CN108960303A - A kind of unmanned plane during flying data exception detection method based on LSTM - Google Patents
A kind of unmanned plane during flying data exception detection method based on LSTM Download PDFInfo
- Publication number
- CN108960303A CN108960303A CN201810639367.6A CN201810639367A CN108960303A CN 108960303 A CN108960303 A CN 108960303A CN 201810639367 A CN201810639367 A CN 201810639367A CN 108960303 A CN108960303 A CN 108960303A
- Authority
- CN
- China
- Prior art keywords
- lstm
- unmanned plane
- during flying
- plane during
- sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Traffic Control Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A kind of unmanned plane during flying data exception detection method based on LSTM, is related to unmanned plane abnormality detection and system health management domain.The present invention is the problem of function generated in the UAV system course of work is complex, and approximation capability is not able to satisfy higher fitting precision demand in order to solve in the detection of unmanned plane during flying data exception.The present invention reconstructs unmanned plane telemetry phase space, obtain input vector and output vector, obtain training sample set and test sample collection, LSTM fundamental forecasting model is built using TensorFlow deep learning Open Framework and carries out parameter and carries out optimizing, optimal L STM model is obtained and then calculates LSTM prediction result;It carries out outlier detection and unusual sequences detection respectively later, is finally completed the detection of unmanned plane during flying data exception.
Description
Technical field
The invention belongs to unmanned plane abnormality detections and system health management domain.
Background technique
Unmanned plane method for detecting abnormality can be divided mainly into three categories, and Knowledge based engineering method for detecting abnormality is based on physics mould
The method for detecting abnormality of type and method for detecting abnormality based on data-driven.First two method need to fully understand domain knowledge or
System structure can be only achieved preferable detection effect, and universality is poor, poor anti jamming capability, and can only generally detect known type
Abnormality detection.Unmanned plane telemetry exception label is few, and Exception Type is complicated, and is difficult to establish its subsystems unified
Physical model.
Method for detecting abnormality based on data-driven belongs to the scope of multi-variate statistical analysis, can be subdivided into based on similitude
Method, Statistics-Based Method, the method based on classification, the method based on prediction, method based on probability.Unmanned plane telemetering
The main form of flying quality is time series, and the method based on prediction is widely used in the abnormality detection of time series, is had
There is adaptivity, the abnormality detection of off-line data or online data can be carried out.Such methods are using the prediction result of data as base
Plinth, therefore outstanding detection effect depends on accurate effective prediction model, such as ARMA, SVM, neural network.But these sides
Method belongs to shallow-layer learning algorithm, for the more complicated function generated in the UAV system course of work, approximation capability
The higher fitting precision demand not being able to satisfy.In recent years, deep learning rapidly develops, this project is using deep learning field
LSTM method completes prediction model, and completes unmanned plane abnormality detection based on prediction result.
Summary of the invention
The present invention is to solve to generate in the UAV system course of work in the detection of unmanned plane during flying data exception
The problem of function is complex, and approximation capability is not able to satisfy higher fitting precision demand now provides a kind of based on LSTM's
Unmanned plane during flying data exception detection method.
A kind of unmanned plane during flying data exception detection method based on LSTM, including abnormal point detecting method and unusual sequences
Detection method:
Abnormal point detecting method specifically:
Step 1: reconstruct unmanned plane telemetry phase space obtains input vector and output vector, and building LSTM is substantially pre-
The training sample set and test sample collection of model are surveyed, and LSTM fundamental forecasting model is built based on training, test sample collection;
Step 2: carrying out optimizing to the parameter of LSTM fundamental forecasting model using grid data service, and by the ginseng after optimizing
Number, which substitutes into, obtains optimal L STM model in LSTM fundamental forecasting model;
Step 3: the sample that test sample is concentrated is input in optimal L STM model, obtains LSTM prediction result;
Step 4: it calculates training sample and concentrates the actual value of sample and the residual error of LSTM prediction result, by training sample set
The residual error average value of middle sample is as normal distribution center μ=mean (etraining), training sample is concentrated to the mark of sample residual
Quasi- difference is used as Variance of Normal Distribution σ=std (etraining), etrainingIndicate that training sample concentrates the residual error of sample, when confidence is general
When rate P=99%, confidence interval is obtained:
[μ-2.6·σ,μ+2.6·σ];
Step 5: judging whether unmanned plane during flying data to be detected belong to confidence interval, is that then the unmanned plane to be detected flies
Row data are normal point, and otherwise the unmanned plane during flying data to be detected are abnormal point;
Unusual sequences detection method specifically:
Step 6: all unmanned plane during flying data to be detected are divided into multiple time serieses sequentially in time, each
Time series contains n data, 6≤n≤15;
Step 7: when fiducial probability is greater than 99%, judge whether the number Q of abnormal point in each time series belongs to
[n-4, n] is that then the time series is abnormal time sequence, and otherwise the time series is normal time series.
