CN104569694B - Electric signal feature extraction and recognition system oriented to aircraft flying process - Google Patents
Electric signal feature extraction and recognition system oriented to aircraft flying process Download PDFInfo
- Publication number
- CN104569694B CN104569694B CN201510043740.8A CN201510043740A CN104569694B CN 104569694 B CN104569694 B CN 104569694B CN 201510043740 A CN201510043740 A CN 201510043740A CN 104569694 B CN104569694 B CN 104569694B
- Authority
- CN
- China
- Prior art keywords
- signal
- sensor
- extracted
- aircraft
- signal characteristics
- 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.)
- Active
Links
Landscapes
- Complex Calculations (AREA)
- Arrangements For Transmission Of Measured Signals (AREA)
Abstract
The invention discloses an electric signal feature extraction and recognition system oriented to aircraft flying process, particularly relates to a collection, processing, extraction and recognition system for electric signals of an aircraft. The system takes the electric signals of the aircraft as an object, is based on the process oriented to flying tasks and is suitable for convention testing of the aircraft at each stage and in various modes and combined testing of multiple aircrafts. By the system, dynamic data, which are more comprehensive and are based on the flying process, of the electric signals can be accumulated at a comprehensive testing stage and serve as a criterion to judge whether electric performance of the aircraft is normal or not, and the dynamic data exceed corresponding remote measuring indicators of a tested object in the aspects of accuracy and instantaneity; as a key link for verifying the aircraft, compared with conventional remote measuring of the aircraft, monitoring accuracy and rate of the system are improved remarkably on a number level, an efficient means is provided for comprehensive testing of the aircraft oriented for complex flying process, and level of judging electric performance working state of the aircraft on a system level is improved.
Description
Technical field
It is particularly a kind of towards spacecraft flight process the present invention relates to a kind of signal characteristics are extracted and identifying system
Signal characteristics are extracted and identifying system, and in particular to the collection of a set of spacecraft electric signal, processs, extraction and identifying system,
Suitable for spacecraft each stage and the conventionally test of various patterns, and the joint test with many spacecrafts.
Background technology
To complete the space mission of arduousness, the application of assembly spacecraft is more prevalent, and the ground connection between each assembly
Cyberrelationship is to ensure that the normal key element of electrical property, and assembly spacecraft reference potential refers to internal each primary power source list
Level between point earth point.Because each equipment has itself certain reference potential thresholding on spacecraft, by benchmark
The monitoring of current potential can in real time reflect the normality of each equipment work.Simultaneously by the calculating to resistance between earth point,
Each device current backflow analysis, and supply network backflow Orientation can be carried out.In conventional spacecraft-testing, not
Having carries out the means of real-time monitoring to reference potential, for the test of earth point is only limitted to the measurement of resistance between earth point, nothing
Method is grounded the real-time monitoring of dotted state, combined resistance value between earth point after can only reflecting single line resistance and combining；
The feature extraction of electric signal and technology of identification are the basic problems of Space Vehicle System, when spacecraft is in work process
When opening electrical equipment, its stability of a system will face the challenge, i.e., the power source bus between equipment by sharing influences each other,
The curent change of some equipment can cause the change of busbar voltage, this cyclically-varying or step mutation to make bus electricity
It is pressed in some frequency ranges and vibration, this exactly spacecraft electricity consumption problem encountered that influences each other occurs.In the past for spacecraft
The monitoring of electricity consumption is only limitted to the voltage monitoring of bus, when fluctuating occurs in busbar voltage, some can not be in real time corresponded to completely
Equipment state changes.In the face of the Complex Spacecraft combined by multi cabin, the scheme before foundation only carries out voltage to bus
Monitoring, can not adapt to the application towards the full working flight process of spacecraft, it is necessary to study new means of testing and work as to meet
Front complex combination type spacecraft and the demand of mass data analysis；
During aerial mission, data management subsystem needs substantial amounts of telemetry parameter and provides for in-orbit flight spacecraft
Important data are supported and ensured, in the integration test stage, it is necessary to which telemetry parameter state is monitored.In the past for remote measurement ginseng
Several tests is to rely on telemetering channel and carries out, and because telemetering channel transmission data is slower, such test can only be feature
Test, for the change of telemetry parameter transient state cannot be captured, and by the real-time monitoring to telemetry parameter, can replace remote measurement to lead to
Road captures substantial amounts of telemetry and is analyzed and judges, it is also possible to verify the correctness that telemetering channel is transmitted.
