore, the data sources for data input and collection in step (2) are mainly divided into the following three types:

[0018] 1) UAV measurement and control data received through UDP and TCP networks and local COM serial ports;

[0019] 2) Original remote control and telemetry DAT data files;

[0020] 3) Batch import of locally stored processed CSV and XML files.

[0021] The flight monitoring method includes two modes: real-time mission mode and playback analysis mode. In mission mode, the measurement and control data comes from 1), and in playback analysis mode, the measurement and control data comes from 2) and 3).

[0022] The advantages and positive effects of the present invention are:

[0023] (1) Compared with the traditional ground station flight monitoring "display and control" function, the general structure of the UAV remote control command / telemetry parameter object is designed to unify the format standard, collect, save, process and analyze the original remote control and telemetry data of the ground station, and present the entire data management process and analysis results in a transparent and visual manner, so that the versatility and intuitiveness of the present invention are enhanced;

[0024] (2) The present invention introduces data analysis and processing, on this basis, mines data rules, and then establishes an aircraft control feature model and a fault prediction and warning model to realize intelligent flight monitoring of ground stations;

[0025] (3) The present invention combines the rules obtained from data analysis with the prior knowledge of experts, and saves the rules in the form of configuration files. The rules are user-oriented and can be modified and improved according to the ideas specified by the user, making the present invention more user-friendly and flexible. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following is a brief introduction to the drawings required for describing the embodiments.

[0027] Figure 1 It is a schematic diagram of the technical solution flow of the present invention.

[0028] Figure 2 It is a schematic diagram of the data analysis and processing workflow of the present invention.

[0029] Figure 3 It is a schematic diagram of the multi-parameter visualized joint verification analysis of the present invention.

[0030] Figure 4 It is a schematic diagram of the flight phase of the UAV of the present invention.

[0031] Figure 5 It is a schematic diagram of the pre-flight automatic processing list of the present invention.

[0032] Figure 6 It is a schematic diagram of the pre-flight automatic processing workflow of the present invention.

[0033] Figure 7 It is a schematic diagram of remote control instruction message push of the present invention. DETAILED DESCRIPTION

[0034] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0035] The present invention is an intelligent flight monitoring method for large-scale unmanned aerial vehicle ground stations. Figure 1 The technical solution shown includes the following steps. The specific content of each step of the technical solution is described in detail below:

[0036] (1) Standardization of remote control command and telemetry parameter object formats;

[0037] The measurement and control data frame formats of various types of large unmanned aerial vehicles currently in use are different and not uniform, which leads to the diverse and cumbersome work of sending outbound remote control commands and parsing and retrieving downlink telemetry parameters. At this time, a universal data object description method is required to adapt to different types of data frame formats. Therefore, the present invention designs and reads a universal remote control command and telemetry parameter attribute configuration table, so that all remote control command / telemetry parameter objects have a unified format and description method.

[0038] 1) For uplink remote control commands, the present invention uniformly describes each remote control command object according to the structure shown in Table 1 below, including the serial number, command name, command code, command type, data length, data location, and whether it is locked (key commands that require secondary confirmation).

[0039] Table 1 General description structure of remote control command object

[0040]

[0041] 2) For downlink telemetry parameters, the present invention provides a unified description of each telemetry parameter object according to the structure shown in Table 2 below, which includes serial number, parameter name, parameter condition, parameter type, parameter unit, big and small ends, starting position, byte length, parameter range, analytical formula and alarm rule.

[0042] Table 2 General description structure of telemetry parameter objects

[0043]

[0044] (2) Data input collection, processing and analysis;

[0045] UAV measurement and control data processing and analysis technology process Figure 2 As shown, it includes three parts: data collection, preprocessing and analysis:

[0046] 1) The data acquisition part realizes the input, judgment, verification and general parameter decoding of source code data;

[0047] 2) Data preprocessing includes data standardization, smoothing and filtering, resampling, feature extraction, anomaly detection, data clipping, etc. Since different measurement and control data analysis models have different requirements for input data, in order to meet the needs of subsequent analysis, the collected raw data needs to be preprocessed in various ways to meet the input requirements of the analysis model.

[0048] 3) The data analysis part realizes the analysis and processing of the input preprocessed data and generates the analysis results.

[0049] (2.1) The sources of flight measurement and control data targeted by the present invention are mainly divided into the following three types:

[0050] 1) UAV measurement and control data received through UDP, TCP network and local COM serial port;

[0051] 2) Original remote control and telemetry DAT data files;

[0052] 3) Batch import of locally stored processed CSV and XML files.

[0053] Therefore, this flight monitoring method includes two modes: real-time mission mode and playback analysis mode. In mission mode, the measurement and control data comes from 1), and in playback analysis mode, it comes from 2) and 3). The measurement and control information of the above three input drones is usually input through a message queue, specifically in the form of frame-format binary source code data. According to the general description structure of the remote control command and telemetry parameter object given in step (1), the frame-format binary source code data that has been judged and verified can be decoded and processed, and encapsulated into a general remote control command and a general telemetry parameter object class, and these parameter object classes support retrieval.