The utility model has the advantages that a kind of unmanned plane during flying data exception detection method based on LSTM of the present invention, to nobody
It is more accurate to the detection effect of segment exception when machine flying quality carries out abnormality detection.Be for segment abnormal in Fig. 3 [303,
343] Unmanned Aerial Vehicle Data, the abnormal segment that the present invention eventually detects be [303,346], it is seen that this method detection effect compared with
Good, approximation capability can satisfy higher fitting precision demand.
Detailed description of the invention
Fig. 1 is a kind of flow chart of unmanned plane during flying data exception detection method based on LSTM;
Fig. 2 is LSTM schematic network structure;
Fig. 3 is in the presence of abnormal Unmanned Aerial Vehicle Data curve graph;
Fig. 4 is test set residual error curve and decision-making value curve graph;
Fig. 5 is that test set abnormality detection result marks curve graph;
1:LSTM node.
Specific embodiment
Present embodiment, a kind of unmanned plane during flying number based on LSTM described in present embodiment are illustrated referring to Fig.1
According to method for detecting abnormality,
Step 1: unmanned plane telemetry phase space reconfiguration.
The characteristics of to adapt to LSTM network structure, carries out phase space reconfiguration for the unmanned plane telemetry of one-dimensional first,
The training sample and test sample of input vector and output vector as LSTM are constructed, as shown in Equation 1.
Wherein X (t) is the input vector constructed from one-dimensional time series, x (t) in the form of time series existing for one-dimensional
Unmanned plane telemetry is in the value of t moment, and phase space reconfiguration length of window is D, and Y (t) is that output vector is X (t) corresponding
True value.What LSTM model needed to complete is exactly to the study of X (t) → Y (t) mapping relations.
Will beforeThe training sample set of a input vector and output vector as LSTM fundamental forecasting model, will be remainingThe test sample collection of a input vector and output vector as LSTM fundamental forecasting model, k indicate input vector and export to
The total number of amount.
Step 2: the training sample set and test sample collection obtained using step 1, using TensorFlow deep learning
Open Framework builds LSTM fundamental forecasting model.LSTM network structure is as shown in Figure 2.
Specifically include four main points:
A, confirmation input dimension.Input degree of enclosing is determined according to number of each moment to LSTM network inputs data point, and
Network is along time expansion number and inputs the product for degree of enclosing equal to phase space reconfiguration length of window.
B, the LSTM structural network number of plies is determined according to the complexity of unmanned plane telemetry and input dimension.This embodiment party
Formula uses three layers of LSTM structure, i.e. only one layer of concealed nodes, and too deep longitudinal number of plies is avoided to lead to over-fitting.
C, activation primitive is selected.Select linear function as activation primitive between hidden layer and output layer.
D, output layer designs.What LSTM was fitted is the mapping pass of [x (t), x (t-1), L, x (t- (D-1))] → x (t+1)
System, every one wheel data of input only correspond to one-dimensional output, therefore all moment do not export knot before the arrival of the last one input data
Fruit, i.e. cancellation network output layer, output layer normally exports prediction result when network last time is along time expansion.Training and test
Output vector in sample.
Step 3: carrying out optimizing to the parameter of LSTM fundamental forecasting model using grid data service, and by the ginseng after optimizing
Number, which substitutes into, obtains optimal L STM model in LSTM fundamental forecasting model.
In LSTM model, pre-set all multi-parameters may all generate final prediction effect and significantly affect.Its
In more crucial three parameters be respectively phase space reconfiguration length of window L, hidden layer node number N and learning rate η.Using net
Lattice searching method carries out optimizing to above-mentioned three kinds of parameters.Grid search refers to that these three parameters respectively ask for its different numerical value,
It intersects, forms the three dimensional search space of parameter.The setting value range for traversing three parameters, finally provides all parameters
Combined test set precision of prediction, and arranged according to sequence from high to low, the parameter combination to rank the first is optimized parameter
Combination.
