CN113029202B - Redundancy strapdown inertial measurement unit fault detection method based on parameter optimization depth confidence model - Google Patents
Redundancy strapdown inertial measurement unit fault detection method based on parameter optimization depth confidence model Download PDFInfo
- Publication number
- CN113029202B CN113029202B CN202110568792.2A CN202110568792A CN113029202B CN 113029202 B CN113029202 B CN 113029202B CN 202110568792 A CN202110568792 A CN 202110568792A CN 113029202 B CN113029202 B CN 113029202B
- Authority
- CN
- China
- Prior art keywords
- confidence model
- depth confidence
- boltzmann machine
- layer
- limiting boltzmann
- 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
- 238000005259 measurement Methods 0.000 title claims abstract description 30
- 238000001514 detection method Methods 0.000 title claims abstract description 16
- 238000005457 optimization Methods 0.000 title claims abstract description 14
- 238000012549 training Methods 0.000 claims abstract description 32
- 238000012360 testing method Methods 0.000 claims abstract description 10
- 230000009191 jumping Effects 0.000 claims description 9
- 238000000034 method Methods 0.000 claims description 9
- 230000001133 acceleration Effects 0.000 claims description 6
- 238000012795 verification Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 abstract description 2
- 210000002569 neuron Anatomy 0.000 description 10
- 230000009286 beneficial effect Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 2
- 238000007796 conventional method Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C25/00—Manufacturing, calibrating, cleaning, or repairing instruments or devices referred to in the other groups of this subclass
- G01C25/005—Manufacturing, calibrating, cleaning, or repairing instruments or devices referred to in the other groups of this subclass initial alignment, calibration or starting-up of inertial devices
Landscapes
- Engineering & Computer Science (AREA)
- Manufacturing & Machinery (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Navigation (AREA)
Abstract
The invention discloses a redundant strapdown inertial measurement unit fault detection method based on a parameter optimization depth confidence model, belonging to the field of aircraft navigation, guidance and control and comprising the following steps of: s1, acquiring output data of the redundant strapdown inertial measurement unit, and constructing a training set and a test set based on the output data; s2, training the deep confidence model by adopting a training set to obtain a trained deep confidence model; s3, optimizing the structural parameters of the trained depth confidence model to obtain a depth confidence model with optimal structural parameters; s4, inputting the test set into a depth confidence model with the optimal structural parameters, and performing redundant strapdown inertial measurement unit fault detection; the invention solves the problems of inaccurate fault modeling, difficult manual feature extraction and limitation on manual parameter adjustment in the conventional redundant strapdown inertial measurement unit fault detection method.
Description
Technical Field
The invention belongs to the field of aircraft navigation, guidance and control, and particularly relates to a redundant strapdown inertial measurement unit fault detection method based on a parameter optimization depth confidence model.
Background
In many important aerospace applications, reliability requirements for aircraft navigation systems are very high. The adoption of the redundant strapdown inertial measurement unit can significantly improve the reliability of the navigation system, provided that the redundant strapdown inertial measurement unit has a high fault tolerance, which can be realized by using a Fault Detection and Isolation (FDI) technology.
In recent years, deep learning has been pursued by more and more researchers due to its excellent classification ability. Under the condition that a system model is unknown, features can be extracted through deep learning model training data, and classification is achieved. The Deep Belief Network (DBN) has good feature learning and classification capabilities as a deep network model, and is widely applied to various fields. At present, the deep confidence network structure is mostly determined by experience or by manually adjusting parameters, and the development of the deep confidence network structure is limited to a certain extent.
Disclosure of Invention
Aiming at the defects in the prior art, the redundancy strapdown inertial measurement unit fault detection method based on the parameter optimization depth confidence model solves the problems that the fault modeling is inaccurate, the characteristics are difficult to extract manually and the manual parameter adjustment is limited in the conventional redundancy strapdown inertial measurement unit fault detection method.
In order to achieve the purpose of the invention, the invention adopts the technical scheme that: a redundant strapdown inertial measurement unit fault detection method based on a parameter optimization depth confidence model comprises the following steps:
s1, acquiring output data of the redundant strapdown inertial measurement unit, and constructing a training set and a test set based on the output data;
s2, training the deep confidence model by adopting a training set to obtain a trained deep confidence model;
s3, optimizing the structural parameters of the trained depth confidence model to obtain a depth confidence model with optimal structural parameters;
and S4, inputting the test set into the depth confidence model with the optimal structural parameters, and detecting the fault of the redundant strapdown inertial measurement unit.
Further, the depth confidence model in step S2 is composed of multiple layers of limiting boltzmann machine units, and the depth confidence model specifically includes a layer 1 limiting boltzmann machine unit, a layer 2 limiting boltzmann machine unit, a layer 3 limiting boltzmann machine unit, a layer 4 limiting boltzmann machine unit, and up to an nth layer limiting boltzmann machine unit, which are connected in sequence.
Further, the step S2 includes the following sub-steps:
s21, training the limiting Boltzmann machine units layer by layer to obtain all the limiting Boltzmann machine units after training;
s22, extracting features of the data output by the trained nth layer limiting Boltzmann machine unit to obtain an output result;
s23, comparing the output result with the actual value to obtain an error value;
and S24, distributing error values from the trained layer 1 limiting Boltzmann machine unit to the trained layer n limiting Boltzmann machine unit in sequence by adopting a back propagation method for parameter fine adjustment, and obtaining the trained deep confidence model.
The beneficial effects of the above further scheme are: the method can automatically extract and learn the characteristics of the data, and overcomes the difficulty of manually extracting the characteristics.
Further, the step S21 includes the following steps:
s211, initializing an iteration parameter r, wherein r = 1;
s212, inputting the training set into the 1 st layer of limiting Boltzmann machine unit, training the 1 st layer of limiting Boltzmann machine unit, updating the weight and the bias of the r layer of limiting Boltzmann machine unit, and obtaining the trained r layer of limiting Boltzmann machine unit;
s213, adding 1 to r, judging whether r is equal to n +1, if yes, finishing the training of all the limiting Boltzmann machine units, and if not, jumping to the step S212.
The beneficial effects of the above further scheme are: the training efficiency is improved, and the problem of local optimization is improved.
Further, the step S3 includes the following sub-steps:
s31, initializing an iteration parameter k, wherein k = 0, and taking the trained depth confidence model as a current optimized depth confidence model;
s32, calculating the fitness of the current optimized depth confidence model;
s33, judging whether the fitness meets a set threshold, if so, taking the current optimized depth confidence model as the depth confidence model with the optimal structural parameters, ending the substeps, if not, adding 1 to k, and jumping to the step S34;
s34, calculating the position and the speed of the structural parameter by adopting the parameter updating model, and carrying out the optimization of the kth iteration on the structural parameter of the current optimized depth confidence model based on the position and the speed of the structural parameter to obtain a kth suboptimal depth confidence model;
s35, taking the k-th sub-optimal depth confidence model as the current optimal depth confidence model, and jumping to the step S32.
The beneficial effects of the above further scheme are: the method can overcome the randomness of manually selecting parameters, reasonably determines the network structure parameters, and is favorable for improving the performance of fault detection of the redundant strapdown inertial measurement unit.
Further, the formula for calculating the fitness of the current optimized depth confidence model in step S32 is as follows:
wherein,fin order to be a degree of fitness,Mfor the total number of verifications of the current optimized deep confidence model,mis as followsmThe result of the secondary verification is that,bin order to correctly classify the number of samples,cthe number of misclassified samples.
Further, in step S34, the parameter update model is:
wherein,αin order to be the inertial weight,β 1is a first one of the acceleration factors, and,β 2is the second one of the acceleration factors, and,v ij k is as followskStructural parameters of sub-iterationiTo (1) ajThe speed of the dimensional variable is then determined,v ij k+1is as followsk+Structural parameters of 1 iterationiTo (1) ajThe speed of the dimensional variable is then determined,x ij k is as followskStructural parameters of sub-iterationiTo (1) ajThe position of the dimension-variable is,x ij k+1is as followsk+Structural parameters of 1 iterationiTo (1) ajThe position of the dimension-variable is,p ij k is as followskStructural parameters of sub-iterationiTo (1) ajThe optimal position of the individual extrema of the dimensional variable,r j k is as followskThe first of the sub-iterationsjThe global optimum position of the dimensional variable,a 1、a 2is a random number between 0 and 1.
The beneficial effects of the above further scheme are: the method can avoid the situation that the parameters fall into the local optimal solution, and obtain the globally optimal parameters.
In conclusion, the beneficial effects of the invention are as follows: the method adopts the parameter optimization depth confidence model to extract the fault characteristics of the redundant strapdown inertial measurement unit, overcomes the problems of inaccurate fault modeling, difficult manual characteristic extraction, difficult parameter selection and the like of the conventional method, realizes the fault detection of the redundant strapdown inertial measurement unit, and is beneficial to improving the reliability of an aircraft navigation system.
Drawings
FIG. 1 is a flow chart of a redundant strapdown inertial measurement unit fault detection method based on a parameter optimized depth confidence model;
FIG. 2 is a block diagram of a depth confidence model;
FIG. 3 is a schematic diagram of a configuration of a regular dodecahedron redundant strapdown inertial mount.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate the understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the invention as defined and defined in the appended claims, and all matters produced by the invention using the inventive concept are protected.
As shown in fig. 1, a redundant strapdown inertial measurement unit fault detection method based on a parameter optimization depth confidence model includes the following steps:
s1, acquiring output data of the redundant strapdown inertial measurement unit, and constructing a training set and a test set based on the output data;
after the output data of the redundant strapdown inertial measurement unit is collected, the fault type of the redundant strapdown inertial measurement unit needs to be marked, and then a training set and a test set are constructed according to the fault type of the redundant strapdown inertial measurement unit.
S2, training the deep confidence model by adopting a training set to obtain a trained deep confidence model;
in step S2, the depth confidence model is composed of multiple layers of limiting boltzmann machine units, and the depth confidence model specifically includes a layer 1 limiting boltzmann machine unit, a layer 2 limiting boltzmann machine unit, a layer 3 limiting boltzmann machine unit, a layer 4 limiting boltzmann machine unit, and up to an nth limiting boltzmann machine unit, which are connected in sequence, as shown in fig. 2.
Step S2 includes the following substeps:
s21, training the limiting Boltzmann machine units layer by layer to obtain all the limiting Boltzmann machine units after training;
step S21 includes the following steps:
s211, initializing an iteration parameter r, wherein r = 1;
s212, inputting the training set into the 1 st layer of limiting Boltzmann machine unit, training the 1 st layer of limiting Boltzmann machine unit, updating the weight and the bias of the r layer of limiting Boltzmann machine unit, and obtaining the trained r layer of limiting Boltzmann machine unit;
s213, adding 1 to r, judging whether r is equal to n +1, if yes, finishing the training of all the limiting Boltzmann machine units, and if not, jumping to the step S212.
S22, extracting features of the data output by the trained nth layer limiting Boltzmann machine unit to obtain an output result;
s23, comparing the output result with the actual value to obtain an error value;
and S24, distributing error values from the trained layer 1 limiting Boltzmann machine unit to the trained layer n limiting Boltzmann machine unit in sequence by adopting a back propagation method for parameter fine adjustment, and obtaining the trained deep confidence model.
S3, optimizing the structural parameters of the trained depth confidence model to obtain a depth confidence model with optimal structural parameters;
step S3 includes the following substeps:
s31, initializing an iteration parameter k, wherein k = 0, and taking the trained depth confidence model as a current optimized depth confidence model;
s32, calculating the fitness of the current optimized depth confidence model:
wherein,fin order to be a degree of fitness,Mfor the total number of verifications of the current optimized deep confidence model,mis as followsmThe result of the secondary verification is that,bin order to correctly classify the number of samples,cthe number of misclassified samples.
S33, judging whether the fitness meets a set threshold, if so, taking the current optimized depth confidence model as the depth confidence model with the optimal structural parameters, ending the substeps, if not, adding 1 to k, and jumping to the step S34;
s34, calculating the position and the speed of the structural parameter by adopting the parameter updating model, and carrying out the optimization of the kth iteration on the structural parameter of the current optimized depth confidence model based on the position and the speed of the structural parameter to obtain a kth suboptimal depth confidence model;
the parameter update model in step S34 is:
wherein,αin order to be the inertial weight,β 1is a first one of the acceleration factors, and,β 2is the second one of the acceleration factors, and,v ij k is as followskStructural parameters of sub-iterationiTo (1) ajThe speed of the dimensional variable is then determined,v ij k+1is as followsk+Structure of 1 iterationParameter(s)iTo (1) ajThe speed of the dimensional variable is then determined,x ij k is as followskStructural parameters of sub-iterationiTo (1) ajThe position of the dimension-variable is,x ij k+1is as followsk+Structural parameters of 1 iterationiTo (1) ajThe position of the dimension-variable is,p ij k is as followskStructural parameters of sub-iterationiTo (1) ajThe optimal position of the individual extrema of the dimensional variable,r j k is as followskThe first of the sub-iterationsjThe global optimum position of the dimensional variable,a 1、a 2is a random number between 0 and 1.
S35, taking the k-th sub-optimal depth confidence model as the current optimal depth confidence model, and jumping to the step S32.
And S4, inputting the test set into the depth confidence model with the optimal structural parameters, and detecting the fault of the redundant strapdown inertial measurement unit.
As shown in fig. 3, a configuration scheme of a regular dodecahedron redundant strapdown inertial measurement unit is taken as an example to describe a method for detecting a fault of a redundant strapdown inertial measurement unit based on a parameter optimization depth confidence model: wherein O is a coordinate origin, XYZ is three orthogonal coordinate axes, m 1-m 6 are 6 inertial sensors, and the included angle between each inertial sensor and the coordinate axes isα。
Taking a gyroscope as an example, the fault types configured by the regular dodecahedron six-table redundant strapdown inertial measurement unit can be divided into: no. 1 gyro fault, No. 2 gyro fault, No. 3 gyro fault, No. 4 gyro fault, No. 5 gyro fault, No. 6 gyro fault and no fault.
The method is characterized in that a 4-layer depth confidence model structure is adopted, the output of six gyroscopes forms an input layer of a depth confidence model, and initial parameter hidden layers (the No. 1 limiting Boltzmann machine unit is outside a visible layer, and the rest are hidden layers) respectively comprise 20, 30, 10 and 7 neurons. Wherein 7 neurons in the last hidden layer respectively represent seven state characteristics of No. 1 gyro fault, No. 2 gyro fault, No. 3 gyro fault, No. 4 gyro fault, No. 5 gyro fault, No. 6 gyro fault, no fault and the like.
And after the training of the deep confidence model is finished by using the training set data, optimizing the structural parameters to obtain the optimal parameters. Inputting the test set data into a depth confidence model with optimal parameters, and checking the state of neurons contained in the last hidden layer:
if the state of the first neuron is 1, the No. 1 gyro is in failure, otherwise, the No. 1 gyro is not in failure; if the state of the second neuron is 1, the No. 2 gyro is in failure, otherwise, the No. 2 gyro is not in failure; if the state of the third neuron is 1, indicating that the No. 3 gyro fails, otherwise, indicating that the No. 3 gyro does not fail; if the state of the fourth neuron is 1, indicating that the No. 4 gyro fails, otherwise, indicating that the No. 4 gyro does not fail; if the state of the fifth neuron is 1, the No. 5 gyro is in failure, otherwise, the No. 5 gyro is not in failure; if the state of the sixth neuron is 1, indicating that the No. 6 gyro fails, otherwise, indicating that the No. 6 gyro does not fail; if the seventh neuron state is 1, all gyros are not in failure.
Claims (1)
1. A redundant strapdown inertial measurement unit fault detection method based on a parameter optimization depth confidence model is characterized by comprising the following steps:
s1, acquiring output data of the redundant strapdown inertial measurement unit, and constructing a training set and a test set based on the output data;
s2, training the deep confidence model by adopting a training set to obtain a trained deep confidence model;
s3, optimizing the structural parameters of the trained depth confidence model to obtain a depth confidence model with optimal structural parameters;
s4, inputting the test set into a depth confidence model with the optimal structural parameters, and performing redundant strapdown inertial measurement unit fault detection;
the depth confidence model in the step S2 is composed of a plurality of layers of limiting Boltzmann machine units, and specifically comprises a layer 1 limiting Boltzmann machine unit, a layer 2 limiting Boltzmann machine unit, a layer 3 limiting Boltzmann machine unit, a layer 4 limiting Boltzmann machine unit and a layer n limiting Boltzmann machine unit which are connected in sequence;
step S2 includes the following substeps:
s21, training the limiting Boltzmann machine units layer by layer to obtain all the limiting Boltzmann machine units after training;
s22, extracting features of the data output by the trained nth layer limiting Boltzmann machine unit to obtain an output result;
s23, comparing the output result with the actual value to obtain an error value;
s24, distributing error values from the trained layer 1 limiting Boltzmann machine unit to the trained layer n limiting Boltzmann machine unit in sequence by adopting a back propagation method for parameter fine adjustment to obtain a trained depth confidence model;
step S21 includes the following steps:
s211, initializing an iteration parameter r, wherein r = 1;
s212, inputting the training set into the 1 st layer of limiting Boltzmann machine unit, training the 1 st layer of limiting Boltzmann machine unit, updating the weight and the bias of the r layer of limiting Boltzmann machine unit, and obtaining the trained r layer of limiting Boltzmann machine unit;
s213, adding 1 to r, judging whether r is equal to n +1, if so, finishing the training of all the limiting Boltzmann machine units, and if not, jumping to the step S212;
step S3 includes the following substeps:
s31, initializing an iteration parameter k, wherein k = 0, and taking the trained depth confidence model as a current optimized depth confidence model;
s32, calculating the fitness of the current optimized depth confidence model;
s33, judging whether the fitness meets a set threshold, if so, taking the current optimized depth confidence model as the depth confidence model with the optimal structural parameters, ending the substeps, if not, adding 1 to k, and jumping to the step S34;
s34, calculating the position and the speed of the structural parameter by adopting the parameter updating model, and carrying out the optimization of the kth iteration on the structural parameter of the current optimized depth confidence model based on the position and the speed of the structural parameter to obtain a kth suboptimal depth confidence model;
s35, taking the k-th sub-optimal depth confidence model as the current optimal depth confidence model, and jumping to the step S32;
in step S32, the formula for calculating the fitness of the current optimized depth confidence model is:
wherein,fin order to be a degree of fitness,Mfor the total number of verifications of the current optimized deep confidence model,mis as followsmThe result of the secondary verification is that,bin order to correctly classify the number of samples,cthe number of samples classified in error;
the parameter update model in step S34 is:
wherein,αin order to be the inertial weight,β 1is a first one of the acceleration factors, and,β 2is the second one of the acceleration factors, and,v ij k is as followskStructural parameters of sub-iterationiTo (1) ajThe speed of the dimensional variable is then determined,v ij k+1is as followsk+Structural parameters of 1 iterationiTo (1) ajThe speed of the dimensional variable is then determined,x ij k is as followskStructural parameters of sub-iterationiTo (1) ajThe position of the dimension-variable is,x ij k+1is as followsk+Structural parameters of 1 iterationiTo (1) ajThe position of the dimension-variable is,p ij k is as followskStructural parameters of sub-iterationiTo (1) ajIndividual extreme values of dimensional variablesThe position of the optimum position is determined,r j k is as followskThe first of the sub-iterationsjThe global optimum position of the dimensional variable,a 1、a 2is a random number between 0 and 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110568792.2A CN113029202B (en) | 2021-05-25 | 2021-05-25 | Redundancy strapdown inertial measurement unit fault detection method based on parameter optimization depth confidence model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110568792.2A CN113029202B (en) | 2021-05-25 | 2021-05-25 | Redundancy strapdown inertial measurement unit fault detection method based on parameter optimization depth confidence model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113029202A CN113029202A (en) | 2021-06-25 |
| CN113029202B true CN113029202B (en) | 2021-09-07 |
Family
ID=76455701
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110568792.2A Active CN113029202B (en) | 2021-05-25 | 2021-05-25 | Redundancy strapdown inertial measurement unit fault detection method based on parameter optimization depth confidence model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113029202B (en) |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190135616A1 (en) * | 2017-11-03 | 2019-05-09 | University Of Dayton | Deep learning software enhanced microelectromechanical systems (mems) based inertial measurement unit (imu) |
| CN109188026B (en) * | 2018-10-25 | 2021-03-09 | 北京航空航天大学 | Automatic calibration deep learning method suitable for MEMS accelerometer |
| CN112766608B (en) * | 2021-02-03 | 2022-03-11 | 燕山大学 | Cement mill system power consumption index prediction method based on extreme learning machine |
-
2021
- 2021-05-25 CN CN202110568792.2A patent/CN113029202B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN113029202A (en) | 2021-06-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112257341B (en) | Customized product performance prediction method based on heterogeneous data difference compensation fusion | |
| CN113705597B (en) | Image processing method, device, computer equipment and readable storage medium | |
| CN113158445B (en) | Prediction algorithm for residual service life of aero-engine with convolution memory residual error self-attention mechanism | |
| CN105975573A (en) | KNN-based text classification method | |
| Sushchenko et al. | Processing of redundant information in airborne electronic systems by means of neural networks | |
| CN110633790B (en) | Method and system for measuring residual oil quantity of airplane oil tank based on convolutional neural network | |
| CN111216126B (en) | Multi-modal perception-based foot type robot motion behavior recognition method and system | |
| CN106886657B (en) | A kind of finite element model method for building up based on kriging function | |
| CN113297994B (en) | Pilot behavior analysis method and system | |
| CN110353694A (en) | A kind of motion recognition method based on Feature Selection | |
| CN114168740B (en) | Transformer concurrency fault diagnosis method based on graph convolution neural network and knowledge graph | |
| CN111914478A (en) | Comprehensive geological drilling well logging lithology identification method | |
| CN112132102A (en) | Intelligent fault diagnosis method combining deep neural network with artificial bee colony optimization | |
| CN111638034A (en) | Strain balance temperature gradient error compensation method and system based on deep learning | |
| CN113411744A (en) | High-precision indoor positioning and tracking method | |
| CN113591553A (en) | Turbo pump migration learning fault intelligent judgment method based on small sample weight optimization | |
| CN114492176B (en) | Kinetic model parameter identification method and system based on generation countermeasure network | |
| CN113029202B (en) | Redundancy strapdown inertial measurement unit fault detection method based on parameter optimization depth confidence model | |
| CN113627075B (en) | Projectile pneumatic coefficient identification method based on adaptive particle swarm optimization extreme learning | |
| CN104833359B (en) | A kind of star chart mode identification method based on discrete Markov characteristic sequence model | |
| CN116894180B (en) | Product manufacturing quality prediction method based on different composition attention network | |
| CN112651499A (en) | Structural model pruning method based on ant colony optimization algorithm and interlayer information | |
| CN116678434A (en) | Three-self laser inertial measurement unit fault detection method based on interval value confidence rule base | |
| CN116680969A (en) | Filler evaluation parameter prediction method and device for PSO-BP algorithm | |
| CN116502522A (en) | BP neural network-based mechanical-elastic separation track prediction method |
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 | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230103 Address after: 710077 Room 12004, Block A, Xi'an National Digital Publishing Base, No.996, Tiangu 7th Road, Yuhua Street Office, Hi tech Zone, Xi'an, Shaanxi Patentee after: XI'AN INNNO AVIATION TECHNOLOGY Co.,Ltd. Address before: Beilin District Shaanxi province Xi'an City friendship road 710002 No. 127 Patentee before: Northwestern Polytechnical University |