[0054] (2.2) The data preprocessing part mainly performs the following processing and preparation work on the data:

[0055] 1) Data standardization: normalize the input parameters and convert the input to [0,1] or [-1,1]. After the original data is processed by data standardization, the dimensional influence between the parameters is eliminated, and the parameters are at the same order of magnitude, which is suitable for comprehensive comparative evaluation and analysis:

[0056]

[0057] Among them, u is the mean of all sample data, σ is the standard deviation of all sample data, x is the input parameter, and x′ is the output. The processed data conforms to the standard normal distribution, that is, the mean is 0 and the standard deviation is 1.

[0058] 2) Smoothing filtering: filter the standardized data to obtain the low-frequency signal component of the parameter data. Use mean smoothing filtering. Let the original signal be I(t) and the smoothing filtering result be I′(t). The calculation formula is as follows:

[0059]

[0060] Where W is the signal processing window, and N(W) represents the window width or the number of discrete signal frames in the window. Signal smoothing filtering can be achieved by sliding the window.

[0061] 3) Resample the filtered data to ensure the integrity of the data structure;

[0062] 4) Abnormal detection and wild value removal processing. The filtering algorithm can effectively remove outliers, but the original signal is smoothed after resampling after filtering, resulting in information loss. The smoothed data needs to be further processed to accurately calculate the outliers and restore the original signal by removing the outliers. The outliers can be calculated with the help of the normal distribution 3σ criterion. Calculate the high-frequency components in the original signal:

[0063]

[0064] Using BOX-COX method Perform normal distribution transformation.

[0065]

[0066] The value of λ can be obtained through maximum likelihood estimation.

[0067] The high frequency signal After normal distribution conversion, the outlier detection threshold Th is calculated using the normal distribution 3σ criterion.

[0068] or

[0069] Therefore, the restored signal is

[0070]

[0071] Using the method based on smoothing filter outlier detection combined with BOX-COX to preprocess the measurement and control data can obtain better preprocessing results, but this method is only suitable for post-processing. If you want to preprocess the data in real time, you need to perform a "windowing" operation on the above process, and use the data distribution of the previous window to estimate the abnormality of the current data. W is the signal processing window:

[0072]

[0073] 5) Data trimming: Clear ground data. The data of the drone in the "ground waiting" stage is special. Some equipment and sensors are not working, resulting in invalid data. When analyzing part of the flight mission data, you can choose to remove the ground data.

[0074] (2.3) The data analysis part realizes the analysis and processing of the input pre-processed data and generates the analysis results:

[0075] 1) Calculate the multi-parameter covariance matrix, and then calculate the correlation coefficient, and output the parameters with strong attribute correlation strength. The value range of the correlation coefficient is (-1,1), which measures the relationship between variables by numbers and has directionality. 1 indicates positive correlation, -1 indicates negative correlation, and the closer to 0, the weaker the correlation. xy The calculation formula is as follows:

[0076]

[0077] Among them, S xy Sample covariance calculation formula:

[0078]

[0079] in, and is the sample mean. S x Sample standard deviation calculation formula:

[0080]

[0081] S y Sample standard deviation calculation formula:

[0082]

[0083] 2) Multi-parameter visualization analysis, such as Figure 3As shown, the present invention supports checking multiple parameter object classes, with time as the horizontal axis and parameter value results as the vertical axis, drawing historical or real-time change trend graphs of multiple parameters, supporting visual joint verification analysis of multiple parameters, supporting simultaneous visualization of multiple parameters in multiple windows with different axes or single window with the same axis, and dragging, scaling, color customization, and adding threshold lines to mark statistical out-of-limit data.

[0084] 3) Data statistical analysis: possessing the ability to perform big data statistical analysis, and statistically displaying the mean, variance, maximum value, standard deviation and other data characteristics of parameters within a certain self-selected range.

[0085] 4) Neural network regression: Use the DNN neural network algorithm model to build a regression model of multiple parameters to be analyzed. The real-time parameter is v t The regression prediction result is u 1t ,u 2t ,…u nt is with v t The health factors associated with the attributes are:

[0086]

[0087] Among them, the function Γ(·) represents the nonlinear fitting model.

[0088] Trend analysis modeling, let The feature description function on the time axis is recorded as Then the parameter v t The degradation model of is:

[0089]

[0090] The degradation model of the equipment is:

[0091]

[0092] in,

[0093] Let L Δm is the average error of the mth model verification. Then let the decision threshold be:

[0094] T h = p·max(L Δm )

[0095] Where p is the set false alarm rate. Using a small number of real fault samples S l To T h To intervene, you need to:

[0096] T h '=p·max(L Δm )+εSl

[0097] If Ω(t)>T h 'An abnormal warning is given.

[0098] Select the trend extrapolation algorithm model K(·), let

[0099] Ω(t+T)=K(Ω(t))

[0100] If Ω(t+T)>T h ', it means that the system will fail after time T.