Step 4: the sample that test sample is concentrated is input in optimal L STM model, obtains LSTM prediction result
Step 5: detection abnormal point.
Predicted value and true value can have residual error, be considered as the random noise that cannot further learn to refine.Random noise has
Several factors codetermine, and meet the normal distribution that mean value is 0 under normal circumstances.Calculate the actual value that training sample concentrates sample
With the residual error of LSTM prediction result, concentrate the residual error average value of sample as normal distribution center μ=mean training sample
(etraining), concentrate the standard deviation of sample residual as Variance of Normal Distribution σ=std (e training sampletraining),
etrainingIt indicates that training sample concentrates the residual error of sample, as fiducial probability P=99%, obtains confidence interval:
[μ-2.6·σ,μ+2.6·σ] (2)
Judge whether unmanned plane during flying data to be detected belong to confidence interval, be, which is
Normal point, otherwise the unmanned plane during flying data to be detected are abnormal point.
Step 6: detection unusual sequences.
All unmanned plane during flying data to be detected are divided into multiple time serieses, each time series sequentially in time
Contain n data, 6≤n≤15;When fiducial probability be greater than 99% when, judge abnormal point in each time series number Q whether
Belong to [n-4, n], is that then the time series is abnormal time sequence, otherwise the time series is normal time series.
Step 7: all abnormal time sequences are taken into union, obtain the abnormal point of all unmanned plane during flying data to be detected
Set.
The Unmanned Aerial Vehicle Data that present embodiment selection has abnormal segment carries out the recruitment evaluation of the method.Unmanned plane number
According to as shown in figure 3, there are segment exceptions in [303,343] section.Itd is proposed method is utilized, sets confidence level as 99%,
I.e. confidence interval is [+2.6 σ of μ -2.6 σ, μ], obtains confidence interval as shown in figure 4, residual error is beyond confidence interval part
Exceptional value.Finally, all abnormal points are marked.According to the method proposed, when degree for occur in 10 time series to
When few 6 exceptional values, this segment is abnormal segment, and finally obtained exception segment is [303,346], and detection effect is preferable, such as
Shown in Fig. 5.
Claims (5)
1. a kind of unmanned plane during flying data exception detection method based on LSTM, which is characterized in that including abnormal point detecting method
With unusual sequences detection method:
Abnormal point detecting method specifically:
Step 1: reconstruct unmanned plane telemetry phase space obtains input vector and output vector, constructs LSTM fundamental forecasting mould
The training sample set and test sample collection of type, and LSTM fundamental forecasting model is built based on training sample set and test sample collection;
Step 2: carrying out optimizing to the parameter of LSTM fundamental forecasting model using grid data service, and by the parameter generation after optimizing
Enter acquisition optimal L STM model in LSTM fundamental forecasting model;
Step 3: the sample that test sample is concentrated is input in optimal L STM model, obtains LSTM prediction result;
Step 4: it calculates training sample and concentrates the actual value of sample and the residual error of LSTM prediction result, training sample is concentrated into sample
This residual error average value is as normal distribution center μ=mean (etraining), training sample is concentrated to the standard deviation of sample residual
As Variance of Normal Distribution σ=std (etraining), etrainingIndicate that training sample concentrates the residual error of sample, as fiducial probability P
When=99%, confidence interval is obtained:
[μ-2.6·σ,μ+2.6·σ];
Step 5: judging whether unmanned plane during flying data to be detected belong to confidence interval, is the then unmanned plane during flying number to be detected
According to for normal point, otherwise the unmanned plane during flying data to be detected are abnormal point;
Unusual sequences detection method specifically:
Step 6: all unmanned plane during flying data to be detected are divided into multiple time serieses, each time sequentially in time
Sequence contains n data, 6≤n≤15;
Step 7: when fiducial probability is greater than 99%, judge whether the number Q of abnormal point in each time series belongs to [n-4,
N], it is that then the time series is abnormal time sequence, otherwise the time series is normal time series.
