CN117091457A - Guided projectile navigation method and system based on deep learning - Google Patents
Guided projectile navigation method and system based on deep learning Download PDFInfo
- Publication number
- CN117091457A CN117091457A CN202310977773.4A CN202310977773A CN117091457A CN 117091457 A CN117091457 A CN 117091457A CN 202310977773 A CN202310977773 A CN 202310977773A CN 117091457 A CN117091457 A CN 117091457A
- Authority
- CN
- China
- Prior art keywords
- deep learning
- guided projectile
- navigation
- ballistic
- ballistic position
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000013135 deep learning Methods 0.000 title claims abstract description 55
- 238000000034 method Methods 0.000 title claims abstract description 50
- 238000001914 filtration Methods 0.000 claims description 24
- 230000004927 fusion Effects 0.000 claims description 14
- 238000012549 training Methods 0.000 claims description 11
- 238000004364 calculation method Methods 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 3
- 238000013473 artificial intelligence Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000010304 firing Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F42—AMMUNITION; BLASTING
- F42B—EXPLOSIVE CHARGES, e.g. FOR BLASTING, FIREWORKS, AMMUNITION
- F42B15/00—Self-propelled projectiles or missiles, e.g. rockets; Guided missiles
- F42B15/01—Arrangements thereon for guidance or control
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/20—Instruments for performing navigational calculations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/11—Complex mathematical operations for solving equations, e.g. nonlinear equations, general mathematical optimization problems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/29—Graphical models, e.g. Bayesian networks
- G06F18/295—Markov models or related models, e.g. semi-Markov models; Markov random fields; Networks embedding Markov models
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- 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
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Abstract
The invention provides a guided projectile navigation method and a guided projectile navigation system based on deep learning, which are suitable for autonomous navigation positioning of the guided projectile in the whole process. The method is suitable for guidance of guided shells in a high dynamic environment.
Description
Technical Field
The invention belongs to the field of guided projectile navigation in a high dynamic environment, and relates to a guided projectile navigation method and system based on deep learning.
Background
The guided projectile high-precision navigation positioning gives the projectile precision guidance capability, and the precision directly restricts the capability of the weapon to precisely strike the target. The flight time of the guided projectile is about 100 seconds, and belongs to the field of short-time navigation. The inertial navigation has autonomy, real-time performance and concealment, can continuously provide all navigation parameters of gestures, speeds and positions, and has very good short-term precision and stability. The shell navigation, guidance and control are generally performed by adopting an inertial navigation mode. Because the global navigation satellite system (Global Navigation Satellite System, GNSS) belongs to a wireless communication system, and satellite navigation signals belong to weak signals, the satellite signals are easy to be interfered by various complex electromagnetic environments when being transmitted in an actual battlefield environment, and are influenced by projectile structures and high-rotation (more than or equal to 20 r/s) motion disturbance, so that satellites are refused, the navigation positioning accuracy is reduced, the combat effectiveness is seriously influenced, and even military combat mission failure is caused. Aiming at the problems of high complexity such as satellite interference and random wind interference, high rotation (more than or equal to 20 r/s), high overload (more than or equal to 10000 g) and high dynamic performance, the whole-process autonomous navigation research of the guided projectile is needed to be carried out.
The method aims at autonomous navigation of the whole process of the guided projectile and mainly surrounds the research of navigation positioning methods of various multi-source sensing information fusion. The us started the precise robust inertial guided ammunition (PRIGM) project in 2015, developed an advanced navigation inertial measurement unit (NGIMU) and Advanced Inertial Microsensor (AIMS) to provide stable navigation performance under extreme conditions, and how this project progressed was unknown from the prior publications. Roux et al, the institute of san lewis law, france, effectively estimates projectile trajectories using an Extended Kalman Filter (EKF) and an incomplete right invariant extended kalman filter (incomplete R-IEKF), which focuses on the study of the filtering algorithm with little analysis of the projectile itself and its environmental characteristics.
With the development of Artificial Intelligence (AI), AI technology is increasingly being used for military applications such as target recognition, predictive maintenance, military training, decision support, or network security. The navigation technology also enters an intelligent era, the data fusion of the multi-source sensor requires the traditional positioning navigation method to advance towards the artificial intelligence direction driven by the data and the model together, and the AI technology is widely applied to the navigation field. Brossard et al propose a noise estimation AI-IMU model based on CNN structure that can adaptively estimate the measurement noise of IMU for pure inertial navigation system. Chen et al propose an IONet method of estimating displacement within a given time window using inertial sensors, superior to algorithms based on standard inertial navigation systems and model-based step estimation. With the development of navigation technology, the organic combination of AI technology and traditional methods is beginning to be a new trend. Chen et al propose DynaNet, a mixed deep learning and time-varying state space model, can carry out end-to-end training, has effectually promoted navigation accuracy. Tang et al propose a combined optimization method VINFNet combining inference modeling and data driving, the method fuses the advantages of a state space potential inference model and a depth generation network, can explicitly simulate the physical process of target motion, and constructs complex posterior distribution of target tracks through a series of reversible mapping, and is superior to the traditional method in terms of convergence, accuracy, robustness and effectiveness. Although AI technology is widely used for ground navigation, it is rarely applied to shell navigation.
Disclosure of Invention
In order to solve the problem of insufficient inertial navigation precision in a complex disturbance environment, the invention provides a guided projectile navigation method and a guided projectile navigation system based on deep learning so as to improve navigation precision.
The technical scheme adopted by the invention is as follows: a guided projectile navigation method based on deep learning comprises the following steps:
acquiring a deep learning network training data set according to a shell flight dynamics equation, and constructing a ballistic position increment prediction model by adopting an LSTM network;
and establishing a ballistic position increment error filtering model by adopting first-order Markov, and carrying out information fusion on a deep learning prediction result and an INS (inertial navigation system) calculation result to realize navigation positioning of the guided projectile.
Further, the method for establishing the ballistic position increment error filtering model by using the first-order Markov specifically comprises the following steps:
the ballistic position delta error is modeled as a first order markov process, namely:
in the formula e p As a result of the ballistic position delta error,is a correlation time matrix, w is white noise, τ x 、τ y 、τ z Respectively representing the related time of x, y and z axes;
with ballistic position delta error e p As a system state variable, constructing a Kalman filtering system equation;
ballistic position delta results using INS solutionAnd deep learning predicted ballistic position delta results Δp k D And constructing an observation equation.
Further, the Kalman filtering system equation is:
X k =Φ k,k-1 X k-1 +W k-1
in which W is k-1 Is white noise phi k,k-1 =I 3×3 -Tt s ,X k =e p 。
Further, the observation equation is:
wherein V is k To observe noise, t s Is the sampling time.
Further, obtaining the deep learning network training data set according to the shell flight dynamics equation specifically includes: and acquiring trajectory under different wind speed environments according to a guided projectile dynamics equation, inverting output data of an inertial device according to the trajectory, and outputting the output data as a deep learning network input and trajectory position increment as a deep learning network output.
Further, the LSTM network adopts 2 layers of LSTM, each layer of nodes has a number of 32, the learning rate is 0.001, the iteration number is 100, the sequence input layer is 6, and the full connection layer is 3.
Further, the ballistic position increment prediction model is:
in the method, in the process of the invention,representing the carrier angular rate measured by the gyroscope, f b Representing the triaxial specific force measured by the accelerometer, f (·) representing +.>And f b And ballistic position delta.
The guided projectile navigation system based on deep learning comprises an LSTM network, an INS module, a filtering fusion module and a guided projectile navigation module, wherein:
the training set of the LSTM network is obtained according to a shell flight dynamics equation, and ballistic position increment is output;
the INS module is used for calculating the trajectory position increment;
the filtering fusion module establishes a ballistic position increment error filtering model by adopting first-order Markov, and carries out information fusion on the deep learning prediction result and the INS calculation result;
and the guided projectile navigation module realizes guided projectile navigation positioning according to the fusion result.
Compared with the prior art, the invention has the beneficial effects that: according to the method, the output results of the gyroscope and the accelerometer are obtained through trajectory inversion and are used as the input of a deep learning network, and the position increment of the trajectory is used as the output of the deep learning network, so that the prediction of the position increment of the trajectory is realized; according to the method, a ballistic position increment error filtering model is constructed by adopting first-order Markov, and the ballistic position increment error is estimated by adopting Kalman filtering, so that the compensation of the ballistic position increment error is realized, and the navigation precision is improved; the invention realizes the autonomous navigation positioning of the whole process of the guided projectile with high precision, and is suitable for the navigation of the guided projectile in a high dynamic environment.
Drawings
Fig. 1 is a schematic diagram of a deep learning guided projectile navigation framework.
Fig. 2 is a schematic diagram of a SINS position estimation framework based on deep learning.
FIG. 3 is a block diagram of a deep learning network model.
Fig. 4 is a flow chart of a deep learning guided projectile navigation framework application.
Detailed Description
The following is a detailed description of the implementation of the present invention with reference to the accompanying drawings.
Referring to fig. 1 and 4, the invention relates to a deep learning guided projectile navigation frame. The frame is mainly divided into two parts: 1) A deep learning network; 2) And (5) information fusion filtering. Firstly, acquiring trajectory under different wind speed environments according to a guided projectile dynamics equation, establishing a deep learning network training set, constructing an INS position reckoning frame based on deep learning, outputting data according to trajectory inversion inertial devices (gyroscopes and accelerometers), taking the data as the input of the deep learning network, taking trajectory position increment as the output of the deep learning network, and completing the training of a deep learning network model. And finally, constructing a ballistic position increment error filtering model, constructing a ballistic position increment error system equation by adopting a first-order Markov process, constructing a measurement equation by combining an INS (inertial navigation system) solution ballistic position increment result and a deep learning network prediction ballistic position increment result, and finishing information fusion by adopting Kalman filtering to realize high-precision navigation positioning of the guided projectile.
The recurrence form of ballistic (i.e. navigation position) information in an inertial navigation system strapdown to a guided projectile is:
wherein P is k Representing the position coordinates of the guided projectile at the moment k; p (P) k-1 Representing the position coordinates of the guided projectile at the moment k-1;and the speed of the guided projectile in the navigation coordinate system at the time k-1 is represented. ΔP k Representing the inertia trajectory increment (i.e., position increment).
This patent adopts the degree of deep learning method to solve the problem of solving the ballistic position of guided projectile, consequently, can write the ballistic position increment as:
in the method, in the process of the invention,representing the carrier angular rate measured by the gyroscope. f (f) b Representing the triaxial specific force measured by the accelerometer. f (·) represents the functional mapping between gyroscope output angular rate and accelerometer output specific force information and ballistic position increment.
From equation (2), it can be seen that the specific position of the guided projectile can be deduced by acquiring only the ballistic position increment. The method is realized by adopting a deep learning method, and a deep learning network training set is constructed by calculating a plurality of trajectories according to ammunition models, meteorological conditions and geographic positions of firing points under different wind speed conditions.
And obtaining n ballistic trajectories according to the projectile flight dynamics equation, and obtaining different ballistic gesture and speed information. The inverted gyroscope output angular rate can be obtained from the Euler equation:
substituting the attitude and the speed into a specific force equation can obtain the inverted accelerometer output specific force:
in the method, in the process of the invention,R e for the average radius of the earth, ω ie Is the rotation angular velocity of the earth, [ Lλh ]] T For the position, respectively representing latitude, longitude and altitude,/->For speed, g represents east, north and heaven direction speeds, respectively n Is the local gravitational acceleration. />Is a gesture conversion matrix from a carrier coordinate system to a navigation coordinate system.
And obtaining output results of the gyroscope and the accelerometer according to ballistic inversion according to the formula (3) and the formula (4). With reference to fig. 2, the output results of the gyroscope and the accelerometer obtained by trajectory inversion are used as the input of the deep learning network, and the position increment of the trajectory is used as the output of the deep learning network, so that the prediction of the position increment of the trajectory is realized.
In processing time series data using deep learning, most methods use Recurrent Neural Networks (RNNs). The method adopts an LSTM deep learning network structure, adopts 2 layers of LSTM, each layer of node number is 32, the learning rate is 0.001, the iteration times are 100, the sequence input layer is 6, the full connection layer is 3, the LSTM network is adopted to realize ballistic position increment prediction, and the obtained result and an INS calculation result are fused by Kalman filtering. The deep learning network structure is shown in fig. 3, and is a prior structure and will not be described in detail herein.
Based on the analysis, the INS position increment result can be obtained through a deep learning means, and the INS can also obtain the position increment result, so that the navigation accuracy is further improved, and the INS position increment result are fused. From the ballistic position recursion process, the ballistic position increment error at the current time is only related to the position increment error at the previous time. Thus, ballistic position delta errors can be modeled as a first order markov process, namely:
in the formula e p Is a ballistic position delta error.Is a correlation time matrix. w is white noise.
With ballistic position delta error e p As a system state variable, a Kalman filtering system equation is constructed, and the discrete form is as follows:
X k =Φ k,k-1 X k-1 +W k-1 (6)
in which W is k-1 Is white noise.
Calculation of ballistic position delta results using INSAnd deep learning predicted ballistic position delta results Δp k D Constructing an observation equation:
wherein V is k To observe noise, t s Is the sampling time.
And constructing a ballistic position incremental error filtering model by combining the formula (6) and the formula (7), estimating the ballistic position incremental error by adopting Kalman filtering, and realizing the compensation of the ballistic position incremental error, thereby improving the navigation precision.
Referring to fig. 2, the present invention obtains gyroscope and accelerometer output results from ballistic inversion. And taking the output results of the gyroscope and the accelerometer obtained by trajectory inversion as the input of the deep learning network, taking the position increment of the trajectory as the output of the deep learning network, and realizing the prediction of the position increment of the trajectory.
The method realizes the autonomous navigation positioning of the guided projectile in the whole process with high precision, and is suitable for the guided projectile navigation in a high dynamic environment.
Claims (8)
1. The guided projectile navigation method based on deep learning is characterized by comprising the following steps:
acquiring a deep learning network training data set according to a shell flight dynamics equation, and constructing a ballistic position increment prediction model by adopting an LSTM network;
and establishing a ballistic position increment error filtering model by adopting first-order Markov, and carrying out information fusion on a deep learning prediction result and an INS (inertial navigation system) calculation result to realize navigation positioning of the guided projectile.
2. The guided projectile navigation method based on deep learning of claim 1, wherein the creating of the ballistic position delta error filtering model using first order markov specifically comprises:
the ballistic position delta error is modeled as a first order markov process, namely:
in the formula e p As a result of the ballistic position delta error,is a correlation time matrix, w is white noise, phi x 、τ y 、τ z Respectively representing the related time of x, y and z axes;
with ballistic position delta error e p As a system state variable, constructing a Kalman filtering system equation;
and constructing an observation equation by adopting the ballistic position increment result calculated by the INS and the deep learning predicted ballistic position increment result.
3. The guided projectile navigation method of claim 2, wherein the Kalman filtering system equation is:
X k Φ k,k-1 X k-1 +W k-1
in which W is k-1 Is white noise phi k,k-1 =I 3×3 -Tt s ,X k =e p ,t s Is the sampling time.
4. A guided projectile navigation method based on deep learning as claimed in claim 3, wherein the observation equation is:
wherein V is k In order to observe the noise it is possible,ballistic position delta results calculated for INS, +.>Ballistic position delta results are predicted for deep learning.
5. The guided projectile navigation method based on deep learning of claim 1, wherein obtaining the deep learning network training data set according to the projectile flight dynamics equation comprises: and acquiring trajectory under different wind speed environments according to a guided projectile dynamics equation, inverting output data of an inertial device according to the trajectory, and outputting the output data as a deep learning network input and trajectory position increment as a deep learning network output.
6. The guided projectile navigation method based on deep learning of claim 1, wherein the LSTM network adopts 2 layers of LSTM, each layer of nodes is 32, learning rate is 0.001, iteration number is 100, sequence input layer is 6, and full connection layer is 3.
7. The guided projectile navigation method of any one of claims 1-6, wherein the ballistic position delta prediction model is:
in the method, in the process of the invention,representing the carrier angular rate measured by the gyroscope, f b Representing the triaxial specific force measured by the accelerometer, f (·) representsAnd f b And ballistic position delta.
8. The guided projectile navigation system based on deep learning is characterized by comprising an LSTM network, an INS module, a filtering fusion module and a guided projectile navigation module, wherein:
the training set of the LSTM network is obtained according to a shell flight dynamics equation, and ballistic position increment is output;
the INS module is used for calculating the trajectory position increment;
the filtering fusion module establishes a ballistic position increment error filtering model by adopting first-order Markov, and carries out information fusion on the deep learning prediction result and the INS calculation result;
and the guided projectile navigation module realizes guided projectile navigation positioning according to the fusion result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310977773.4A CN117091457B (en) | 2023-08-03 | 2023-08-03 | Guided projectile navigation method and system based on deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310977773.4A CN117091457B (en) | 2023-08-03 | 2023-08-03 | Guided projectile navigation method and system based on deep learning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN117091457A true CN117091457A (en) | 2023-11-21 |
CN117091457B CN117091457B (en) | 2024-02-13 |
Family
ID=88770862
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310977773.4A Active CN117091457B (en) | 2023-08-03 | 2023-08-03 | Guided projectile navigation method and system based on deep learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117091457B (en) |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6163021A (en) * | 1998-12-15 | 2000-12-19 | Rockwell Collins, Inc. | Navigation system for spinning projectiles |
US20040133346A1 (en) * | 2003-01-08 | 2004-07-08 | Bye Charles T. | Attitude change kalman filter measurement apparatus and method |
CN111813146A (en) * | 2020-07-01 | 2020-10-23 | 大连理工大学 | Reentry prediction-correction guidance method based on BP neural network prediction voyage |
AU2020103939A4 (en) * | 2020-12-08 | 2021-02-11 | Xi'an University Of Architecture And Technology | Polar Integrated Navigation Algorithm of SINS / GPS Based on Grid Framework |
CN114186477A (en) * | 2021-11-02 | 2022-03-15 | 南京长峰航天电子科技有限公司 | Elman neural network-based orbit prediction algorithm |
CN114398827A (en) * | 2022-01-05 | 2022-04-26 | 北京理工大学 | Virtual gyroscope construction method based on deep learning |
CN114689047A (en) * | 2022-06-01 | 2022-07-01 | 鹏城实验室 | Deep learning-based integrated navigation method, device, system and storage medium |
-
2023
- 2023-08-03 CN CN202310977773.4A patent/CN117091457B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6163021A (en) * | 1998-12-15 | 2000-12-19 | Rockwell Collins, Inc. | Navigation system for spinning projectiles |
US20040133346A1 (en) * | 2003-01-08 | 2004-07-08 | Bye Charles T. | Attitude change kalman filter measurement apparatus and method |
CN111813146A (en) * | 2020-07-01 | 2020-10-23 | 大连理工大学 | Reentry prediction-correction guidance method based on BP neural network prediction voyage |
AU2020103939A4 (en) * | 2020-12-08 | 2021-02-11 | Xi'an University Of Architecture And Technology | Polar Integrated Navigation Algorithm of SINS / GPS Based on Grid Framework |
CN114186477A (en) * | 2021-11-02 | 2022-03-15 | 南京长峰航天电子科技有限公司 | Elman neural network-based orbit prediction algorithm |
CN114398827A (en) * | 2022-01-05 | 2022-04-26 | 北京理工大学 | Virtual gyroscope construction method based on deep learning |
CN114689047A (en) * | 2022-06-01 | 2022-07-01 | 鹏城实验室 | Deep learning-based integrated navigation method, device, system and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN117091457B (en) | 2024-02-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Dai et al. | An INS/GNSS integrated navigation in GNSS denied environment using recurrent neural network | |
CN102445200B (en) | Microminiature personal combined navigation system as well as navigating and positioning method thereof | |
CN106871928B (en) | Strap-down inertial navigation initial alignment method based on lie group filtering | |
CN104655131B (en) | Inertial navigation Initial Alignment Method based on ISTSSRCKF | |
CN108362288B (en) | Polarized light SLAM method based on unscented Kalman filtering | |
CN109931955B (en) | Initial alignment method of strap-down inertial navigation system based on state-dependent lie group filtering | |
CN108387236B (en) | Polarized light SLAM method based on extended Kalman filtering | |
CN106643715A (en) | Indoor inertial navigation method based on bp neural network improvement | |
CN105737823A (en) | GPS/SINS/CNS integrated navigation method based on five-order CKF | |
Bezick et al. | Inertial navigation for guided missile systems | |
CN109059914B (en) | Projectile roll angle estimation method based on GPS and least square filtering | |
CN105806363A (en) | Alignment method of an underwater large misalignment angle based on SINS (Strapdown Inertial Navigation System)/DVL (Doppler Velocity Log) of SRQKF (Square-root Quadrature Kalman Filter) | |
CN115248038B (en) | SINS/BDS combined navigation engineering algorithm under emission system | |
CN112797985A (en) | Indoor positioning method and indoor positioning system based on weighted extended Kalman filtering | |
CN111238469A (en) | Unmanned aerial vehicle formation relative navigation method based on inertia/data chain | |
Liu et al. | Interacting multiple model UAV navigation algorithm based on a robust cubature Kalman filter | |
CN115855049A (en) | SINS/DVL navigation method based on particle swarm optimization robust filtering | |
CN111964675A (en) | Intelligent aircraft navigation method for blackout area | |
CN111982126B (en) | Design method of full-source BeiDou/SINS elastic state observer model | |
CN109211232A (en) | A kind of shell Attitude estimation method based on least squares filtering | |
CN112325878A (en) | Ground carrier combined navigation method based on UKF and air unmanned aerial vehicle node assistance | |
CN110873577B (en) | Underwater rapid-acting base alignment method and device | |
CN113029123A (en) | Multi-AUV collaborative navigation method based on reinforcement learning | |
CN117091457B (en) | Guided projectile navigation method and system based on deep learning | |
Al-Jlailaty et al. | Efficient attitude estimators: A tutorial and survey |
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 |