[0101] 5) Staged fitting: Through nonlinear fitting and linear fitting, each flight stage of the UAV is fitted in sections to obtain the ideal parameter fitting curve.

[0102] (3) Automatic processing before flight;

[0103] For the entire UAV flight mission, the “flight phase” telemetry parameters can be roughly divided into the following categories using the statistical analysis of step (2): Figure 4 As for the ground waiting stage before takeoff, the UAV flight a priori law is combined with the remote control operation characteristics and telemetry ideal fitting curve obtained in step (2) to generate the various command operations and telemetry status feedback confirmation process required for the ground waiting stage before flight, and the processing and analysis results are integrated into the structure frame object attributes and saved in the attribute configuration table in step (1). Figure 5 The pre-flight automatic processing module shown in the figure shows the various command operations and status feedback confirmation processes required during the ground waiting stage before takeoff in the form of a list. When the process reaches the corresponding stage, the module pushes the contents of the command pop-up window to be sent to the operator. The workflow is as follows: Figure 6 shown.

[0104] (4) Intelligent message push;

[0105] For the air mission phase, the telemetry parameter structure frame object that stores the alarm rules formulated by the experts is integrated with the remote control operation characteristics and telemetry trend perception prediction results obtained through data analysis and processing in step (2). When a specific parameter is abnormal or exceeds the limit, a fault alarm and a designated remote control command push service are performed. The push content can be divided into two aspects: one is the uplink remote control command, and the other is the downlink telemetry fault information. The push form is as follows: Figure 7 As shown, the comprehensive warning content supports multi-condition editing, that is, multiple trigger conditions must be met to push a certain warning content. In summary, it provides intelligent decision-making push for remote control operation and fault warning prediction of the aircraft, and the push form includes using the QTextToSpeech class in the TTS module to convert text into speech and then perform voice broadcast.

[0106] (5) Flight mission evaluation;

[0107] When the flight mission is completed, the data of the entire process is saved in CSV / XML format, which can provide a playback display of the aircraft platform parameters during the entire flight process. And by comparing the remote control command and telemetry parameter data with the telemetry ideal fitting curve obtained in step (2), and combining the alarm rules in step (4), illegal data points are marked and counted to evaluate the completion of the flight mission.

Claims

1. An intelligent flight monitoring method for a large UAV ground station, characterized in that: The specific steps include: (1) Read the remote control command and telemetry parameter attribute configuration table, and set the standard unified rules for the remote control command and telemetry parameter object format; (2) According to the rules set in step (1), the input data is collected, preprocessed, and analyzed. The remote control operation during the entire flight mission is obtained through data processing and analysis of the remote control data. The ideal fitting curve and trend perception prediction results of the data are obtained through data processing and analysis of the telemetry data. (3) For the pre-takeoff ground waiting stage, the UAV flight priori rules are combined with the remote control operation characteristics and telemetry ideal fitting curve obtained in step (2) to generate the various command operations and telemetry status feedback confirmation process required for the pre-flight ground waiting stage, and the processing and analysis results are integrated into the structure frame object attributes and saved in the attribute configuration table in step (1); (4) For the air mission phase, the telemetry parameter structure frame object storing the warning rules formulated by the experts is integrated with the remote control operation characteristics and telemetry trend perception prediction results obtained through data analysis and processing in step (2). When a specific parameter is abnormal or exceeds the limit, a fault alarm and a designated remote control command push service are performed; (5) When the flight mission is completed, the data of the entire process is saved, and the remote control command and telemetry parameter data are compared with the telemetry ideal fitting curve obtained in step (2). In combination with the alarm rules in step (4), illegal data points are marked and counted to evaluate the completion of the flight mission.

2. The intelligent flight monitoring method for a large UAV ground station according to claim 1 is characterized in that: Step (2) specifically includes the following steps: (2.1) Judge and verify the input source code data, then decode it and encapsulate it into universal remote control command and universal telemetry parameter object classes; (2.2) Preprocessing the encapsulated universal remote control command and universal telemetry parameter object class according to the rules set in step (1), including data standardization, smoothing filtering, resampling, feature extraction, anomaly detection and data clipping; (2.3) Extract features and correlation coefficients from the preprocessed data, perform multi-parameter visualization analysis, and perform regression and fitting to generate analysis results. Among them, after eliminating empty instructions and invalid data from the remote control data, the remote control operation during the entire flight mission is obtained, and the telemetry data is processed and analyzed to obtain the ideal fitting curve and trend perception prediction of the data.

3. According to claim 1, an intelligent flight monitoring method for a large UAV ground station, wherein: The data sources for data input and collection in step (2) are mainly divided into the following three types: 1) UAV measurement and control data received through UDP and TCP networks and local COM serial ports; 2) Original telecontrol and telemetry DAT data files; 3) Batch import of locally stored processed CSV and XML files; The flight monitoring method includes two modes: real-time mission mode and playback analysis mode. In mission mode, the measurement and control data comes from 1), and in playback analysis mode, it comes from 2) and 3).