2. a kind of unmanned plane during flying data exception detection method based on LSTM according to claim 1, which is characterized in that
After step 7, all abnormal time sequences are taken into union, obtain the abnormal point set of all unmanned plane during flying data to be detected
It closes.
3. a kind of unmanned plane during flying data exception detection method based on LSTM according to claim 1 or 2, feature exist
In input vector X (t) and output vector Y (t) expression formula of unmanned plane telemetry phase space are as follows in step 1:
Wherein, x (t) be in the form of time series existing for one-dimensional unmanned plane telemetry t moment value, D be phase space weight
Structure length of window.
4. a kind of unmanned plane during flying data exception detection method based on LSTM according to claim 1 or 2, feature exist
In building the method for LSTM fundamental forecasting model in step 1 are as follows:
Input degree of enclosing determined to the number of LSTM network inputs data point according to each moment, and network along time expansion number with
The product for inputting degree of enclosing is equal to phase space reconfiguration length of window;
The LSTM structural network number of plies is determined according to the complexity of unmanned plane telemetry and input dimension;
Select linear function as activation primitive between hidden layer and output layer;
Output layer output prediction result when making LSTM network last time along time expansion.
5. a kind of unmanned plane during flying data exception detection method based on LSTM according to claim 1 or 2, feature exist
In, the parameter of the model of LSTM fundamental forecasting described in step 2 include: phase space reconfiguration length of window, hidden layer node number and
Learning rate.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810639367.6A CN108960303B (en) | 2018-06-20 | 2018-06-20 | Unmanned aerial vehicle flight data anomaly detection method based on LSTM |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810639367.6A CN108960303B (en) | 2018-06-20 | 2018-06-20 | Unmanned aerial vehicle flight data anomaly detection method based on LSTM |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108960303A true CN108960303A (en) | 2018-12-07 |
CN108960303B CN108960303B (en) | 2021-05-07 |
Family
ID=64490435
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810639367.6A Active CN108960303B (en) | 2018-06-20 | 2018-06-20 | Unmanned aerial vehicle flight data anomaly detection method based on LSTM |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108960303B (en) |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109711440A (en) * | 2018-12-13 | 2019-05-03 | 新奥数能科技有限公司 | A kind of data exception detection method and device |
CN109934337A (en) * | 2019-03-14 | 2019-06-25 | 哈尔滨工业大学 | A kind of detection method of the spacecraft telemetry exception based on integrated LSTM |
CN110458195A (en) * | 2019-07-10 | 2019-11-15 | 河海大学 | A kind of abnormal data discriminating method based on the fusion of more criterions |
CN110569925A (en) * | 2019-09-18 | 2019-12-13 | 南京领智数据科技有限公司 | LSTM-based time sequence abnormity detection method applied to electric power equipment operation detection |
CN110705658A (en) * | 2019-08-15 | 2020-01-17 | 哈尔滨工业大学 | Multivariable regression-based unmanned aerial vehicle flight parameter data fault detection and recovery method |
CN111241744A (en) * | 2020-01-07 | 2020-06-05 | 浙江大学 | Low-pressure casting machine time sequence data abnormity detection method based on bidirectional LSTM |
CN111260029A (en) * | 2020-01-13 | 2020-06-09 | 北京工业大学 | Credibility analysis method for air quality data |
CN111562996A (en) * | 2020-04-11 | 2020-08-21 | 北京交通大学 | Method and system for detecting time sequence abnormality of key performance index data |
CN111786738A (en) * | 2020-07-01 | 2020-10-16 | 中国人民解放军陆军工程大学 | Anti-interference learning network structure based on long-term and short-term memory and learning method |
CN112085866A (en) * | 2020-08-14 | 2020-12-15 | 陕西千山航空电子有限责任公司 | Airplane abnormal state identification method based on flight parameter data |
CN112306722A (en) * | 2019-07-26 | 2021-02-02 | 北京京东乾石科技有限公司 | Method, device, equipment and computer readable medium for identifying fault |
CN113218537A (en) * | 2021-05-25 | 2021-08-06 | 中国南方电网有限责任公司超高压输电公司广州局 | Training method, device, equipment and storage medium of temperature anomaly detection model |
CN114037012A (en) * | 2021-11-09 | 2022-02-11 | 四川大学 | Flight data anomaly detection method based on correlation analysis and deep learning |
CN116088349A (en) * | 2023-03-07 | 2023-05-09 | 东方空间技术(山东)有限公司 | LSTM-based rocket semi-physical simulation data anomaly evaluation method and decision device |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106055885A (en) * | 2016-05-26 | 2016-10-26 | 哈尔滨工业大学 | Anomaly detection method of flight data of unmanned aerial vehicle based on over-sampling projection approximation basis pursuit |
EP3261024A2 (en) * | 2016-06-23 | 2017-12-27 | Siemens Healthcare GmbH | Method and system for vascular disease detection using recurrent neural networks |
CN107748899A (en) * | 2017-11-03 | 2018-03-02 | 哈尔滨工业大学深圳研究生院 | A kind of target classification of the two dimensional image based on LSTM sentences knowledge method |
CN107833208A (en) * | 2017-10-27 | 2018-03-23 | 哈尔滨工业大学 | A kind of hyperspectral abnormity detection method based on changeable weight depth own coding |
CN108053653A (en) * | 2018-01-11 | 2018-05-18 | 广东蔚海数问大数据科技有限公司 | Vehicle behavior Forecasting Methodology and device based on LSTM |
CN108063698A (en) * | 2017-12-15 | 2018-05-22 | 东软集团股份有限公司 | Unit exception detection method and device, program product and storage medium |
-
2018
- 2018-06-20 CN CN201810639367.6A patent/CN108960303B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106055885A (en) * | 2016-05-26 | 2016-10-26 | 哈尔滨工业大学 | Anomaly detection method of flight data of unmanned aerial vehicle based on over-sampling projection approximation basis pursuit |
EP3261024A2 (en) * | 2016-06-23 | 2017-12-27 | Siemens Healthcare GmbH | Method and system for vascular disease detection using recurrent neural networks |
CN107833208A (en) * | 2017-10-27 | 2018-03-23 | 哈尔滨工业大学 | A kind of hyperspectral abnormity detection method based on changeable weight depth own coding |
CN107748899A (en) * | 2017-11-03 | 2018-03-02 | 哈尔滨工业大学深圳研究生院 | A kind of target classification of the two dimensional image based on LSTM sentences knowledge method |
CN108063698A (en) * | 2017-12-15 | 2018-05-22 | 东软集团股份有限公司 | Unit exception detection method and device, program product and storage medium |
CN108053653A (en) * | 2018-01-11 | 2018-05-18 | 广东蔚海数问大数据科技有限公司 | Vehicle behavior Forecasting Methodology and device based on LSTM |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109711440A (en) * | 2018-12-13 | 2019-05-03 | 新奥数能科技有限公司 | A kind of data exception detection method and device |
CN109934337A (en) * | 2019-03-14 | 2019-06-25 | 哈尔滨工业大学 | A kind of detection method of the spacecraft telemetry exception based on integrated LSTM |
CN110458195A (en) * | 2019-07-10 | 2019-11-15 | 河海大学 | A kind of abnormal data discriminating method based on the fusion of more criterions |
CN112306722A (en) * | 2019-07-26 | 2021-02-02 | 北京京东乾石科技有限公司 | Method, device, equipment and computer readable medium for identifying fault |
CN110705658A (en) * | 2019-08-15 | 2020-01-17 | 哈尔滨工业大学 | Multivariable regression-based unmanned aerial vehicle flight parameter data fault detection and recovery method |
CN110705658B (en) * | 2019-08-15 | 2022-03-15 | 哈尔滨工业大学 | Multivariable regression-based unmanned aerial vehicle flight parameter data fault detection and recovery method |
CN110569925A (en) * | 2019-09-18 | 2019-12-13 | 南京领智数据科技有限公司 | LSTM-based time sequence abnormity detection method applied to electric power equipment operation detection |
CN110569925B (en) * | 2019-09-18 | 2023-05-26 | 南京领智数据科技有限公司 | LSTM-based time sequence abnormality detection method applied to power equipment operation detection |
CN111241744A (en) * | 2020-01-07 | 2020-06-05 | 浙江大学 | Low-pressure casting machine time sequence data abnormity detection method based on bidirectional LSTM |
CN111241744B (en) * | 2020-01-07 | 2022-08-09 | 浙江大学 | Low-pressure casting machine time sequence data abnormity detection method based on bidirectional LSTM |
CN111260029A (en) * | 2020-01-13 | 2020-06-09 | 北京工业大学 | Credibility analysis method for air quality data |
