CN113780352B - Satellite receiver health management method based on neural network - Google Patents
Satellite receiver health management method based on neural network Download PDFInfo
- Publication number
- CN113780352B CN113780352B CN202110913850.0A CN202110913850A CN113780352B CN 113780352 B CN113780352 B CN 113780352B CN 202110913850 A CN202110913850 A CN 202110913850A CN 113780352 B CN113780352 B CN 113780352B
- Authority
- CN
- China
- Prior art keywords
- neuron
- self
- organizing map
- euclidean distance
- test 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.)
- Active
Links
- 230000036541 health Effects 0.000 title claims abstract description 40
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 9
- 238000007726 management method Methods 0.000 title abstract description 11
- 210000002569 neuron Anatomy 0.000 claims abstract description 106
- 238000012360 testing method Methods 0.000 claims abstract description 97
- 239000013598 vector Substances 0.000 claims abstract description 75
- 230000001537 neural effect Effects 0.000 claims abstract description 64
- 230000002159 abnormal effect Effects 0.000 claims abstract description 28
- 238000010606 normalization Methods 0.000 claims abstract description 14
- 238000012545 processing Methods 0.000 claims abstract description 14
- 238000012549 training Methods 0.000 claims description 64
- 238000000034 method Methods 0.000 claims description 20
- 238000004590 computer program Methods 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 abstract description 6
- 238000012544 monitoring process Methods 0.000 abstract description 5
- 238000003745 diagnosis Methods 0.000 abstract description 3
- 238000012423 maintenance Methods 0.000 abstract description 3
- 230000005856 abnormality Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000002860 competitive effect Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000012821 model calculation Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
- G01S19/01—Satellite radio beacon positioning systems transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
- G01S19/13—Receivers
- G01S19/23—Testing, monitoring, correcting or calibrating of receiver elements
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biophysics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Evolutionary Biology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Monitoring And Testing Of Transmission In General (AREA)
Abstract
The invention provides a satellite receiver health management method based on a neural network, which comprises the following steps: s10, collecting test data of a satellite receiver under a normal working condition and test data of the satellite receiver under an unknown working condition; s20, carrying out normalization processing on test data under normal working conditions, test data under unknown working conditions and weight vectors of each neuron in the self-organizing map neural grid model; s30, establishing a self-organizing map neural grid model; s40, obtaining a trained self-organizing map neural grid model; s50, acquiring Euclidean distance between each test sample in the test sample set and the corresponding best matched neuron; s60, acquiring an abnormal index quantity of each test sample; s70, obtaining the health degree of the test sample; s80, acquiring the health state of the satellite receiver. The invention can solve the technical problem that the traditional fault diagnosis and maintenance support technology can not realize the health condition monitoring of the satellite receiver in a long-time running state rapidly and accurately.
Description
Technical Field
The invention relates to the technical field of satellite navigation health management, in particular to a satellite receiver health management method based on a neural network.
Background
The global satellite navigation system (GNSS) is a satellite-based navigation positioning system, can provide high-precision positioning, speed measurement and time service data for users, has the advantages of all weather, wide coverage, no accumulated error and the like, has wide application market in the military and civil fields, but has precision which is easy to be interfered, such as multipath effect, antenna shielding, electromagnetic interference and the like. Satellite navigation and positioning are the basis of high-efficiency operation in the fields of modern weaponry, civil aviation and the like, and the problem that the performance of a satellite navigation system cannot meet the requirement of long-time operation of weaponry and civil aircraft is needed to be solved.
With the rapid development of science and technology, the integration level, complexity and intelligent degree of weaponry are rapidly increased, and the traditional fault diagnosis and maintenance support technology cannot rapidly and accurately realize the health condition monitoring of the long-time running state of the satellite receiver.
Disclosure of Invention
The invention provides a satellite receiver health management method based on a neural network, which can solve the technical problem that the traditional fault diagnosis and maintenance guarantee technology cannot realize the health condition monitoring of the long-time running state of a satellite receiver quickly and accurately.
According to an aspect of the present invention, there is provided a satellite receiver health management method based on a neural network, the method comprising:
S10, collecting test data of a satellite receiver under a normal working condition and test data of the satellite receiver under an unknown working condition;
s20, carrying out normalization processing on test data under normal working conditions, test data under unknown working conditions and weight vectors of each neuron in the self-organizing map neural grid model, taking the test data under the normal working conditions after normalization processing as a normal training sample set, and taking the test data under the unknown working conditions after normalization processing as a test sample set;
s30, initializing each neuron in the self-organizing map neural grid model, and establishing the self-organizing map neural grid model;
S40, training the self-organizing map neural grid model by using a normal training sample set to obtain a trained self-organizing map neural grid model;
s50, importing a test sample set into a trained self-organizing map neural grid model, and acquiring Euclidean distance between each test sample in the test sample set and a corresponding best matching neuron;
S60, acquiring an abnormal index quantity of each test sample according to the Euclidean distance between each test sample and the corresponding best matched neuron, the minimum Euclidean distance average value trained by the normal training sample set and the minimum Euclidean distance standard deviation trained by the normal training sample set;
S70, under the condition that the abnormal index quantity of the test sample is larger than zero, obtaining the health degree of the corresponding test sample according to each abnormal index quantity larger than zero and the scale parameter;
s80, acquiring the health state of the satellite receiver according to the health degree of the test samples with all abnormal index amounts larger than zero.
Preferably, the abnormal index amount of each test sample is obtained by:
H＝|MQE_T-MQE_ave|-3δ；
wherein, H represents the abnormal index quantity of the current test sample, MQE _t represents the euclidean distance between the current test sample and the corresponding best matching neuron, MQE _ave represents the minimum euclidean distance average value trained by the normal training sample set, and δ represents the minimum euclidean distance standard deviation trained by the normal training sample set.
Preferably, the Euclidean distance between each test sample and the corresponding best matching neuron is obtained by:
Where x i represents the ith test sample and x BMU represents the best matching neuron corresponding to the ith test sample.
Preferably, the minimum Euclidean distance standard deviation trained by the normal training sample set is obtained by the following formula:
Where N h represents the number of sample vectors in the normal training sample set, MQE i represents the euclidean distance between the i-th sample vector in the normal training sample set and the corresponding best matching neuron.
Preferably, the health of each test sample is obtained by:
where D represents the health of the current test sample and c represents the scale parameter.
Preferably, the scale parameter is obtained by:
Where N h represents the number of sample vectors in the normal training sample set, MQE i represents the euclidean distance between the i-th sample vector in the normal training sample set and the corresponding best matching neuron.
Preferably, S40, training the self-organizing map neural mesh model by using the normal training sample set, and obtaining the trained self-organizing map neural mesh model includes:
S41, selecting any sample vector from a normal training sample set, and importing the currently selected sample vector into a self-organizing map neural grid model;
s42, calculating Euclidean distance between a weight vector of each neuron in the self-organizing map neural grid model and a currently selected sample vector, taking the minimum Euclidean distance as a radius of a winning field, and taking a neuron corresponding to the minimum Euclidean distance as an optimal matched neuron;
s43, judging whether the radius of the winning field is equal to 0, if so, obtaining a trained self-organizing map neural grid model, otherwise, turning to S44;
s44, updating the weight vector of the best matched neuron and the weight vector of the neuron adjacent to the best matched neuron;
s45, selecting any sample vector from the residual sample vectors of the normal training sample set, importing the currently selected sample vector into the self-organizing map neural grid model, and turning to S42.
Preferably, the Euclidean distance between the weight vector of each neuron in the self-organizing map neural mesh model and the currently selected sample vector is calculated by:
Where d k (x) represents the Euclidean distance between the weight vector of the kth neuron in the self-organizing map neural mesh model and the currently selected sample vector, k represents the kth neuron in the self-organizing map neural mesh model, S.k represents the total number of neurons in the self-organizing map neural mesh model, x k represents the currently selected sample vector, and w k represents the weight vector of the kth neuron.
Preferably, the weight vector of the best matching neuron and the weight vector of the neuron adjacent to the best matching neuron are updated by:
wl(p+1)＝wl(p)+α(p)h(nBMU,nl,p)(yc-wl(p))；
Where w l (p+1) represents the weight vector of the best matching neuron or the neuron adjacent to the best matching neuron after updating, w l (p) represents the weight vector of the best matching neuron or the neuron adjacent to the best matching neuron, α (p) represents the learning rate at the time of training step p, h (n BMU,nl, p) represents the neighborhood function between the best matching neuron and the first adjacent neuron, and y c represents the normal training sample set.
According to a further aspect of the present invention there is provided a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing any of the methods described above when executing the computer program.
By adopting the technical scheme, the data is normalized, so that model calculation is facilitated, and the model classification accuracy is improved to a certain extent; training the self-organizing map neural grid model by using a normal training sample set to obtain a trained self-organizing map neural grid model, introducing a test sample set into the trained self-organizing map neural grid model to obtain an abnormal index quantity of the test sample, and obtaining the health degree of the test sample according to the abnormal index quantity of the test sample, thereby more rapidly and accurately solving the health condition monitoring in the long-time operation process of the satellite receiver system.
Drawings
The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention. It is evident that the drawings in the following description are only some embodiments of the present invention and that other drawings may be obtained from these drawings without inventive effort for a person of ordinary skill in the art.
FIG. 1 illustrates a flow chart of a method for neural network-based satellite receiver health management, provided in accordance with one embodiment of the present invention;
FIG. 2 illustrates a schematic diagram of a self-organizing map neural mesh model provided in accordance with one embodiment of the present invention;
FIG. 3 illustrates a schematic diagram of an anomaly index quantity and a sample tag provided in accordance with one embodiment of the present invention;
Fig. 4 shows a schematic diagram of a health assessment state provided according to an embodiment of the present invention.
Detailed Description
It should be noted that, without conflict, the embodiments of the present application and features of the embodiments may be combined with each other. The following description of the embodiments of the present application will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present application, but not all embodiments. The following description of at least one exemplary embodiment is merely exemplary in nature and is in no way intended to limit the application, its application, or uses. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the present application. As used herein, the singular is also intended to include the plural unless the context clearly indicates otherwise, and furthermore, it is to be understood that the terms "comprises" and/or "comprising" when used in this specification are taken to specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof.
The relative arrangement of the components and steps, numerical expressions and numerical values set forth in these embodiments do not limit the scope of the present invention unless it is specifically stated otherwise. Meanwhile, it should be understood that the sizes of the respective parts shown in the drawings are not drawn in actual scale for convenience of description. Techniques, methods, and apparatus known to one of ordinary skill in the relevant art may not be discussed in detail, but should be considered part of the specification where appropriate. In all examples shown and discussed herein, any specific values should be construed as merely illustrative, and not a limitation. Thus, other examples of the exemplary embodiments may have different values. It should be noted that: like reference numerals and letters denote like items in the following figures, and thus once an item is defined in one figure, no further discussion thereof is necessary in subsequent figures.
As shown in fig. 1, the present invention provides a satellite receiver health management method based on a neural network, the method comprising:
S10, collecting test data of a satellite receiver under a normal working condition and test data of the satellite receiver under an unknown working condition;
s20, carrying out normalization processing on test data under normal working conditions, test data under unknown working conditions and weight vectors of each neuron in the self-organizing map neural grid model, taking the test data under the normal working conditions after normalization processing as a normal training sample set, and taking the test data under the unknown working conditions after normalization processing as a test sample set;
s30, initializing each neuron in the self-organizing map neural grid model, and establishing the self-organizing map neural grid model;
S40, training the self-organizing map neural grid model by using a normal training sample set to obtain a trained self-organizing map neural grid model;
s50, importing a test sample set into a trained self-organizing map neural grid model, and acquiring Euclidean distance between each test sample in the test sample set and a corresponding best matching neuron;
S60, acquiring an abnormal index quantity of each test sample according to the Euclidean distance between each test sample and the corresponding best matched neuron, the minimum Euclidean distance average value trained by the normal training sample set and the minimum Euclidean distance standard deviation trained by the normal training sample set;
S70, under the condition that the abnormal index quantity of the test sample is larger than zero, obtaining the health degree of the corresponding test sample according to each abnormal index quantity larger than zero and the scale parameter;
s80, acquiring the health state of the satellite receiver according to the health degree of the test samples with all abnormal index amounts larger than zero.
According to the invention, through carrying out normalization processing on the data, model calculation is facilitated, and the model classification accuracy is improved to a certain extent; training the self-organizing map neural grid model by using a normal training sample set to obtain a trained self-organizing map neural grid model, introducing a test sample set into the trained self-organizing map neural grid model to obtain an abnormal index quantity of the test sample, and obtaining the health degree of the test sample according to the abnormal index quantity of the test sample, thereby more rapidly and accurately solving the health condition monitoring in the long-time operation process of the satellite receiver system.
In order to meet the requirements of informatization war on quick and reliable operations of weaponry, the invention combines the technology of fault prediction and health management (Prognostics AND HEALTH MANAGEMENT, PHM) with a satellite receiver system, and the PHM technology utilizes the acquired data to evaluate the satellite receiver in time so as to acquire the health state of the satellite receiver, thereby avoiding the occurrence of faults when the satellite receiver runs for a long time, and causing personnel energy loss and military and economic losses.
According to one embodiment of the present invention, in S20, normalization processing is performed on the test data under the normal working condition, the test data under the unknown working condition, and the weight vector of each neuron in the self-organizing map neural mesh model, so as to avoid the excessive difference between the test data and the weight vector value. In the present invention, normalization processing is performed by the following formula:
wherein value represents a normalized feature value, the value range is [ -1,1], lower represents-1, upper represents 1, max represents the maximum value of the feature column, and min represents the minimum value of the feature column.
In accordance with one embodiment of the present invention, each neuron in the self-organizing map neural mesh model is initialized, i.e., a small initial value is assigned to the weight of each neuron in the competitive layer of the self-organizing map neural mesh model, at S30.
According to an embodiment of the present invention, S40, training the self-organizing map neural mesh model by using the normal training sample set, to obtain a trained self-organizing map neural mesh model includes:
S41, selecting any sample vector from a normal training sample set, and importing the currently selected sample vector into a self-organizing map neural grid model;
s42, calculating Euclidean distance between a weight vector of each neuron in the self-organizing map neural grid model and a currently selected sample vector, taking the minimum Euclidean distance as a radius of a winning field, and taking a neuron corresponding to the minimum Euclidean distance as an optimal matched neuron;
s43, judging whether the radius of the winning field is equal to 0, if so, obtaining a trained self-organizing map neural grid model, otherwise, turning to S44;
s44, updating the weight vector of the best matched neuron and the weight vector of the neuron adjacent to the best matched neuron;
s45, selecting any sample vector from the residual sample vectors of the normal training sample set, importing the currently selected sample vector into the self-organizing map neural grid model, and turning to S42.
In the present invention, the number of neurons is generally determined based on the number of training samples. Specifically, the number of neurons in the self-organizing map neural mesh model is determined by:
Where n represents the number of neurons and m train represents the number of training samples.
In accordance with one embodiment of the present invention, in S42, the euclidean distance between the weight vector of each neuron in the self-organizing map neural mesh model and the currently selected sample vector is calculated by:
Where d k (x) represents the Euclidean distance between the weight vector of the kth neuron in the self-organizing map neural mesh model and the currently selected sample vector, k represents the kth neuron in the self-organizing map neural mesh model, S.k represents the total number of neurons in the self-organizing map neural mesh model, x k represents the currently selected sample vector, and w k represents the weight vector of the kth neuron.
In accordance with one embodiment of the present invention, in S44, the weight vector of the best matching neuron and the weight vector of the neuron adjacent to the best matching neuron are updated by:
wl(p+1)＝wl(p)+α(p)h(nBMU,nl,p)(yc-wl(p))；
Where w l (p+1) represents the weight vector of the best matching neuron or the neuron adjacent to the best matching neuron after updating, w l (p) represents the weight vector of the best matching neuron or the neuron adjacent to the best matching neuron, α (p) represents the learning rate at the time of training step p, h (n BMU,nl, p) represents the neighborhood function between the best matching neuron and the first adjacent neuron, and y c represents the normal training sample set.
Wherein,
Where d BMU,l represents the distance between the best matching neuron n BMU and the first neighboring neuron n l, and r represents the corresponding winning domain radius at the p-th training step.
According to one embodiment of the present invention, in S60, an abnormality index amount for each test sample is obtained by:
H＝|MQE_T-MQE_ave|-3δ；
wherein, H represents the abnormal index quantity of the current test sample, MQE _t represents the euclidean distance between the current test sample and the corresponding best matching neuron, MQE _ave represents the minimum euclidean distance average value trained by the normal training sample set, and δ represents the minimum euclidean distance standard deviation trained by the normal training sample set.
In the case that the abnormal index quantity H of the test sample is more than 0, the detection result is in an abnormal state; in the case where the abnormality index amount H of the test sample is 0 or less, the detection result is in a normal state.
The Euclidean distance between each test sample and the corresponding best matching neuron is obtained by the following formula:
Where x i represents the ith test sample and x BMU represents the best matching neuron corresponding to the ith test sample.
The minimum Euclidean distance standard deviation trained by the normal training sample set is obtained through the following formula:
Where N h represents the number of sample vectors in the normal training sample set, MQE i represents the euclidean distance between the i-th sample vector in the normal training sample set and the corresponding best matching neuron.
According to one embodiment of the present invention, in S70, the health of each test sample is obtained by:
where D represents the health of the current test sample and c represents the scale parameter.
Wherein the scale parameter is obtained by:
Where N h represents the number of sample vectors in the normal training sample set, MQE i represents the euclidean distance between the i-th sample vector in the normal training sample set and the corresponding best matching neuron.
In the present invention, the greater the abnormal index amount of the test sample, the lower the health thereof, that is, the higher the degree of abnormality.
The method of the present invention will be described in detail with reference to telemetry data acquired by a satellite receiver.
In this embodiment, the collected test data under the normal working condition is telemetry data of the known satellite receiver in the normal state; the collected test data under unknown working conditions are satellite receiver telemetry data of random test. Setting the data value range after normalization processing as [ -1,1]. The number of neurons in the competitive layer of the self-organizing map neural grid model is 144 (the corresponding grid side length is 12), the initial learning rate alpha is 0.6, the initial winning domain radius r is 2, the maximum training times is 12000, and the established self-organizing map neural grid model is shown in fig. 2.
Through the data, the minimum Euclidean distance average MQE _ave=0.42 trained by the normal training sample set can be obtained, the minimum Euclidean distance standard deviation delta=0.17 trained by the normal training sample set is obtained, and the test accuracy is 90%.
Thus, an abnormal index quantity H of each test sample can be obtained, and when H is less than or equal to 0, the sample label is classified as 1 to represent a normal state; when H >0, the sample tag is classified as-1, indicating an abnormal state, as shown in FIG. 3.
And obtaining the health degree of the corresponding test sample according to each abnormal index quantity and the scale parameter which are larger than zero, as shown in fig. 4.
The invention also provides a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing any of the methods described above when executing the computer program.
Spatially relative terms, such as "above … …," "above … …," "upper surface on … …," "above," and the like, may be used herein for ease of description to describe one device or feature's spatial location relative to another device or feature as illustrated in the figures. It will be understood that the spatially relative terms are intended to encompass different orientations in use or operation in addition to the orientation depicted in the figures. For example, if the device in the figures is turned over, elements described as "above" or "over" other devices or structures would then be oriented "below" or "beneath" the other devices or structures. Thus, the exemplary term "above … …" may include both orientations "above … …" and "below … …". The device may also be positioned in other different ways (rotated 90 degrees or at other orientations) and the spatially relative descriptors used herein interpreted accordingly.
In addition, the terms "first", "second", etc. are used to define the components, and are only for convenience of distinguishing the corresponding components, and the terms have no special meaning unless otherwise stated, and therefore should not be construed as limiting the scope of the present invention.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (9)
1. A method for satellite receiver health management based on a neural network, the method comprising:
S10, collecting test data of a satellite receiver under a normal working condition and test data of the satellite receiver under an unknown working condition;
s20, carrying out normalization processing on test data under normal working conditions, test data under unknown working conditions and weight vectors of each neuron in the self-organizing map neural grid model, taking the test data under the normal working conditions after normalization processing as a normal training sample set, and taking the test data under the unknown working conditions after normalization processing as a test sample set;
s30, initializing each neuron in the self-organizing map neural grid model, and establishing the self-organizing map neural grid model;
S40, training the self-organizing map neural grid model by using a normal training sample set to obtain a trained self-organizing map neural grid model;
s50, importing a test sample set into a trained self-organizing map neural grid model, and acquiring Euclidean distance between each test sample in the test sample set and a corresponding best matching neuron;
S60, acquiring an abnormal index quantity of each test sample according to the Euclidean distance between each test sample and the corresponding best matched neuron, the minimum Euclidean distance average value trained by the normal training sample set and the minimum Euclidean distance standard deviation trained by the normal training sample set;
S70, under the condition that the abnormal index quantity of the test sample is larger than zero, obtaining the health degree of the corresponding test sample according to each abnormal index quantity larger than zero and the scale parameter;
S80, acquiring the health state of the satellite receiver according to the health degree of the test sample with all abnormal index amounts larger than zero;
wherein the abnormal index amount of each test sample is obtained by:
H＝|MQE_T-MQE_ave|-3δ；
wherein, H represents the abnormal index quantity of the current test sample, MQE _t represents the euclidean distance between the current test sample and the corresponding best matching neuron, MQE _ave represents the minimum euclidean distance average value trained by the normal training sample set, and δ represents the minimum euclidean distance standard deviation trained by the normal training sample set.
2. The method of claim 1, wherein the euclidean distance between each test sample and the corresponding best matching neuron is obtained by:
Where x i represents the ith test sample and x BMU represents the best matching neuron corresponding to the ith test sample.
3. The method of claim 1, wherein the minimum euclidean distance standard deviation trained out of the normal training sample set is obtained by:
Where N h represents the number of sample vectors in the normal training sample set, MQE i represents the euclidean distance between the i-th sample vector in the normal training sample set and the corresponding best matching neuron.
4. A method according to any one of claims 1 to 3, wherein the health of each test sample is obtained by:
where D represents the health of the current test sample and c represents the scale parameter.
5. The method of claim 4, wherein the scale parameter is obtained by:
Where N h represents the number of sample vectors in the normal training sample set, MQE i represents the euclidean distance between the i-th sample vector in the normal training sample set and the corresponding best matching neuron.
6. The method of claim 1, wherein training the self-organizing map neural mesh model using the normal training sample set to obtain a trained self-organizing map neural mesh model comprises:
S41, selecting any sample vector from a normal training sample set, and importing the currently selected sample vector into a self-organizing map neural grid model;
s42, calculating Euclidean distance between a weight vector of each neuron in the self-organizing map neural grid model and a currently selected sample vector, taking the minimum Euclidean distance as a radius of a winning field, and taking a neuron corresponding to the minimum Euclidean distance as an optimal matched neuron;
s43, judging whether the radius of the winning field is equal to 0, if so, obtaining a trained self-organizing map neural grid model, otherwise, turning to S44;
s44, updating the weight vector of the best matched neuron and the weight vector of the neuron adjacent to the best matched neuron;
s45, selecting any sample vector from the residual sample vectors of the normal training sample set, importing the currently selected sample vector into the self-organizing map neural grid model, and turning to S42.
7. The method of claim 6, wherein the euclidean distance between the weight vector of each neuron in the self-organizing map neural mesh model and the currently selected sample vector is calculated by:
Where d k (x) represents the Euclidean distance between the weight vector of the kth neuron in the self-organizing map neural mesh model and the currently selected sample vector, k represents the kth neuron in the self-organizing map neural mesh model, S.k represents the total number of neurons in the self-organizing map neural mesh model, x k represents the currently selected sample vector, and w k represents the weight vector of the kth neuron.
8. The method of claim 6, wherein the weight vector of the best matching neuron and the weight vector of the neuron adjacent to the best matching neuron are updated by:
wl(p+1)＝wl(p)+α(p)h(nBMU,nl,p)(yc-wl(p))；
Where w l (p+1) represents the weight vector of the best matching neuron or the neuron adjacent to the best matching neuron after updating, w l (p) represents the weight vector of the best matching neuron or the neuron adjacent to the best matching neuron, α (p) represents the learning rate at the time of training step p, h (n BMU,nl, p) represents the neighborhood function between the best matching neuron and the first adjacent neuron, y c represents the normal training sample set, n BMU represents the best matching neuron, and n l represents the adjacent neuron.
9. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the method of any of claims 1 to 8 when executing the computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110913850.0A CN113780352B (en) | 2021-08-10 | 2021-08-10 | Satellite receiver health management method based on neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110913850.0A CN113780352B (en) | 2021-08-10 | 2021-08-10 | Satellite receiver health management method based on neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113780352A CN113780352A (en) | 2021-12-10 |
| CN113780352B true CN113780352B (en) | 2024-06-14 |
Family
ID=78837252
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110913850.0A Active CN113780352B (en) | 2021-08-10 | 2021-08-10 | Satellite receiver health management method based on neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113780352B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107967489A (en) * | 2017-11-29 | 2018-04-27 | 中国科学院空间应用工程与技术中心 | A kind of method for detecting abnormality and system |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4356716B2 (en) * | 2006-08-03 | 2009-11-04 | パナソニック電工株式会社 | Abnormality monitoring device |
| US10311356B2 (en) * | 2013-09-09 | 2019-06-04 | North Carolina State University | Unsupervised behavior learning system and method for predicting performance anomalies in distributed computing infrastructures |
| CN110298830A (en) * | 2019-06-24 | 2019-10-01 | 天津大学 | Fresh concrete based on convolutional neural networks vibrates apparent mass recognition methods |
| CN111221849B (en) * | 2019-12-26 | 2023-05-30 | 北京航天测控技术有限公司 | Satellite fault positioning method, device, terminal and computer readable medium |
| CN112816884B (en) * | 2021-03-01 | 2022-08-02 | 中国人民解放军国防科技大学 | Method, device and equipment for monitoring health state of satellite lithium ion battery |
| CN112949753B (en) * | 2021-03-26 | 2023-10-24 | 西安交通大学 | Binary relation-based satellite telemetry time sequence data anomaly detection method |
-
2021
- 2021-08-10 CN CN202110913850.0A patent/CN113780352B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107967489A (en) * | 2017-11-29 | 2018-04-27 | 中国科学院空间应用工程与技术中心 | A kind of method for detecting abnormality and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113780352A (en) | 2021-12-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109947086B (en) | Mechanical fault migration diagnosis method and system based on counterstudy | |
| CN101907681B (en) | Analog circuit dynamic online failure diagnosing method based on GSD-SVDD | |
| CN110954132B (en) | GRNN-assisted self-adaptive Kalman filtering navigation fault identification method | |
| CN109031374A (en) | Difference pseudo-range corrections abnormal signal monitoring method suitable for continuous operation of the reference station | |
| CN113112510B (en) | Semantic segmentation forest fire detection method, controller and storage medium | |
| CN106441279A (en) | Robot positioning method based on self-positioning and edge searching and system | |
| CN112817318B (en) | Multi-unmanned-boat collaborative search control method and system | |
| CN114969990A (en) | Multi-model fused avionic product health assessment method | |
| CN113780352B (en) | Satellite receiver health management method based on neural network | |
| CN116310913B (en) | Natural resource investigation monitoring method and device based on unmanned aerial vehicle measurement technology | |
| CN113780573B (en) | Anti-noise high-precision distance measuring method and device | |
| Rambli et al. | Outlier detection in a circular regression model | |
| Favenza et al. | A machine learning approach to GNSS scintillation detection: Automatic soft inspection of the events | |
| CN103903258B (en) | Method for detecting change of remote sensing image based on order statistic spectral clustering | |
| CN113720333B (en) | Indoor multipoint co-location method based on graph attention network | |
| CN112966710B (en) | FY-3D infrared hyperspectral cloud detection method based on linear discriminant analysis | |
| CN107203493A (en) | Multiple target battle field situation method based on complicated ratio evaluation method | |
| CN112464997A (en) | Multi-sensor collaborative monitoring method and system based on fuzzy-greedy search decision | |
| CN107608333B (en) | A kind of diagnosticability appraisal procedure based on equivalent depression of order | |
| CN111553421A (en) | SAR equipment task failure cause reasoning method based on double-layer nested structure | |
| CN115185814B (en) | Multi-defect positioning method, system and equipment based on two-dimensional program frequency spectrum | |
| CN113419140B (en) | Distribution network line fault diagnosis method and device considering human-computer-object cooperation | |
| CN117725523B (en) | Thunderstorm weather early warning method and early warning system based on physical constraint | |
| CN114943344B (en) | Navigation scene perception general model construction method combining SVM and HMM | |
| CN116486178B (en) | Defect detection method and device, electronic equipment and storage medium |
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 |