The mode of operation complexity of spacecraft is various, and the structure of combination and various mode of operation cause the electrical characteristics of spacecraft
It is particularly critical, therefore in integration test and stage in orbit, for spacecraft is extracted towards the signal characteristics of flight course
It is most important with recognizing.At present, China's spacecraft substantially belongs to blank field for the reference potential monitoring of Complex Spacecraft；It is right
Voltage is concentrated mainly in the monitoring of spacecraft electric signal characteristic, by transient responses such as the ripples of oscillograph observation signal.Make
For contact type measurement, the signal input negative terminal of measuring apparatus is directly connected to tested spacecraft negative terminal, and measurement result is easily by tested
Object is disturbed, and can only do pattern monitoring to tested spacecraft, it is impossible to realize data storage, figure shows and signal transacting
Concurrent working, test spectrum analysis ability is weak, and the species of the signal transacting that can be selected to is extremely limited with function；For distant
The test of survey parameter relies on substantially telemetering channel to be carried out, and due to the restriction of transmission speed, can only be confined to functional test, nothing
Method captures the transient changing of telemetry parameter.
The content of the invention
Present invention solves the technical problem that being：Overcome the deficiencies in the prior art, there is provided one kind is towards spacecraft flight mistake
The signal characteristics of journey are extracted and identifying system, and using contactless mode the electric signal of spacecraft is obtained, and are adopted using drop
The mode of sample is carried out being transferred to software processing module through fiber optical transceiver after data sampling and carries out the identification of flight course, maximum
The demand of extraction and the identification of spacecraft flight process signal characteristics is met in degree.
The present invention technical solution be：A kind of signal characteristics towards spacecraft flight process extract and recognize system
System, including：Sensor assembly, I-V conversion modules, Isolation Amplifier Module, isolation matching module, data acquisition module, the first light
Fine transceiver, the second fiber optical transceiver and software processing module；
The sensor assembly includes current sensor, reference potential sensor, voltage sensor, remote measurement voltage sensor
With telemeter communication sensor, current sensor, reference potential sensor, voltage sensor, remote measurement voltage sensor and remote measurement are led to
News sensor obtains respectively the current signal of spacecraft, reference potential signal, voltage signal, remote measurement voltage signal and telemeter communication
Signal；
Current signal is converted into being transferred to data acquisition module after voltage signal and carries out data signal and adopt by I-V conversion modules
Collect, isolation amplification/matching module by reference potential signal, voltage signal and remote measurement voltage signal changed so as to fit in
The input range of acquisition module, and realize the isolation matching conversion of telemeter communication signal；
Data acquisition module is connected by the first fiber optical transceiver and the second fiber optical transceiver with master controller, will be collected
Signal transmission processed to software processing module；
The software processing module is pre-processed using down-sampled and digital filtering method to the electric signal for receiving
Afterwards signal characteristics are extracted using continuous wavelet transform, if the data for receiving are to receive first, using the electric signal for extracting
Feature enters the cluster of horizontal electrical signal, and by the corresponding signal characteristics of each cluster result for obtaining and the emulation data being previously implanted
Stored as expertise；If the data for receiving are non-receiving first, using the signal characteristics for extracting to electric signal
Classified, using it is unidentified go out electric signal corresponding to flight course and the signal characteristics that extract enter as expertise
Row storage, the electrical signal types that output is identified simultaneously are stored.
The data acquisition module carries out digital signal acquiring using down-sampled method.
The current sensor and telemeter communication sensor obtain signal using contactless method.
The reference potential sensor, voltage sensor and remote measurement voltage sensor obtain letter using the method for light-coupled isolation
Number.
The clustering algorithm of the software processing module is K-means clustering algorithms.
Transformation parameter K in the K-means algorithms is determined that threshold value d is 0.4 by known state.
Sorting algorithm is K- nearest-neighbors method and support vector machine method in the software processing module.
The range of choice of K values is in the K- nearest-neighbors method：30～40.