CN111562996A (en) * | 2020-04-11 | 2020-08-21 | 北京交通大学 | Method and system for detecting time sequence abnormality of key performance index data |
CN111786738A (en) * | 2020-07-01 | 2020-10-16 | 中国人民解放军陆军工程大学 | Anti-interference learning network structure based on long-term and short-term memory and learning method |
CN111786738B (en) * | 2020-07-01 | 2022-06-03 | 中国人民解放军陆军工程大学 | Anti-interference learning network structure based on long-term and short-term memory and learning method |
CN112085866A (en) * | 2020-08-14 | 2020-12-15 | 陕西千山航空电子有限责任公司 | Airplane abnormal state identification method based on flight parameter data |
CN113218537A (en) * | 2021-05-25 | 2021-08-06 | 中国南方电网有限责任公司超高压输电公司广州局 | Training method, device, equipment and storage medium of temperature anomaly detection model |
CN113218537B (en) * | 2021-05-25 | 2024-04-05 | 中国南方电网有限责任公司超高压输电公司广州局 | Training method, training device, training equipment and training storage medium for temperature anomaly detection model |
CN114037012A (en) * | 2021-11-09 | 2022-02-11 | 四川大学 | Flight data anomaly detection method based on correlation analysis and deep learning |
CN114037012B (en) * | 2021-11-09 | 2023-04-07 | 四川大学 | Flight data anomaly detection method based on correlation analysis and deep learning |
CN116088349A (en) * | 2023-03-07 | 2023-05-09 | 东方空间技术(山东)有限公司 | LSTM-based rocket semi-physical simulation data anomaly evaluation method and decision device |
Also Published As
Publication number | Publication date |
---|---|
CN108960303B (en) | 2021-05-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108960303A (en) | A kind of unmanned plane during flying data exception detection method based on LSTM | |
CN109255505B (en) | Short-term load prediction method of multi-model fusion neural network | |
CN107220734A (en) | CNC Lathe Turning process Energy Consumption Prediction System based on decision tree | |
CN104820716B (en) | Equipment Reliability appraisal procedure based on data mining | |
CN109657945B (en) | Industrial production process fault diagnosis method based on data driving | |
CN108197648A (en) | A kind of Fault Diagnosis Method of Hydro-generating Unit and system based on LSTM deep learning models | |
CN109146162B (en) | A kind of probability wind speed forecasting method based on integrated Recognition with Recurrent Neural Network | |
CN111079836B (en) | Process data fault classification method based on pseudo label method and weak supervised learning | |
CN101576604B (en) | Method for diagnosing failures of analog circuit based on heterogeneous information fusion | |
CN107274011A (en) | The equipment state recognition methods of comprehensive Markov model and probability net | |
CN102622510A (en) | System and method for quantitative management of software defects | |
CN107704962B (en) | Steam flow interval prediction method based on incomplete training data set | |
CN103995237A (en) | Satellite power supply system online fault diagnosis method | |
CN104751229A (en) | Bearing fault diagnosis method capable of recovering missing data of back propagation neural network estimation values | |
CN111191726B (en) | Fault classification method based on weak supervision learning multilayer perceptron | |
CN103218660B (en) | A kind of airway selection method based on extensive fuzzy competition nerve net | |
CN107290965A (en) | Adaptive soft-sensor Forecasting Methodology based on local weighted Bayesian network | |
CN108830325A (en) | A kind of vibration information classification of landform recognition methods based on study | |
CN106372724A (en) | Artificial neural network algorithm | |
CN111488946A (en) | Radar servo system fault diagnosis method based on information fusion | |
CN108133090A (en) | A kind of high-end complex equipment analysis method for reliability of reliability susceptibility driving | |
CN103678886B (en) | A kind of satellite Bayesian network health based on ground test data determines method | |
CN109829627A (en) | A kind of safe confidence appraisal procedure of Electrical Power System Dynamic based on integrated study scheme | |
Akinci et al. | Wind speed correlation between neighboring measuring stations | |
CN104656620A (en) | Comprehensive evaluation system for remanufacturing of heavy-duty machine tool |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |