CN112505737A - GNSS/INS combined navigation method based on Elman neural network online learning assistance - Google Patents

GNSS/INS combined navigation method based on Elman neural network online learning assistance Download PDF

Info

Publication number
CN112505737A
CN112505737A CN202011281769.7A CN202011281769A CN112505737A CN 112505737 A CN112505737 A CN 112505737A CN 202011281769 A CN202011281769 A CN 202011281769A CN 112505737 A CN112505737 A CN 112505737A
Authority
CN
China
Prior art keywords
gnss
neural network
output
navigation
ins
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
Application number
CN202011281769.7A
Other languages
Chinese (zh)
Other versions
CN112505737B (en
Inventor
王庆
杨高朝
张欢
张坤
宋爱国
赵国普
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Southeast University
Original Assignee
Southeast University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Southeast University filed Critical Southeast University
Priority to CN202011281769.7A priority Critical patent/CN112505737B/en
Publication of CN112505737A publication Critical patent/CN112505737A/en
Application granted granted Critical
Publication of CN112505737B publication Critical patent/CN112505737B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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
    • G01S19/38—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system
    • G01S19/39—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system the satellite radio beacon positioning system transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
    • G01S19/42—Determining position
    • G01S19/45—Determining position by combining measurements of signals from the satellite radio beacon positioning system with a supplementary measurement
    • G01S19/47—Determining position by combining measurements of signals from the satellite radio beacon positioning system with a supplementary measurement the supplementary measurement being an inertial measurement, e.g. tightly coupled inertial
    • 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
    • G01C21/10—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration
    • G01C21/12—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning
    • G01C21/16—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning by integrating acceleration or speed, i.e. inertial navigation
    • G01C21/165—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning by integrating acceleration or speed, i.e. inertial navigation combined with non-inertial navigation instruments
    • 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
    • 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
    • G06N3/045—Combinations of networks
    • 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
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • 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
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Biomedical Technology (AREA)
  • General Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Software Systems (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Automation & Control Theory (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Navigation (AREA)

Abstract

The invention discloses a GNSS/INS combined navigation method based on the assistance of Elman neural network online learning, which comprises two parts: (1) the method comprises the steps of realizing a conventional GNSS/MEMS-INS integrated navigation algorithm, designing a Kalman filter to fuse GNSS signals and inertial navigation data, and outputting fused navigation data; (2) the method comprises the steps of designing a neural network model on the basis of the step (1), and then respectively taking inertial navigation data obtained in the step (1) and data output by a Kalman filter as sample input and sample output of a training neural network to train the neural network model.

Description

GNSS/INS combined navigation method based on Elman neural network online learning assistance
Technical Field
The invention belongs to the technical field of integrated navigation methods, and particularly relates to a GNSS/INS integrated navigation method based on the assistance of Elman neural network online learning.
Background
In the aspect of navigation technology, the most applications are currently available, and the most mature navigation modes include inertial navigation and satellite navigation. The GNSS satellite navigation has the advantages of being global, all-weather and high in long-time positioning accuracy, but has the defect that signals are easily interfered and shielded. In strong electromagnetic environment and when high-rise shelters from, signal quality worsens to its output frequency is limited, generally is 1-10Hz, and the output is discontinuous, and on the occasion that needs quick update information, for example on the unmanned aerial vehicle system that mobility and real-time requirement are higher, GNSS satellite navigation's shortcoming just highlights. The INS inertial navigation system is a fully autonomous navigation mode, so that the system has strong concealment and anti-interference capability, continuous output information and high positioning accuracy in a short time. However, due to the characteristics of the MEMS-INS device, the gyroscope and the accelerometer have errors such as initial zero offset and random drift, and the errors become larger and larger with the cumulative effect of time, so that the positioning accuracy is poor for a long time, and finally the attitude and position information of the unmanned aerial vehicle cannot be accurately reflected.
In order to overcome the defects of the two types of navigation, the common method is to fuse the satellite navigation signals and the inertial navigation signals through kalman filtering, and make up the respective defects by using the respective advantages. However, under some special environmental conditions, such as a signal blocking area and an environment with many obstructions, the satellite signal may be lost, and at this time, the navigation system only depends on pure inertial navigation, and the error of the navigation data becomes larger and larger as time goes on. Therefore, there is a need to develop a method that can replace the function of GNSS and complete the output of navigation data in cooperation with inertial navigation in the case of GNSS signal loss.
Disclosure of Invention
In order to solve the problems, the invention discloses a GNSS/INS combined navigation method based on the assistance of the Elman neural network on-line learning, and provides a method for predicting the output error of an inertial navigation system of an unmanned aerial vehicle navigation system under the condition that a GNSS signal is lost, and compensating and correcting the output of the inertial navigation system by using error data. The inertial navigation system can output accurate navigation data under the assistance of a neural network algorithm under the condition that GNSS signals of the navigation system are lost.
In order to achieve the purpose, the technical scheme of the invention is as follows:
a GNSS/INS combined navigation method based on the assistance of Elman neural network online learning comprises the following steps:
1. the method comprises the steps of realizing a conventional GNSS/MEMS-INS integrated navigation algorithm, designing a Kalman filter to fuse GNSS signals and inertial navigation data, and outputting fused navigation data;
2. designing a neural network model on the basis of the step 1, then respectively taking the inertial navigation data obtained in the step 1 and data output by a Kalman filter as sample input and sample output of a training neural network, training the neural network model, predicting an inertial navigation output error by using the trained neural network model when a GNSS signal is lost, and compensating and correcting the inertial navigation by using the predicted error.
Further, the GNSS/INS integrated navigation method based on the Elman neural network online learning assistance provided by the invention specifically comprises the following steps:
s1, mounting the IMU and the GNSS equipment, and matching the IMU coordinate axis with the carrier axis; the LGM851 vehicle-mounted integrated navigation positioner based on the Beidou satellite communication is adopted as a navigation calculation unit, the data safety is ensured by adopting the LoRa communication, a serial port data line is connected with a computer and an integrated navigation system, and the computer is used for processing the integrated navigation data in real time;
s2, calibrating external parameters of the INS and the GNSS antenna, calibrating the installation external parameters of the INS and the GNSS receiver antenna by using a total station, and inputting the calibrated external parameters into the system as initial values;
s3, mounting the combined equipment on a trolley, testing in an experimental site, and starting to collect and process data, wherein the method specifically comprises the following steps:
initial alignment of IMU under GNSS assistance under S3-1 dynamics
The initial alignment is done by means of dynamic movements, since the GNSS is able to give the velocity of the carrier relative to the local geographical coordinate system, assuming that the position of the GNSS receiver under ECEF has been acquiredePGNSSAnd velocityeVGNSSVelocity is calculated by formula
Figure BDA0002781043450000021
Conversion to geographical coordinate systemnVG，
Figure BDA0002781043450000022
Determining yaw and pitch angles by projection of velocity in a geographic coordinate system, the formula is as follows:
Figure BDA0002781043450000023
since for the vehicle carrier the roll is directly assigned a value of 0, while a large initial variance is given.
S3-2 inertial navigation solution based on ECEF coordinate system
S3-2-1, in order to initialize the IMU better, firstly, the carrier is moved slowly, the neural network learning part is not started temporarily, and after the initialization is finished, the neural network learning part is started; reading carrier parameters including angular velocity information acquired by inertial navigation module
Figure BDA0002781043450000024
Acceleration vector
Figure BDA0002781043450000025
Acquisition of GNSS reception using a combined systemPosition and speed of the machineePk+1,GAndeVk+1,G。
s3-2-2, obtaining real-time rotation quaternion by solving quaternion differential equationeq＝[q0 q1 q2 q3]T；
Figure BDA0002781043450000026
S3-2-3, carrier acceleration vector obtained according to step S3-2-1
Figure BDA0002781043450000027
And the attitude quaternion solved in step S3-2-2
Figure BDA0002781043450000028
Solving the differential equation of the following formula to obtain the speed information of the carrier in three directions under the navigation coordinate system:
Figure BDA0002781043450000029
wherein V is [ V ]E VN VU]TRespectively the speed in the middle east, north and sky directions of the geographic coordinate system,
Figure BDA00027810434500000210
is the angular velocity of the earth's rotation,eg is the acceleration of gravity under a geographic coordinate system;
s3-2-4, respectively calculating position parameters of inertial navigation output according to the following formula;
ePk+1,INS≈ePk,EKF+(eVk,EKF+eVk+1,INS)Δt/2 (5)
constraint updating under assistance of S3-3 Elman neural network
When the GNSS signals are good, the neural network is connected into the system to predict the output error of the inertial navigation and compensate and correct the output of the inertial navigation.
S3-3-1Elman neural network design
The mathematical model of the Elman neural network is:
Figure BDA0002781043450000031
wherein: y (k) is the output of the output layer; x (k-1) is an external input of the input layer; u. ofc(k) Is the output of the receiving layer; u (k) is the output of the hidden layer;
Figure BDA0002781043450000032
connecting the input layer with the hidden layer by a weight value;
Figure BDA0002781043450000033
connecting the weight values of the bearer layer and the hidden layer;
Figure BDA0002781043450000034
the weights are connected between the hidden layer and the output layer. The hidden layer of the Elman neural network can adopt a bipolar Sigmoid function shown in an expression (7) as an excitation function, and the output layer adopts a Pureline activation function, so that an expression (8) can be obtained:
f(a)＝(1-e-a)/(1+e-a) (7)
Figure BDA0002781043450000035
the following equations (6) to (8) can be derived:
Figure BDA0002781043450000036
wherein f (a) epsilon (-1, 1);
Figure BDA0002781043450000037
and
Figure BDA0002781043450000038
respectively representing the network connection weight at the moment k-1 and the moment k-2. u. ofc(k) And the connection weight of the historical time is related, so that the characteristic of dynamic recursion is embodied.
The Elman neural network adopts a BP algorithm to correct the weight, and the learning index function adopts an error sum of squares function:
Figure BDA0002781043450000039
s3-3-2, performing online learning according to the outputs of the inertial sensor and the EKF;
s3-3-2-1, using the GNSS output position and speed to correct the prediction data obtained from the inertial sensor in step S1, and using Kalman filtering to obtain the filtered prediction data
Figure BDA00027810434500000310
And
Figure BDA00027810434500000311
as shown in formula (11).
The GNSS measurement status in EKF is updated as follows:
Figure BDA00027810434500000312
wherein
Figure BDA0002781043450000041
And
Figure BDA0002781043450000042
respectively, the inertial sensor output values;ePk+1,GandeVk+1,Goutput values for GNSS position and velocity, respectively;
Figure BDA0002781043450000043
and
Figure BDA0002781043450000044
respectively, a Kalman filterWave solution.
S3-3-2-2 synchronously trains the neural network model. Designing a neural network model on the basis of the step S3-3-2-1, then obtaining the relative change value between the epochs of the inertial navigation data and the output data of the Kalman filter from the step S3-3-2-1, and when the number of the GNSS available satellites is large, designing a neural network model according to the step S3-3-2-1, and obtaining the relative change value between the epochs of the inertial navigation data and the output data of the Kalman filter
Figure BDA0002781043450000045
And
Figure BDA0002781043450000046
the signal is respectively used as the input and the output of an Elman neural network and added into a training sample to be subjected to online learning, so that the network has the capability of predicting speed and position variation, the position, the rotation quaternion and the speed of the epoch are obtained according to the position and the speed at the previous moment, and when the training error meets a certain threshold value, the training model is reliable, and the system can be assisted to forecast;
in order to ensure the timeliness of the system, the number of samples is limited by adopting a fixed window method, and when the number of samples reaches a certain number, the older samples are removed out of the window in order to ensure the timeliness of the samples.
Figure BDA0002781043450000047
Figure BDA0002781043450000048
Wherein
Figure BDA0002781043450000049
Respectively, the inter-epoch variation of the output value of the inertial sensor;
Figure BDA00027810434500000410
respectively are the variation between epochs of the output values after Kalman filtering;
Figure BDA00027810434500000411
and
Figure BDA00027810434500000412
respectively is an output value of the inertial sensor at the moment k;
Figure BDA00027810434500000413
and
Figure BDA00027810434500000414
and respectively outputting solutions for the k moment Kalman filtering.
S3-3-3, when online learning meets a threshold value and can carry out prediction, classifying according to satellite signal occlusion and non-occlusion;
s3-3-3-1, when the GNSS satellite signal is blocked, the INS can work normally due to the attribute of the INS; GNSS data at the moment before the invalidation is an initial value of filtering, which is equivalent to an accurate value, so that divergence of INS precision caused by long navigation time can be avoided; when the GNSS signal is lost, the trained neural network model is used for predicting the output error of the inertial navigation, and the prediction error is used for compensating and correcting the inertial navigation, and firstly, the error is obtained by pre-integrating according to an inertial sensor
Figure BDA00027810434500000415
As Elman neural network for prediction, the prediction data of the neural network is output
Figure BDA00027810434500000416
And
Figure BDA00027810434500000417
EKF updates were performed as observation constraints.
Neural network predicted variance
Figure BDA00027810434500000418
And
Figure BDA00027810434500000419
the obtained predicted values of the position, the rotation quaternion and the speed at the time k +1 are:
Figure BDA0002781043450000051
the constraint is updated as:
Figure BDA0002781043450000052
wherein,
Figure BDA0002781043450000053
and
Figure BDA0002781043450000054
are the state vectors output at step S3-1, respectively, from the inertial sensors;
Figure BDA0002781043450000055
and
Figure BDA0002781043450000056
the predicted values of the position, the rotational quaternion and the velocity at the time k +1 obtained by the neural network are respectively.
S3-3-3-2 when GNSS signals are normal, two strategies are available: 1. directly updating the position, the speed and the rotation quaternion at the moment when the inertial sensor outputs k +1 as a state vector, and then performing constraint updating by using the position and the speed output by the GNSS (global navigation satellite system), as shown in step S3-3-2; 2. inputting the position, the speed and the variation of the rotation quaternion at the moment when the inertial sensor outputs k +1 into an Elman neural network, and updating the state vector of the EKF by taking the position, the rotation quaternion and the variation of the speed output by the Elman neural network as a pseudo observation value, as shown in a formula (16); constraint updates are then made based on the GNSS output position and velocity, as shown in equation (17); finally obtained by inertial sensors
Figure BDA0002781043450000057
And
Figure BDA0002781043450000058
and obtained by using EKF
Figure BDA0002781043450000059
Figure BDA00027810434500000510
And
Figure BDA00027810434500000511
and adding the training data into an online learning system of a neural network to perform real-time learning training.
The EKF status is updated as:
Figure BDA00027810434500000512
the GNSS measurement status in EKF is updated as follows:
Figure BDA00027810434500000513
Figure BDA00027810434500000514
and
Figure BDA00027810434500000515
respectively predicting the variation according to the output neural network
Figure BDA00027810434500000516
And
Figure BDA00027810434500000517
the predicted values of the position, the rotation quaternion and the speed at the k +1 moment are obtained;ePk+1,GandeVk+1,Goutput values for GNSS position and velocity, respectively.
The invention has the beneficial effects that:
the invention discloses a GNSS/INS combined navigation method based on the on-line learning assistance of an Elman neural network, which provides a method for predicting the output error of an inertial navigation system of an unmanned aerial vehicle navigation system under the condition of GNSS signal loss and compensating and correcting the output of the inertial navigation system by using error data under the special environment conditions, such as a signal blocking area, an environment with more shelters or a long sheltering time. The inertial navigation system can output accurate navigation data under the assistance of a neural network algorithm under the condition that GNSS signals of the navigation system are lost.
Drawings
FIG. 1 is a flow chart of a GNSS/INS integrated navigation based on a genetic neural network (there is a sufficient number of GNSS satellites available).
FIG. 2 is a flow chart of GNSS/INS integrated navigation based on the genetic neural network (when the number of GNSS satellites is insufficient).
Fig. 3 is a diagram of the Elman neural network architecture.
Fig. 4 is a dynamic quasi-resolution graph.
FIG. 5 is a diagram of an integrated navigation device.
Detailed Description
The present invention will be further illustrated with reference to the accompanying drawings and specific embodiments, which are to be understood as merely illustrative of the invention and not as limiting the scope of the invention.
The invention relates to a GNSS/INS combined navigation method based on the Elman neural network online learning assistance, which specifically comprises the following steps:
s1, mounting the IMU and the GNSS equipment, and fitting the IMU coordinate axis with the carrier axis as much as possible in order to reduce the lever arm error; FIG. 5 is a diagram of a GNSS/INS based integrated navigation hardware system. The invention adopts the LGM851 vehicle-mounted integrated navigation positioner based on the Beidou satellite communication as a navigation calculation unit, adopts LoRa communication to ensure data safety, uses a serial port data line to connect a computer and an integrated navigation system, and uses the computer to process the integrated navigation data in real time.
S2, calibrating external parameters of the INS and the GNSS antenna, calibrating the installation external parameters of the INS and the GNSS receiver antenna by using a total station, and inputting the calibrated external parameters into the system as initial values;
s3, mounting the combined equipment on a trolley, testing in an experiment site of a four-storey building school district of university in the east and south, and starting to collect and process data, wherein the method specifically comprises the following steps:
initial alignment of IMU under GNSS assistance under S3-1 dynamics
In the GNSS/INS integrated navigation system, initial alignment is one step of comparison, and the inertial navigation system adopts a dead reckoning algorithm to realize continuous autonomous navigation so as to obtain full navigation parameters (position, speed and attitude information), so that the acquisition of the initial state of the first epoch is particularly critical. Meanwhile, the GNSS can conveniently give out the position and the speed (if the receiver can output Doppler frequency shift, the instantaneous speed can be directly solved through Doppler integration, if not, the average speed can be obtained through position difference between receiver epochs), so the key task of initial alignment falls into the determination of the attitude. Therefore, initial alignment of the IMU, mainly referred to as pose initialization, poses accuracy becomes a major factor affecting navigation accuracy. For the static coarse alignment initialization of the low-precision MEMS IMU, the gyroscope has low measurement precision and cannot sense the rotational angular velocity of the earth, so the coarse alignment can only depend on dynamic motion to complete the initial alignment. Since GNSS can give the velocity of the carrier relative to the local geographical coordinate system (system n). It is assumed that the position of the GNSS receiver under ECEF has been acquiredePGNSSAnd velocityeVGNSSThe speed can be calculated by formula
Figure BDA0002781043450000071
(can be directly solved) to be converted into a geographic coordinate systemnVG. As shown in particular in fig. 4.
Figure BDA0002781043450000072
Determining yaw and pitch angles by projection of velocity under n system, the schematic and formula are as follows:
Figure BDA0002781043450000073
since for the vehicle carrier, the roll can be directly assigned a value of 0, while a larger initial variance is given.
S3-2 inertial navigation solution based on ECEF coordinate system
S3-2-1, in order to initialize IMU better, firstly, the carrier is moved slowly, the neural network learning part is not started temporarily, and after the initialization is completed, the neural network learning part is started. Reading carrier parameters including angular velocity information acquired by inertial navigation module
Figure BDA0002781043450000074
Acceleration vector
Figure BDA0002781043450000075
Obtaining position and velocity of a GNSS receiver using a combined systemePk+1,GAndeVk+1,G。
s3-2-2, obtaining real-time rotation quaternion by solving quaternion differential equationeq＝[q0 q1 q2 q3]T；
Figure BDA0002781043450000076
S3-2-3, carrier acceleration vector obtained according to step S3-2-1
Figure BDA0002781043450000077
And the attitude quaternion solved in step S3-2-2
Figure BDA0002781043450000078
Solving the differential equation of the following formula to obtain the speed information of the carrier in three directions under the navigation coordinate system:
Figure BDA0002781043450000079
wherein V is [ V ]E VN VU]TRespectively the speed in the middle east, north and sky directions of the geographic coordinate system,
Figure BDA00027810434500000710
is the angular velocity of the earth's rotation,eg is the acceleration of gravity under a geographic coordinate system;
and S3-2-4, respectively obtaining the position parameters of the inertial navigation output by the following formula.
ePk+1,INS≈ePk,EKF+(eVk,EKF+eVk+1,INS)Δt/2 (22)
Constraint updating under assistance of S3-3 Elman neural network
When the GNSS signals are good, the neural network is connected into the system to predict the output error of the inertial navigation and compensate and correct the output of the inertial navigation. As shown in fig. 2.
S3-3-1Elman neural network design
The Elman neural network is a typical dynamic recurrent neural network, which is characterized in that on the basis of a basic structure of a BP network, a carrying layer is added on a hidden layer and used as a one-step delay operator to achieve the purpose of memory, so that a system has the capability of adapting to time-varying characteristics, the global stability of the network is enhanced, and the Elman neural network has stronger computing capability than a feedforward neural network and can be used for solving the problem of quick optimization. The Elman neural network structure is shown in fig. 1, and has 4 layers, namely an input layer, a hidden layer, a carrying layer and an output layer. As shown in particular in figure 3. The mathematical model of the Elman neural network is:
Figure BDA0002781043450000081
wherein: y (k) is the output of the output layer; x (k-1) is an external input of the input layer; u. ofc(k) Is the output of the receiving layer; u (k) is the output of the hidden layer;
Figure BDA0002781043450000082
connecting the input layer with the hidden layer by a weight value;
Figure BDA0002781043450000083
connecting the weight values of the bearer layer and the hidden layer;
Figure BDA0002781043450000084
the weights are connected between the hidden layer and the output layer. The hidden layer of the Elman neural network can adopt a bipolar Sigmoid function shown in an expression (7) as an excitation function, and the output layer adopts a Pureline activation function, so that an expression (8) can be obtained:
f(a)＝(1-e-a)/(1+e-a)
(24)
Figure BDA0002781043450000085
the following equations (6) to (8) can be derived:
Figure BDA0002781043450000086
wherein f (a) epsilon (-1, 1);
Figure BDA0002781043450000087
and
Figure BDA0002781043450000088
respectively representing the network connection weight at the moment k-1 and the moment k-2. u. ofc(k) And the connection weight of the historical time is related, so that the characteristic of dynamic recursion is embodied.
The Elman neural network adopts a BP algorithm to correct the weight, and the learning index function adopts an error sum of squares function:
Figure BDA0002781043450000089
s3-3-2, performing online learning according to the outputs of the inertial sensor and the EKF;
s3-3-2-1 uses the GNSS output position and velocity to predict data obtained by the inertial sensor in step S1Corrected, filtered by Kalman filtering
Figure BDA00027810434500000810
And
Figure BDA00027810434500000811
as shown in formula (11).
The GNSS measurement status in EKF is updated as follows:
Figure BDA00027810434500000812
wherein
Figure BDA00027810434500000813
And
Figure BDA00027810434500000814
respectively, the inertial sensor output values;ePk+1,GandeVk+1,Goutput values for GNSS position and velocity, respectively;
Figure BDA00027810434500000815
and
Figure BDA00027810434500000816
respectively kalman filter solutions.
S3-3-2-2 synchronously trains the neural network model. Designing a neural network model on the basis of the step S3-3-2-1, then solving the inertial navigation data obtained in the step S3-3-2-1 and the relative change value between epochs of the output data of the Kalman filter (namely the relative change quantity of the position, the rotation quaternion and the speed between k and k +1 epochs, an inertial sensor can be solved according to pre-integration, and the relative change quantity after Kalman filtering can be solved according to the formulas (12) and (13)), and specifically realizing the steps as shown in the figure 1, when the number of satellites available in the GNSS is large, carrying out the steps of
Figure BDA0002781043450000091
And
Figure BDA0002781043450000092
and respectively used as input and output of an Elman neural network and added into a training sample to perform online learning, and if the lever arm is compensated, the influence of time delay errors between the GNSS and the INS is temporarily not considered. The network has the capability of predicting speed and position variation, the position, the rotation quaternion and the speed of the epoch are obtained according to the position and the speed at the previous moment, and when the training error meets a certain threshold value, the training model is reliable, so that the system can be assisted to predict;
in order to ensure the timeliness of the system, the number of samples is limited by adopting a fixed window method, and when the number of samples reaches a certain number, the older samples are removed out of the window in order to ensure the timeliness of the samples.
Figure BDA0002781043450000093
Figure BDA0002781043450000094
Wherein
Figure BDA0002781043450000095
Respectively, the inter-epoch variation of the output value of the inertial sensor;
Figure BDA0002781043450000096
respectively are the variation between epochs of the output values after Kalman filtering;
Figure BDA0002781043450000097
and
Figure BDA0002781043450000098
respectively is an output value of the inertial sensor at the moment k;
Figure BDA0002781043450000099
and
Figure BDA00027810434500000910
and respectively outputting solutions for the k moment Kalman filtering.
S3-3-3, when online learning meets a threshold value and can carry out prediction, classifying according to satellite signal occlusion and non-occlusion;
s3-3-3-1 when GNSS satellite signals are occluded, the INS can still work normally due to its properties. The GNSS data at the moment before the invalidation is the initial value of the filtering, which is equivalent to the accurate value, so that the dispersion of the INS precision caused by long navigation time can be avoided. When the GNSS signal is lost, the trained neural network model is used to predict the inertial navigation output error, and the inertial navigation is compensated and corrected by using the predicted error, which is specifically implemented as the step shown in fig. 2. First of all obtained by pre-integration of inertial sensors
Figure BDA00027810434500000911
As Elman neural network for prediction, the prediction data of the neural network is output
Figure BDA00027810434500000912
And
Figure BDA00027810434500000913
EKF updates were performed as observation constraints.
Neural network predicted variance
Figure BDA00027810434500000914
And
Figure BDA00027810434500000915
the obtained predicted values of the position, the rotation quaternion and the speed at the time k +1 are:
Figure BDA00027810434500000916
the constraint is updated as:
Figure BDA0002781043450000101
wherein,
Figure BDA0002781043450000102
and
Figure BDA0002781043450000103
are the state vectors output at step S3-1, respectively, from the inertial sensors;
Figure BDA0002781043450000104
and
Figure BDA0002781043450000105
the predicted values of the position, the rotational quaternion and the velocity at the time k +1 obtained by the neural network are respectively.
S3-3-3-2 when GNSS signals are normal, we have two strategies to apply. 1. Directly updating the position, the speed and the rotation quaternion at the moment when the inertial sensor outputs k +1 as a state vector, and then performing constraint updating by using the position and the speed output by the GNSS (global navigation satellite system), as shown in step S3-3-2; 2. inputting the position, the speed and the variation of the rotation quaternion at the moment when the inertial sensor outputs k +1 into an Elman neural network, and updating the state vector of the EKF by taking the position, the rotation quaternion and the variation of the speed output by the Elman neural network as a pseudo observation value, as shown in a formula (16); constraint updates are then made based on the GNSS output position and velocity, as shown in equation (17); finally obtained by inertial sensors
Figure BDA0002781043450000106
And
Figure BDA0002781043450000107
and obtained by using EKF
Figure BDA0002781043450000108
And
Figure BDA0002781043450000109
online learning system for joining to neural networkAnd performing real-time learning training. As in fig. 1.
The EKF status is updated as:
Figure BDA00027810434500001010
the GNSS measurement status in EKF is updated as follows:
Figure BDA00027810434500001011
Figure BDA00027810434500001012
and
Figure BDA00027810434500001013
respectively predicting the variation according to the output neural network
Figure BDA00027810434500001014
And
Figure BDA00027810434500001015
the predicted values of the position, the rotation quaternion and the speed at the k +1 moment are obtained;ePk+1,GandeVk+1,Goutput values for GNSS position and velocity, respectively.
The technical means disclosed in the invention scheme are not limited to the technical means disclosed in the above embodiments, but also include the technical scheme formed by any combination of the above technical features.

Claims (2)

1. A GNSS/INS combined navigation method based on the assistance of Elman neural network online learning is characterized in that: the method comprises the following steps:
(1) the method comprises the steps of realizing a conventional GNSS/MEMS-INS integrated navigation algorithm, designing a Kalman filter to fuse GNSS signals and inertial navigation data, and outputting fused navigation data;
(2) designing a neural network model on the basis of the step (1), then respectively taking the inertial navigation data obtained in the step (1) and the data output by the Kalman filter as sample input and sample output of a training neural network, training the neural network model, predicting an inertial navigation output error by using the trained neural network model when the GNSS signal is lost, and compensating and correcting the inertial navigation by using the predicted error.
2. The integrated GNSS/INS navigation method based on the Elman neural network online learning assistance as claimed in claim 1, wherein: the method specifically comprises the following steps:
s1, mounting the IMU and the GNSS equipment, and matching the IMU coordinate axis with the carrier axis; the LGM851 vehicle-mounted integrated navigation positioner based on the Beidou satellite communication is adopted as a navigation calculation unit, the data safety is ensured by adopting the LoRa communication, a serial port data line is connected with a computer and an integrated navigation system, and the computer is used for processing the integrated navigation data in real time;
s2, calibrating external parameters of the INS and the GNSS antenna, calibrating the installation external parameters of the INS and the GNSS receiver antenna by using a total station, and inputting the calibrated external parameters into the system as initial values;
s3, mounting the combined equipment on a trolley, testing in an experimental site, and starting to collect and process data, wherein the method specifically comprises the following steps:
initial alignment of IMU under GNSS assistance under S3-1 dynamics
The initial alignment is done by means of dynamic movements, since the GNSS is able to give the velocity of the carrier relative to the local geographical coordinate system, assuming that the position of the GNSS receiver under ECEF has been acquiredePGNSSAnd velocityeVGNSSVelocity is calculated by formula
Figure FDA0002781043440000011
Conversion to geographical coordinate systemnVG，
Figure FDA0002781043440000012
Determining yaw and pitch angles by projection of velocity in a geographic coordinate system, the formula is as follows:
Figure FDA0002781043440000013
as for the vehicle carrier, the roll is directly assigned with 0, and a larger initial variance is given;
s3-2 inertial navigation solution based on ECEF coordinate system
S3-2-1, in order to initialize the IMU better, firstly, the carrier is moved slowly, the neural network learning part is not started temporarily, and after the initialization is finished, the neural network learning part is started; reading carrier parameters including angular velocity information acquired by inertial navigation module
Figure FDA0002781043440000014
Acceleration vector
Figure FDA0002781043440000015
Obtaining position and velocity of a GNSS receiver using a combined systemePk+1,GAndeVk+1,G，
s3-2-2, obtaining real-time rotation quaternion by solving quaternion differential equationeq＝[q0 q1 q2 q3]T；
eqk+1,INS＝eqk,EKF(I+[w×]) (3)
S3-2-3, carrier acceleration vector obtained according to step S3-2-1
Figure FDA0002781043440000021
And the attitude quaternion solved in step S3-2-2
Figure FDA0002781043440000022
Solving the differential equation of the following formula to obtain three of the carrier under the navigation coordinate systemSpeed information in each direction:
Figure FDA0002781043440000023
wherein V is [ V ]E VN VU]TRespectively the speed in the middle east, north and sky directions of the geographic coordinate system,
Figure FDA0002781043440000024
is the angular velocity of the earth's rotation,eg is the acceleration of gravity under a geographic coordinate system;
s3-2-4, respectively calculating position parameters of inertial navigation output according to the following formula;
ePk+1,INS≈ePk,EKF+(eVk,EKF+eVk+1,INS)Δt/2 (5)
constraint updating under assistance of S3-3 Elman neural network
When the GNSS signal is good, the neural network is connected into the system to predict the output error of the inertial navigation and compensate and correct the output of the inertial navigation,
s3-3-1Elman neural network design
The mathematical model of the Elman neural network is:
Figure FDA0002781043440000025
wherein: y (k) is the output of the output layer; x (k-1) is an external input of the input layer; u. ofc(k) Is the output of the receiving layer; u (k) is the output of the hidden layer; w is a1Connecting the input layer with the hidden layer by a weight value; w is a2Connecting the weight values of the bearer layer and the hidden layer; w is a3For connecting the weight values of the hidden layer and the output layer, the hidden layer of the Elman neural network adopts a bipolar Sigmoid function shown in formula (7) as an excitation function, and the output layer adopts a Pureline activation function to obtain a formula (8):
f(a)＝(1-e-a)/(1+e-a) (7)
y(k)＝w3u(k) (8)
the following equations (6) to (8) are derived:
uc(k)＝f(w2(k-1)uc(k-1)+w1(k-2)uc(k-2)) (9)
wherein f (a) epsilon (-1, 1); w is a2(k-1) and w1(k-2) represents the network connection weight at k-1 and k-2, respectively, uc(k) The connection weight value of the history time is related to the connection weight value of the history time, the characteristic of dynamic recursion is embodied,
the Elman neural network adopts a BP algorithm to correct the weight, and the learning index function adopts an error sum of squares function:
Figure FDA0002781043440000031
s3-3-2, performing online learning according to the outputs of the inertial sensor and the EKF;
s3-3-2-1, correcting the prediction data obtained by the inertial sensor in the step S1 by adopting the position and the speed output by the GNSS, and obtaining the filtered prediction data through Kalman filtering
Figure FDA0002781043440000032
And
Figure FDA0002781043440000033
as shown in the formula (11),
the GNSS measurement status in EKF is updated as follows:
Figure FDA0002781043440000034
wherein
Figure FDA0002781043440000035
And
Figure FDA0002781043440000036
respectively, the inertial sensor output values;ePk+1,GandeVk+1,Goutput values for GNSS position and velocity, respectively;
Figure FDA0002781043440000037
and
Figure FDA0002781043440000038
respectively, a kalman filter solution is used,
s3-3-2-2 synchronously training the neural network model, designing the neural network model on the basis of the step S3-3-2-1, and then obtaining the relative change value between the inertial navigation data obtained in the step S3-3-2-1 and the epoch of the Kalman filter output data, when the number of satellites available for the GNSS is large, the method comprises the steps of
Figure FDA0002781043440000039
And
Figure FDA00027810434400000310
respectively serving as input and output of an Elman neural network, adding the input and output into a training sample to perform online learning, enabling the network to have the capability of predicting speed and position variation, obtaining the position, rotation quaternion and speed of the epoch according to the position and speed at the previous moment, and when a training error meets a certain threshold value, indicating that a training model is reliable and assisting a system to perform prediction;
in order to ensure the timeliness of the system, the number of samples is limited by adopting a fixed window method, when the number of samples reaches a certain number, in order to ensure the timeliness of the samples, the older samples are removed from a window,
Figure FDA00027810434400000311
Figure FDA00027810434400000312
wherein
Figure FDA00027810434400000313
Respectively, the inter-epoch variation of the output value of the inertial sensor;
Figure FDA00027810434400000314
respectively are the variation between epochs of the output values after Kalman filtering;
Figure FDA00027810434400000315
and
Figure FDA00027810434400000316
respectively is an output value of the inertial sensor at the moment k;
Figure FDA00027810434400000317
and
Figure FDA00027810434400000318
respectively outputting solutions for the Kalman filtering at the k moment,
s3-3-3, when online learning meets a threshold value for prediction, classifying according to satellite signal occlusion and non-occlusion;
s3-3-3-1, when the GNSS satellite signal is blocked, the INS can work normally due to the attribute of the INS; GNSS data at the moment before the invalidation is an initial value of filtering, which is equivalent to an accurate value, so that divergence of INS precision caused by long navigation time can be avoided; when the GNSS signal is lost, the trained neural network model is used for predicting the output error of the inertial navigation, and the prediction error is used for compensating and correcting the inertial navigation, and firstly, the error is obtained by pre-integrating according to an inertial sensor
Figure FDA0002781043440000041
As Elman neural network for prediction, the prediction data of the neural network is output
Figure FDA0002781043440000042
And
Figure FDA0002781043440000043
the EKF update is performed as an observation constraint,
neural network predicted variance
Figure FDA0002781043440000044
And
Figure FDA0002781043440000045
the obtained predicted values of the position, the rotation quaternion and the speed at the time k +1 are:
Figure FDA0002781043440000046
the constraint is updated as:
Figure FDA0002781043440000047
wherein,
Figure FDA0002781043440000048
and
Figure FDA0002781043440000049
are the state vectors output at step S3-1, respectively, from the inertial sensors;
Figure FDA00027810434400000410
and
Figure FDA00027810434400000411
the predicted values of the position, the rotation quaternion and the speed at the moment k +1 obtained by the neural network respectively,
s3-3-3-2 when GNSS signals are normal, two strategies are available: 1.directly updating the position, the speed and the rotation quaternion at the moment when the inertial sensor outputs k +1 as a state vector, and then performing constraint updating by using the position and the speed output by the GNSS (global navigation satellite system), as shown in step S3-3-2; 2. inputting the position, the speed and the variation of the rotation quaternion at the moment when the inertial sensor outputs k +1 into an Elman neural network, and updating the state vector of the EKF by taking the position, the rotation quaternion and the variation of the speed output by the Elman neural network as a pseudo observation value, as shown in a formula (16); constraint updates are then made based on the GNSS output position and velocity, as shown in equation (17); finally obtained by inertial sensors
Figure FDA00027810434400000412
And
Figure FDA00027810434400000413
and obtained by using EKF
Figure FDA00027810434400000414
Figure FDA00027810434400000415
And
Figure FDA00027810434400000416
adding the training agent into an online learning system of a neural network to carry out real-time learning training,
the EKF status is updated as:
Figure FDA00027810434400000417
the GNSS measurement status in EKF is updated as follows:
Figure FDA0002781043440000051
Figure FDA0002781043440000052
and
Figure FDA0002781043440000053
respectively predicting the variation according to the output neural network
Figure FDA0002781043440000054
And
Figure FDA0002781043440000055
the predicted values of the position, the rotation quaternion and the speed at the k +1 moment are obtained;ePk+1,GandeVk+1,Goutput values for GNSS position and velocity, respectively.
CN202011281769.7A 2020-11-16 2020-11-16 GNSS/INS integrated navigation method Active CN112505737B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011281769.7A CN112505737B (en) 2020-11-16 2020-11-16 GNSS/INS integrated navigation method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011281769.7A CN112505737B (en) 2020-11-16 2020-11-16 GNSS/INS integrated navigation method

Publications (2)

Publication Number Publication Date
CN112505737A true CN112505737A (en) 2021-03-16
CN112505737B CN112505737B (en) 2024-03-01

Family

ID=74956392

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011281769.7A Active CN112505737B (en) 2020-11-16 2020-11-16 GNSS/INS integrated navigation method

Country Status (1)

Country Link
CN (1) CN112505737B (en)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113252031A (en) * 2021-04-28 2021-08-13 燕山大学 NARX neural network assisted integrated navigation method
CN113340298A (en) * 2021-05-24 2021-09-03 南京航空航天大学 Inertial navigation and dual-antenna GNSS external reference calibration method
CN113687396A (en) * 2021-09-26 2021-11-23 重庆赛迪奇智人工智能科技有限公司 Positioning processing method and device, positioning equipment, vehicle and storage medium
CN114216459A (en) * 2021-12-08 2022-03-22 昆山九毫米电子科技有限公司 ELM-assisted GNSS/INS integrated navigation unmanned target vehicle positioning method
CN115326063A (en) * 2022-07-07 2022-11-11 江苏大块头智驾科技有限公司 Inertial device signal filtering method based on deep learning
CN116224407A (en) * 2023-05-06 2023-06-06 山东科技大学 GNSS and INS integrated navigation positioning method and system
CN116381753A (en) * 2023-06-01 2023-07-04 北京航空航天大学 Neural network assisted navigation method of GNSS/INS integrated navigation system during GNSS interruption
CN116499469A (en) * 2023-06-28 2023-07-28 北京航空航天大学 GNSS/INS combined navigation method utilizing neural network model on-line learning and compensation
WO2024012006A1 (en) * 2022-07-15 2024-01-18 腾讯科技（深圳）有限公司 Positioning precision estimation method and apparatus, electronic device and storage medium

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106980133A (en) * 2017-01-18 2017-07-25 中国南方电网有限责任公司超高压输电公司广州局 The GPS INS Combinated navigation methods and system for being compensated and being corrected using neural network algorithm
CN107390246A (en) * 2017-07-06 2017-11-24 电子科技大学 A kind of GPS/INS Combinated navigation methods based on genetic neural network
CN109000640A (en) * 2018-05-25 2018-12-14 东南大学 Vehicle GNSS/INS Combinated navigation method based on discrete Grey Neural Network Model
CN110487271A (en) * 2019-09-26 2019-11-22 哈尔滨工程大学 Elman neural network aiding tight integration air navigation aid when a kind of GNSS signal is obstructed
US10809388B1 (en) * 2019-05-01 2020-10-20 Swift Navigation, Inc. Systems and methods for high-integrity satellite positioning

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106980133A (en) * 2017-01-18 2017-07-25 中国南方电网有限责任公司超高压输电公司广州局 The GPS INS Combinated navigation methods and system for being compensated and being corrected using neural network algorithm
CN107390246A (en) * 2017-07-06 2017-11-24 电子科技大学 A kind of GPS/INS Combinated navigation methods based on genetic neural network
CN109000640A (en) * 2018-05-25 2018-12-14 东南大学 Vehicle GNSS/INS Combinated navigation method based on discrete Grey Neural Network Model
US10809388B1 (en) * 2019-05-01 2020-10-20 Swift Navigation, Inc. Systems and methods for high-integrity satellite positioning
CN110487271A (en) * 2019-09-26 2019-11-22 哈尔滨工程大学 Elman neural network aiding tight integration air navigation aid when a kind of GNSS signal is obstructed

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113252031A (en) * 2021-04-28 2021-08-13 燕山大学 NARX neural network assisted integrated navigation method
CN113340298A (en) * 2021-05-24 2021-09-03 南京航空航天大学 Inertial navigation and dual-antenna GNSS external reference calibration method
CN113340298B (en) * 2021-05-24 2024-05-17 南京航空航天大学 Inertial navigation and dual-antenna GNSS external parameter calibration method
CN113687396A (en) * 2021-09-26 2021-11-23 重庆赛迪奇智人工智能科技有限公司 Positioning processing method and device, positioning equipment, vehicle and storage medium
CN114216459A (en) * 2021-12-08 2022-03-22 昆山九毫米电子科技有限公司 ELM-assisted GNSS/INS integrated navigation unmanned target vehicle positioning method
CN114216459B (en) * 2021-12-08 2024-03-15 昆山九毫米电子科技有限公司 ELM-assisted GNSS/INS combined navigation unmanned target vehicle positioning method
CN115326063A (en) * 2022-07-07 2022-11-11 江苏大块头智驾科技有限公司 Inertial device signal filtering method based on deep learning
WO2024012006A1 (en) * 2022-07-15 2024-01-18 腾讯科技（深圳）有限公司 Positioning precision estimation method and apparatus, electronic device and storage medium
CN116224407A (en) * 2023-05-06 2023-06-06 山东科技大学 GNSS and INS integrated navigation positioning method and system
US12019170B1 (en) 2023-05-06 2024-06-25 Shandong University Of Science And Technology GNSS and INS integrated navigation positioning method and system thereof
CN116381753A (en) * 2023-06-01 2023-07-04 北京航空航天大学 Neural network assisted navigation method of GNSS/INS integrated navigation system during GNSS interruption
CN116381753B (en) * 2023-06-01 2023-08-15 北京航空航天大学 Neural network assisted navigation method of GNSS/INS integrated navigation system during GNSS interruption
CN116499469B (en) * 2023-06-28 2023-09-08 北京航空航天大学 GNSS/INS combined navigation method utilizing neural network model on-line learning and compensation
CN116499469A (en) * 2023-06-28 2023-07-28 北京航空航天大学 GNSS/INS combined navigation method utilizing neural network model on-line learning and compensation

Also Published As

Publication number Publication date
CN112505737B (en) 2024-03-01

Similar Documents

Publication Publication Date Title
CN112505737B (en) GNSS/INS integrated navigation method
CN111156994B (en) INS/DR & GNSS loose combination navigation method based on MEMS inertial component
CN107270893B (en) Lever arm and time asynchronous error estimation and compensation method for real estate measurement
CN109991636A (en) Map constructing method and system based on GPS, IMU and binocular vision
CN101846734B (en) Agricultural machinery navigation and position method and system and agricultural machinery industrial personal computer
CN110779521A (en) Multi-source fusion high-precision positioning method and device
CN104729506A (en) Unmanned aerial vehicle autonomous navigation positioning method with assistance of visual information
CN109000640B (en) Vehicle GNSS/INS integrated navigation method based on discrete grey neural network model
CN106980133A (en) The GPS INS Combinated navigation methods and system for being compensated and being corrected using neural network algorithm
CN113029139B (en) Airport flight area vehicle differential Beidou/SINS combined navigation method based on motion detection
CN109059909A (en) Satellite based on neural network aiding/inertial navigation train locating method and system
CN112146655B (en) Elastic model design method for BeiDou/SINS tight integrated navigation system
CN113203418A (en) GNSSINS visual fusion positioning method and system based on sequential Kalman filtering
CN110954102A (en) Magnetometer-assisted inertial navigation system and method for robot positioning
CN114545472B (en) Navigation method and device of GNSS/INS combined system
CN113503872A (en) Low-speed unmanned vehicle positioning method based on integration of camera and consumption-level IMU
CN114526731A (en) Inertia combination navigation direction positioning method based on moped
CN113074757A (en) Calibration method for vehicle-mounted inertial navigation installation error angle
CN115060257A (en) Vehicle lane change detection method based on civil-grade inertia measurement unit
CN114136310B (en) Autonomous error suppression system and method for inertial navigation system
CN113008229B (en) Distributed autonomous integrated navigation method based on low-cost vehicle-mounted sensor
CN112461236B (en) Vehicle-mounted high-precision fault-tolerant integrated navigation method and system
CN110375740B (en) Vehicle navigation method, device, equipment and storage medium
CN117029878A (en) Train navigation rapid initial alignment system based on double-antenna assistance
CN116576849A (en) Vehicle fusion positioning method and system based on GMM assistance

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