The penalty factor of the SVMs takes respectively different from parameter γ of kernel function between 0.0l to 100
The classification accuracy rate of SVMs, takes classification accuracy rate highest (C, γ) under value, relatively different (C, γ)) as support to
The parameter of amount machine.
Compared with the prior art, the invention has the advantages that：
(1) present invention connects data acquisition module and software processing module using fiber optical transceiver, it is to avoid ground survey
Circuit excessively obtains " signal energy " from spacecraft in circuit, and avoids interfering with each other between multiple TCH test channels；
(2) sensor in the present invention using non-contacting mode obtains spacecraft electric signal, it is to avoid measuring circuit is given
Circuit increases common-mode voltage, series mode voltage, conducting, leakage branch road on spacecraft；
(3) present invention carries out Feature extraction and recognition for the electrical characteristic data of spacecraft Dynamic Signal first, using straight
Connect the technological means such as time domain extraction, frequency domain character are extracted and wavelet coefficient is extracted and complete electrical characteristics extraction, calculated based on Bayes
The standard vector ownership of method identification measured signal, effective integration signal transacting, classification, cluster and SVMs technology are solved
During complex task, the problem that feature extraction is difficult, identification is difficult is associated after electrical characteristics signal coupling excitation, with traditional algorithm phase
Than accuracy of identification improves 15%, in the leading level in the world, with great military affairs, economic benefit and promotional value.
Description of the drawings
Fig. 1 is the system block diagram of the present invention；
Fig. 2 is the system connection diagram in the present invention；
Fig. 3 is the software processing flow figure in the present invention.
Specific embodiment
The specific embodiment of the present invention is further described in detail below in conjunction with the accompanying drawings.
The system block diagram of the present invention is illustrated in figure 1, Fig. 2 is the system connection diagram in the present invention；From Fig. 1 and Fig. 2
Understand, a kind of signal characteristics towards spacecraft flight process proposed by the present invention are extracted and identifying system, including：Sensor
Module, I-V conversion modules, isolation matching/amplification module, data acquisition module, the first fiber optical transceiver, the second fiber optical transceiver
And software processing module；
The sensor assembly includes current sensor (LEM IT 60-S ULTRASTAB), reference potential sensor (NI
9220-24bit), voltage sensor (NI 9220-16bit), remote measurement voltage sensor (NI 9220-16bit) and telemeter communication
Sensor (DDC 61580, B3226), current sensor, reference potential sensor, voltage sensor, remote measurement voltage sensor and
Telemeter communication sensor obtains respectively the current signal of spacecraft, reference potential signal, voltage signal, remote measurement voltage signal and distant
Survey communication signal；Current sensor and telemeter communication sensor obtain signal, reference potential sensing using contactless method
Device, voltage sensor and remote measurement voltage sensor obtain signal using the method for light-coupled isolation；
I-V conversion modules (LEM IT 60-S ULTRASTAB) are converted into current signal to be transferred to number after voltage signal
Digital signal acquiring is carried out according to acquisition module (NI 9138), Isolation Amplifier Module (ANALOG DEVICE AD202) is electric by benchmark
Position signal, voltage signal and remote measurement voltage signal are carried out being transferred to data acquisition module (NI 9138) after isolation amplification and enter line number
Word signals collecting, isolation matching module (ANALOG DEVICE AD202) carries out telemeter communication signal to transmit after isolation matching
Digital signal acquiring is carried out to data acquisition module (NI 9138)；The data acquisition module is carried out using down-sampled method
Digital signal acquiring.
Data acquisition module (NI 9138) is by the first fiber optical transceiver (SICOM3424P) and the second fiber optical transceiver
(SICOM3424P) it is connected with software processing module, the signal transmission for collecting is processed to software processing module.
Reference potential monitoring device is mainly used in the measurement of each subsystem benchmark level point of spacecraft.Connect with mainly including device
Oral area point, isolation amplifier section, data acquiring portion.The device can include 16 passages, and device earth signal interface section passes through
Each equipment ground point in single device is introduced monitoring device by the long low-heat potential band spade lug plate cables of 3m；Isolation enlarging section
Sorting with the Isolation Amplifier Module of 16 maturations realize 16 passage -100mV～100mV reference voltage signals be converted into -10V～
The voltage signal of 10V；Reference voltage signal is converted into data signal and is easy to pass by data acquiring portion by general collecting device
It is defeated.Isolation amplification is powered with data acquiring portion using unified battery；
Current monitoring device is by the Signal Matching to each load supplying input/output interface of spacecraft, current-voltage
Conversion, finally carry out voltage measurement, realize load, the real-time monitoring of supply input output interface electric current and signature analysis.Should
Device includes 16 passages, and front end at equipment power inlet end by sealing in the long cable of 20cm with ring-type current sensors
Line, monitoring device is incorporated into by ring-type current sensors output end by current measurement signal；Then by I-V conversion sections,
Ring-type current sensors output current signal is converted into by precision resister to be suitable for the signal that general collecting device is gathered；
Finally signal is sent into into general number and adopt equipment, realize the collection of signal.Isolation is amplified with data acquiring portion using unified battery
Power supply；
Voltage monitor is by the monitoring to each load supplying interface voltage of spacecraft, including star ground interface, isolation
Amplify, the part such as data acquisition is realized to equipment working state monitoring and the early warning and warning of failure.The device is logical including 16
Road, monitoring device is introduced in spacecraft proximally by flip-flop cable by load voltage signal；Then amplified by isolation and realized
So as to fit in the input range of general collecting device, isolation amplification is put from general for the isolation conversion of each voltage signal
Big module realizes the isolation conversion of -50V～50V to -10V～10V signals；General several device for picking are finally adopted, signal is realized
Transmission with storage.Isolation amplification is powered with data acquiring portion using unified battery；
Remote measurement voltage monitor is by the monitoring to each power supply interface voltage of spacecraft, including star ground interface, isolation
Amplify, the part such as data acquisition is realized to equipment working state monitoring and the early warning and warning of failure.The device is logical including 16
Road, monitoring device is introduced in spacecraft proximally by flip-flop cable by remote measurement voltage signal；Then amplified by isolation and realized
So as to fit in the input range of general collecting device, isolation is amplified from general for the isolation conversion of each remote measurement voltage signal
Amplification module realize 0V～5V to 0V～10V signals isolation conversion；General several device for picking are finally adopted, signal is realized
Transmission and storage.Isolation amplification is powered with data acquiring portion using unified battery.
Telemeter communication link monitoring device mainly realizes the monitoring of remote measurement subcarrier signal waveform and specificity analysis, including
Amplify with circuit, isolation, the part such as data acquisition realizes to telemeter communication link working state, the device is provided and communicated all the way
Link monitoring passage.
Fig. 3 is the software processing flow figure in the present invention.As can be seen from Figure 3, the processing procedure of processing module has in the present invention
Body is as follows：
(1) pretreatment of electric signal, is premise and the basis that can reach millivolt level monitoring accuracy of the invention.Data signal
Pretreatment mainly includes three parts, is respectively：Down-sampled part, digital filtering part, electric signal event capturing part；
Down-sampled part：Because hardware acquisition system has a very strong acquisition capacity, the sample rate of electric signal is 10KHz, band
Wide 0～400Hz, therefore ONLINE RECOGNITION system carries out one on the premise of identification accuracy is had substantially no effect on to acquired data
Individual down-sampled process.So as to the data of calculating needed for reducing, amount of calculation is greatly reduced, it is down-sampled to enter according to actual needs
Row is adjusted.
Digital filtering part：During electric signal event recognition, the data for being collected may also have substantial amounts of making an uproar
Sound, in order to improve the accuracy of identification, digital filtering is essential.Herein we are used by the way of wavelet threshold denoising
The method of wavelet decomposition, goes to be reconstructed respectively after primary signal low-and high-frequency is separated, and has preferable effect to being mutated noise.Threshold
In value denoising, threshold function table is embodied to the wavelet coefficient different disposal strategy above and below threshold value, is closed in threshold denoising
One step of key.For electric signal, the process of wavelet threshold denoising is：First electric signal is resolved into into high frequency and low frequency two parts, point
Wavelet transformation is carried out to gained signal after solution, required denoising threshold can be calculated by the result of decomposition and wavelet transformation
Value, the wavelet coefficient after hard threshold function or soft-threshold denoising function pair wavelet transformation carries out denoising.Finally will
Wavelet coefficient after denoising is reconstructed, so as to obtain denoising after signal.
Electric signal event capturing part：The event of electric signal event generation, Zhi Houtong are obtained by way of threshold triggers
The window function for crossing a fixed size captures the electrical signal data of a period of time.The effect of capture portion is after the event occurred
Complete electric signal waveform is obtained, the length of window function is secured, has unified standard, the identification operation after this ensure that is only
Carry out in the case where event occurs.
(2) signal characteristics extraction is carried out to pretreated signal in step (1), in the present system, is become using small echo
The feature extracting method for changing.Mathematically, small echo is the elementary cell of constructed fuction orthogonal space base, is had in energy
The function of enabled condition is met in limit space L2 (R), is said from the angle of signal transacting, frequency division when small echo (conversion) is strong
Analysis (process) instrument, develops on the basis of Fourier transform shortcoming is overcome.
Wavelet transformation has the advantages that multiple dimensioned/many resolution, can be by thick and thin ground process signal.While calculating speed
Hurry up, computation complexity is relatively low, and relevance grade is wide, it is adaptable to non-stationary signal, decomposes and reconstruct has good fidelity.It is little
Wave conversion is divided into two big class：Wavelet transform (DWT) and continuous wavelet conversion (CWT).Both differs primarily in that, even
Continuous conversion is operated in all possible zooming and panning, and discrete transform is using the particular subset of all zooming and panning values.
In the present system in order to obtain all of feature, changed using continuous wavelet.
The electric signal that wavelet transformation is really input into carries out the result acquired in convolution algorithm, the result with mother wavelet function
For wavelet coefficient, sign is the electric signal of input and the similarity of small echo.Feature is extracted with wavelet transformation unified feature
Standard, while having the preferable free degree, the present invention uses DB4 small echos as wavelet transform function, carries out three layers of wavelet transformation,
Time domain, the mutation of frequency domain electric signal, the feature extraction of step are realized, the accuracy classified afterwards is favorably improved.
(3) if the data for receiving are to receive first, into step (4), otherwise, into step (5)；
(4) electric signal is clustered using the signal characteristics extracted in step (2), it is therefore an objective to find one species electricity
The center of signal, is both the common form for finding such electric signal.Cluster result is obtained, each cluster result for obtaining is corresponding
Signal characteristics and the emulation data being previously implanted are stored as expertise；
Carrying out clustering processing by the raw electrical signal data to early stage storage can obtain standard library module, be that system is accurate
Standby pre-knowledge, is the basis that can be identified to electric signal.
Cluster analysis is the analysis to some features, based on similitude, is characterized with distance.Spy in a cluster
Than having more similitudes between feature not in same cluster between levying.The representative method of partitioning has K-means, K-
Medoids algorithms.Central point is taken as the mean value of all data points in current class, the algorithm can be more in K-means
Cluster centre is accurately obtained, so as to obtain the standard of each classification.To choose from current class in K-medoids algorithms to
Other all points (in current class) apart from sum minimum as central point.K-means algorithms are easily subject to some dirty numbers
According to impact, dirty data refers to the data that there is larger difference in same category, and K-medoids algorithms overcome K-means calculations
Method is poor to the classification effect with deviation point, but amount of calculation considerable influence computational efficiency, therefore, the system is selected
K-means algorithms.This algorithm with classification inside similarity highest, the minimum criterion of the direct similarity of classification, here similar
Degree had both referred to the distance between sample.
The idiographic flow of K-means algorithms is as follows：An object is arbitrarily selected to make respectively from k classification of artificial classification
For the initial cluster center of the category；
A) according to the average (the cluster centre object of the category) of each clustering object, each object and these centers are calculated
The distance of object；And according to minimum range corresponding object is divided again；
B) average (being set to cluster centre object) of each (changing) cluster is recalculated；
C) when certain condition, such as function convergence is met (cluster centre object no longer changes), then algorithm terminates；Such as
Really bar part is unsatisfactory for, and returns to step b).
The selection of K-means algorithmic transformations parameter K and threshold value d all has crucial impact, the present invention for classification results
Transformation parameter K is defined by known state, and threshold value d is set to 0.4.
Signal to the distance of different cluster centres routinely selects Euclidean distance, but in the present invention, uses Hamming distance
From.Hamming distance is more suitable for for data signal, and Hamming distance is otherwise known as signal distance, in electric signal identification, often
There are some errors, therefore the signal in certain error can be integrated.It is now very suitable using Hamming distance, while
Can be exported with the similarity that is used for of maximum Hamming using Hamming distance.
(5) electric signal is classified using the signal characteristics extracted in step (2), obtains the similarity letter of electric signal
Numerical value.In the conventional system, mainly used in the present system with the Naive Bayes Classification Algorithm based on temporal signatures
K- nearest neighbor algorithms and SVMs.
K- nearest-neighbors methods (KNN algorithms) are the methods that the immediate training sample in feature space is classified.
Classified using vector space model, the mutual similarity of case of identical category is high, by the phase calculated with known class case
The possible classification of unknown classification case is assessed like degree.Similarity refers to distance or vector similarity, if a sample
Great majority in the sample of k most like (i.e. closest in feature space) in feature space belong to some classification, then
The sample falls within this classification.In KNN algorithms, selected neighbours are the objects correctly classified.The method is fixed
The classification belonging to sample to be divided only is determined in class decision-making according to the classification of one or several closest samples.Due to KNN
Algorithm mainly by neighbouring sample limited around, rather than by differentiating that the method for class field is determining generic therefore right
For the more sample set to be divided of the intersection or overlap of class field, KNN algorithms are more suitable for compared with additive method.The side of true defining K value
Method first takes a relatively low initial value, and then basis takes in practice impact of the different K values to classification results and is adjusted.According to this
From the point of view of the implementation result of invention, spacecraft K values usually select proper between 30 to 40.
Support vector machine method is built upon on the VC of Statistical Learning Theory dimensions theory and Structural risk minization basis
, according to limited sample information is in the complexity (i.e. to the study precision of specific training sample) of model and learning ability is (i.e.
Recognize the ability of arbitrary sample without error) between seek optimal compromise, in the hope of obtaining best Generalization Ability.SVMs
By in the space of maps feature vectors to a more higher-dimension, a largest interval hyperplane is set up in this space.Separating
The both sides of the hyperplane of data have two hyperplane parallel to each other, and separating hyperplane makes the distance of two parallel hyperplane most
Bigization.It is assumed that distance or gap between parallel hyperplane are bigger, the overall error of grader is less.SVMs is typically adopted,
Man-to-man identification tactic.In this strategy, it is required for training a grader between any two class in N class samples, therefore needs N altogether
(N-1)/2 SVM classifier.The classification that highest poll is obtained after classification is both classification results.The penalty factor of SVMs
Different values are taken between 0.0l to 100 respectively from parameter γ of kernel function, SVMs divides relatively under different (C, γ)
Class accuracy, takes classification accuracy rate highest (C, γ)) as the parameter of SVMs, using the state recognition of spacecraft.
(6) if similarity function value be less than default threshold value, it is unidentified go out electric signal, the flight corresponding to the electric signal
Process is not recorded, and the flight course corresponding to the electric signal and the signal characteristics for extracting are deposited as expertise
Storage, if similarity function value is more than or equal to default threshold value, identifies electric signal, exports electrical signal types and is stored.
The content not being described in detail in description of the invention belongs to the known technology of professional and technical personnel in the field.
Claims (9)
1. a kind of signal characteristics towards spacecraft flight process are extracted and identifying system, it is characterised in that included：Sensor
Module, I-V conversion modules, Isolation Amplifier Module, isolation matching module, data acquisition module, the first fiber optical transceiver, the second light
Fine transceiver and software processing module；
The sensor assembly includes current sensor, reference potential sensor, voltage sensor, remote measurement voltage sensor and distant
Communication sensor is surveyed, current sensor, reference potential sensor, voltage sensor, remote measurement voltage sensor and telemeter communication are passed
Sensor obtains respectively the current signal of spacecraft, reference potential signal, voltage signal, remote measurement voltage signal and telemeter communication letter
Number；
Current signal is converted into being transferred to data acquisition module after voltage signal by I-V conversion modules carries out digital signal acquiring,
Isolation Amplifier Module carries out reference potential signal, voltage signal and remote measurement voltage signal to be transferred to data acquisition after isolation amplification
Module carries out digital signal acquiring, and isolation matching module carries out telemeter communication signal to be transferred to data acquisition module after isolation matching
Block carries out digital signal acquiring；
Data acquisition module is connected by the first fiber optical transceiver and the second fiber optical transceiver with software processing module, will be collected
Signal transmission processed to software processing module；
The software processing module is sharp after being pre-processed to the electric signal for receiving using down-sampled and digital filtering method
Signal characteristics are extracted with continuous wavelet transform, if the data for receiving are to receive first, using the signal characteristics for extracting
Enter the cluster of horizontal electrical signal, and using the corresponding signal characteristics of each cluster result for obtaining and the emulation data being previously implanted as
Expertise is stored；If the data for receiving are non-receiving first, electric signal is carried out using the signal characteristics for extracting
Classification, using it is unidentified go out electric signal corresponding to flight course and the signal characteristics that extract deposited as expertise
Storage, the electrical signal types that output is identified simultaneously are stored.
2. a kind of signal characteristics towards spacecraft flight process according to claim 1 are extracted and identifying system, its
It is characterised by：The data acquisition module carries out digital signal acquiring using down-sampled method.
3. a kind of signal characteristics towards spacecraft flight process according to claim 1 are extracted and identifying system, its
It is characterised by：The current sensor and telemeter communication sensor obtain signal using contactless method.
4. a kind of signal characteristics towards spacecraft flight process according to claim 1 are extracted and identifying system, its
It is characterised by：The reference potential sensor, voltage sensor and remote measurement voltage sensor are obtained using the method for light-coupled isolation
Signal.
5. a kind of signal characteristics towards spacecraft flight process according to claim 1 are extracted and identifying system, its
It is characterised by：The clustering algorithm of the software processing module is K-means clustering algorithms.
6. a kind of signal characteristics towards spacecraft flight process according to claim 5 are extracted and identifying system, its
It is characterised by：Transformation parameter K in the K-means algorithms is determined that threshold value d is 0.4 by known state.
7. a kind of signal characteristics towards spacecraft flight process according to claim 1 are extracted and identifying system, its
It is characterised by：Sorting algorithm is K- nearest-neighbors method and support vector machine method in the software processing module.
8. a kind of signal characteristics towards spacecraft flight process according to claim 7 are extracted and identifying system, its
It is characterised by：The range of choice of K values is in the K- nearest-neighbors method：30～40.
9. a kind of signal characteristics towards spacecraft flight process according to claim 7 are extracted and identifying system, its
It is characterised by：The penalty factor of the SVMs takes respectively different from parameter γ of kernel function between 0.0l to 100
The classification accuracy rate of SVMs under value, relatively different (C, γ), take classification accuracy rate highest (C, γ) as support to
The parameter of amount machine.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510043740.8A CN104569694B (en) | 2015-01-28 | 2015-01-28 | Electric signal feature extraction and recognition system oriented to aircraft flying process |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510043740.8A CN104569694B (en) | 2015-01-28 | 2015-01-28 | Electric signal feature extraction and recognition system oriented to aircraft flying process |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104569694A CN104569694A (en) | 2015-04-29 |
| CN104569694B true CN104569694B (en) | 2017-05-10 |
Family
ID=53086228
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510043740.8A Active CN104569694B (en) | 2015-01-28 | 2015-01-28 | Electric signal feature extraction and recognition system oriented to aircraft flying process |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104569694B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106446464B (en) * | 2016-11-08 | 2019-05-24 | 上海宇航系统工程研究所 | Visual spacecraft launching site margin of error Density functional calculations method |
| CN108418728B (en) * | 2018-02-12 | 2019-02-26 | 北京空间技术研制试验中心 | A kind of system detection method for spacecraft launching site aerial mission |
| CN109034250A (en) * | 2018-07-31 | 2018-12-18 | 佛山科学技术学院 | A kind of information representation system of spacecraft monitoring big data |
| CN109298300A (en) * | 2018-12-04 | 2019-02-01 | 武汉康电电气有限公司 | One kind being used for part discharge test system and method when high-tension cable pressure test |
| CN109962742B (en) * | 2019-03-29 | 2021-09-03 | 西安工业大学 | Portable telemetering data monitoring platform and monitoring method |
| CN110058133B (en) * | 2019-04-15 | 2021-03-02 | 杭州拓深科技有限公司 | Feedback mechanism-based electric circuit fault arc false alarm optimization method |
| CN111104374B (en) * | 2019-12-25 | 2023-04-07 | 中国航空工业集团公司沈阳飞机设计研究所 | Actually measured vibration data processing method and system |
| CN111680748B (en) * | 2020-06-08 | 2024-02-02 | 中国人民解放军63920部队 | Spacecraft state mode identification method and identification device |
| CN112487362B (en) * | 2020-12-03 | 2022-12-23 | 上海卫星工程研究所 | Satellite step remote parameter stability monitoring method and system based on K-Means + + algorithm |
| CN114506471B (en) * | 2022-03-17 | 2023-06-20 | 北京机电工程研究所 | First-order bending frequency determining method for aircraft suspension |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101619989A (en) * | 2008-07-04 | 2010-01-06 | 中国铁路通信信号上海工程有限公司 | System and method for acquiring and analyzing remote data of bridge |
| CN102033546A (en) * | 2010-11-09 | 2011-04-27 | 上海交通大学 | Low-altitude airship flight control system and flight control method thereof |
| CN102147967A (en) * | 2010-02-10 | 2011-08-10 | 上海卫星工程研究所 | Satellite data acquisition and transmission method |
-
2015
- 2015-01-28 CN CN201510043740.8A patent/CN104569694B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101619989A (en) * | 2008-07-04 | 2010-01-06 | 中国铁路通信信号上海工程有限公司 | System and method for acquiring and analyzing remote data of bridge |
| CN102147967A (en) * | 2010-02-10 | 2011-08-10 | 上海卫星工程研究所 | Satellite data acquisition and transmission method |
| CN102033546A (en) * | 2010-11-09 | 2011-04-27 | 上海交通大学 | Low-altitude airship flight control system and flight control method thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104569694A (en) | 2015-04-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104569694B (en) | Electric signal feature extraction and recognition system oriented to aircraft flying process | |
| CN105258947B (en) | A kind of Fault Diagnosis of Roller Bearings under operating mode disturbed conditions based on compressed sensing | |
| CN103728551B (en) | A kind of analog-circuit fault diagnosis method based on cascade integrated classifier | |
| CN109974782B (en) | Equipment fault early warning method and system based on big data sensitive characteristic optimization selection | |
| CN105841961A (en) | Bearing fault diagnosis method based on Morlet wavelet transformation and convolutional neural network | |
| CN105956623A (en) | Epilepsy electroencephalogram signal classification method based on fuzzy entropy | |
| CN105204493B (en) | A kind of rotating machinery condition monitoring and fault diagnosis method | |
| CN111062250A (en) | Multi-subject motor imagery electroencephalogram signal identification method based on depth feature learning | |
| CN104021238A (en) | Lead-acid power battery system fault diagnosis method | |
| CN110298085A (en) | Analog-circuit fault diagnosis method based on XGBoost and random forests algorithm | |
| CN112307950A (en) | Method for extracting and intelligently analyzing detail characteristic quantity of GIS vibration defect identification | |
| CN105701470A (en) | Analog circuit fault characteristic extraction method based on optimal wavelet packet decomposition | |
| CN110443117B (en) | Wind turbine generator fault diagnosis method | |
| CN110353673A (en) | A kind of brain electric channel selection method based on standard mutual information | |
| CN111678699B (en) | Early fault monitoring and diagnosing method and system for rolling bearing | |
| CN103558519A (en) | GIS partial discharge ultrasonic signal identification method | |
| CN103323228A (en) | Mining drill fault intelligent identification method | |
| CN116894187A (en) | Gear box fault diagnosis method based on deep migration learning | |
| CN103267652B (en) | Intelligent online diagnosis method for early failures of equipment | |
| CN216848010U (en) | Cable partial discharge online monitoring device for edge calculation | |
| CN111259949A (en) | Fault identification model construction method, model and identification method for aircraft environmental control system | |
| CN111310719B (en) | Unknown radiation source individual identification and detection method | |
| Xue et al. | A novel framework for motor bearing fault diagnosis based on multi-transformation domain and multi-source data | |
| CN109726770A (en) | A kind of analog circuit fault testing and diagnosing method | |
| CN110764152B (en) | Device and method for rapid detection and identification of unmanned aerial vehicle |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |