CN111290007A - BDS/SINS combined navigation method and system based on neural network assistance - Google Patents

BDS/SINS combined navigation method and system based on neural network assistance Download PDF

Info

Publication number
CN111290007A
CN111290007A CN202010125926.9A CN202010125926A CN111290007A CN 111290007 A CN111290007 A CN 111290007A CN 202010125926 A CN202010125926 A CN 202010125926A CN 111290007 A CN111290007 A CN 111290007A
Authority
CN
China
Prior art keywords
error
sins
bds
model
value
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.)
Pending
Application number
CN202010125926.9A
Other languages
Chinese (zh)
Inventor
蔡成林
李帅
彭滨
贾伟
潘海涛
刘元成
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.)
Guilin University of Electronic Technology
Original Assignee
Guilin University of Electronic Technology
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 Guilin University of Electronic Technology filed Critical Guilin University of Electronic Technology
Priority to CN202010125926.9A priority Critical patent/CN111290007A/en
Publication of CN111290007A publication Critical patent/CN111290007A/en
Pending legal-status Critical Current

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
    • G06N3/045—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

Landscapes

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

Abstract

The invention provides a BDS/SINS combined navigation method and system based on neural network assistance, comprising the following steps: establishing an SINS navigation system error model, and obtaining error data by using the SINS navigation system error model; establishing a BDS/SINS combined navigation model, taking error data as an initial value of the BDS/SINS combined navigation model, obtaining an original value and an estimated value of a target receiver by using the BDS/SINS combined navigation model, and taking the difference value of the original value and the estimated value as a measurement vector; establishing an unscented Kalman filter, inputting the measurement vector into the unscented Kalman filter, and performing filtering processing on the BDS/SINS integrated navigation model by using the unscented Kalman filter; carrying out optimization training on the BDS/SINS combined navigation model after filtering by using a neural network to obtain an optimized BDS/SINS combined navigation model; and carrying out tracking navigation. The locking-losing phenomenon of the BDS signals can be effectively inhibited, and the positioning of the vehicles in the tunnel, the viaduct, the mountain area and the dense environment of the high-rise building is facilitated.

Description

BDS/SINS combined navigation method and system based on neural network assistance
Technical Field
The invention mainly relates to the technical field of navigation, in particular to a BDS/SINS combined navigation method and system based on neural network assistance.
Background
Currently, the main disadvantage of inertial navigation is that positioning errors accumulate over time, so that it is difficult to work independently for a long time, and there are two approaches to solve this problem: the precision of the inertial navigation system is improved, and the other method adopts a combined navigation technology. The accuracy of the INS is improved by means of new materials and new processes, which requires a lot of manpower and material resources, and the improvement of the accuracy of the inertial sensor is limited.
The Beidou Satellite Navigation System (BeiDou Navigation Satellite System, BDS) is a national strategy in China, and with 12 months and 16 days in this year (2019), the fifth twelve and fifty-three Beidou Navigation satellites are successfully launched by using a Changcheng third carrier rocket in the West Chang Satellite launching center in a one-rocket two-star mode in China. The successful launching of the two satellites is that the China Beidou satellite navigation plan formally enters a milestone-all circular orbit satellites in the earth are launched completely, which means that the Beidou global system core constellation is deployed and completed, and a solid foundation is laid for the Beidou No. three system to be built completely in 2020. The problems of signal lock losing and low accuracy exist in some current navigation systems.
Disclosure of Invention
The invention aims to solve the technical problem of the prior art and provides a BDS/SINS combined navigation method and system based on neural network assistance.
The technical scheme for solving the technical problems is as follows: a BDS/SINS combined navigation method based on neural network assistance comprises the following steps:
establishing an SINS navigation system error model, and obtaining error data by using the SINS navigation system error model;
combining a BDS navigation model and an SINS navigation system error model in a loose combination mode to obtain a BDS/SINS combined navigation model, taking the error data as an initial value of the BDS/SINS combined navigation model, obtaining an original value and an estimated value of a target receiver by using the BDS/SINS combined navigation model, calculating a difference value of the original value and the estimated value, and taking the difference value as a measurement vector;
establishing an unscented Kalman filter, inputting the measurement vector into the unscented Kalman filter, and performing filtering processing on the BDS/SINS combined navigation model by using the unscented Kalman filter;
carrying out optimization training on the BDS/SINS combined navigation model after filtering by using a Back Propagation neural network to obtain an optimized BDS/SINS combined navigation model;
and tracking and navigating the target receiver by using the optimized BDS/SINS combined navigation model.
Another technical solution of the present invention for solving the above technical problems is as follows: a BDS/SINS combined navigation system based on neural network assistance comprises:
the error value calculation module is used for establishing an SINS navigation system error model and obtaining error data by using the SINS navigation system error model;
the integrated navigation model establishing module is used for combining a BDS navigation model and an SINS navigation system error model in a loose combination mode to obtain a BDS/SINS integrated navigation model of the BDS/SINS integrated navigation model, using the error data as an initial value of the BDS/SINS integrated navigation model, obtaining an original value and an estimated value of a target receiver by using the BDS/SINS integrated navigation model, calculating a difference value between the original value and the estimated value, and using the difference value as a measurement vector;
the filtering processing module is used for establishing an unscented Kalman filter, inputting the measurement vector into the unscented Kalman filter, and performing filtering processing on the BDS/SINS integrated navigation model by using the unscented Kalman filter;
the optimization module is used for carrying out optimization training on the BDS/SINS combined navigation model after filtering processing by using a Back Propagation neural network to obtain an optimized BDS/SINS combined navigation model;
and the tracking module is used for tracking and navigating the target receiver by using the optimized BDS/SINS combined navigation model.
Another technical solution of the present invention for solving the above technical problems is as follows: a BDS/SINS combined navigation system based on neural network assistance, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the computer program, when executed by the processor, implementing a navigation method as described above.
Another technical solution of the present invention for solving the above technical problems is as follows: a computer readable storage medium storing a computer program which, when executed by a processor, implements a neural network assistance-based BDS/SINS combined navigation method as described above.
The invention has the beneficial effects that: the error models of the BDS navigation system and the SINS navigation system are effectively combined, the error value is used as an initial value of the combined navigation model, an original value and an estimated value of a target receiver are obtained, difference calculation is carried out on the original value and the estimated value, filtering processing is carried out on the difference, optimization processing is carried out, the phenomenon that the BDS signals are often unlocked can be effectively inhibited, and positioning of vehicles in the dense environment of tunnels, viaducts, mountainous areas and high buildings is facilitated.
Drawings
Fig. 1 is a schematic flow chart of a navigation method according to an embodiment of the present invention;
FIG. 2 is a block diagram of a navigation system according to an embodiment of the present invention;
fig. 3 is a schematic diagram of an overall scheme of a BDS/SINS pine combination according to an embodiment of the present invention;
fig. 4 is a schematic diagram of solving an error model of the SINS navigation system according to an embodiment of the present invention;
fig. 5 is a schematic diagram of Back Propagation neural network training according to an embodiment of the present invention.
Detailed Description
The principles and features of this invention are described below in conjunction with the following drawings, which are set forth by way of illustration only and are not intended to limit the scope of the invention.
Fig. 1 is a schematic flow chart of a navigation method according to an embodiment of the present invention;
as shown in fig. 1, a BDS/SINS combined navigation method based on neural network assistance includes:
establishing an SINS navigation system error model, and obtaining an error value by using the SINS navigation system error model;
combining a BDS navigation model and an SINS navigation system error model in a loose combination mode to obtain a BDS/SINS combined navigation model, taking the error data as an initial value of the BDS/SINS combined navigation model, obtaining an original value and an estimated value of a target receiver by using the BDS/SINS combined navigation model, calculating a difference value of the original value and the estimated value, and taking the difference value as a measurement vector;
establishing an unscented Kalman filter, inputting the measurement vector into the unscented Kalman filter, and performing filtering processing on the BDS/SINS combined navigation model by using the unscented Kalman filter;
carrying out optimization training on the BDS/SINS combined navigation model after filtering by using a Back Propagation neural network to obtain an optimized BDS/SINS combined navigation model;
and tracking and navigating the target receiver by using the optimized BDS/SINS combined navigation model.
It should be understood that the BDS/SINS combined navigation model is represented as a combination of a BDS navigation model and a SINS navigation system error model.
In the embodiment, the error models of the BDS navigation system and the SINS navigation system are effectively combined, the error value is used as the initial value of the combined navigation model, the original value and the estimated value of the target receiver are obtained, the difference value between the original value and the estimated value is calculated, filtering processing is performed through the difference value, optimization processing is performed, the phenomenon that the BDS signals are often unlocked can be effectively inhibited, and positioning of vehicles in the dense environment of tunnels, viaducts, mountainous areas and high buildings is facilitated.
Alternatively, as an embodiment of the present invention, assume that the ideal SINS attitude matrix from the navigation coordinate system (n system) to the carrier coordinate system (b system) is
Figure BDA0002394370860000041
And the attitude matrix solved in the navigation computer is
Figure BDA0002394370860000051
There is a deviation between the two. For transformation matrix
Figure BDA0002394370860000052
And
Figure BDA0002394370860000053
it is believed that their b-systems are coincident and will be
Figure BDA0002394370860000054
The corresponding navigation coordinate system is called as the calculation navigation coordinate system, which is abbreviated as n' system, so the calculation posture is also often arrayed as
Figure BDA0002394370860000055
Therefore, the temperature of the molten metal is controlled,
Figure BDA0002394370860000056
and
Figure BDA0002394370860000057
the deviation therebetween is a deviation between n' and n. Taking n system as a reference coordinate system, recording the equivalent rotation vector (misalignment angle error) from n system to n' as phi, and recording the gyro measurement error as phi
Figure BDA0002394370860000058
The calculation error of navigation system is recorded as
Figure BDA0002394370860000059
The process of obtaining the error value by using the SINS navigation system error model comprises the following steps:
obtaining a misalignment angle error value of an SINS navigation system error model through a first formula, wherein the first formula is as follows:
Figure BDA00023943708600000510
wherein n is a navigation reference coordinate system, b is a carrier coordinate system, i is an inertia coordinate system,
Figure BDA00023943708600000511
is the rotation of n relative to i,
Figure BDA00023943708600000512
is the angular velocity of b relative to i,
Figure BDA00023943708600000513
for the misalignment angle error value to be the value,
Figure BDA00023943708600000528
in the form of a differential of the misalignment angle error, let delta be the error sign,
Figure BDA00023943708600000514
for n is the error of the calculation,
Figure BDA00023943708600000515
the value of the error is measured by the gyroscope,
Figure BDA00023943708600000516
is an attitude matrix;
obtaining a speed error of an SINS navigation system error model through a second formula, wherein the second formula is as follows:
Figure BDA00023943708600000517
wherein v isnIs a velocity value, δ is an error sign, δ vnIn order to be able to determine the speed error,
Figure BDA00023943708600000518
in order to differentiate the speed error,
Figure BDA00023943708600000519
in the form of a matrix of poses,
Figure BDA00023943708600000520
in the form of an accelerometer measurement,
Figure BDA00023943708600000521
in order for the accelerometer to measure the error,
Figure BDA00023943708600000522
is the angular velocity of the earth's rotation,
Figure BDA00023943708600000523
n is the angular velocity of rotation,
Figure BDA00023943708600000524
an error is calculated for the rotational angular velocity of the earth,
Figure BDA00023943708600000525
calculating an error for the navigation system rotation;
calculating the longitude error, the latitude error and the altitude error of an SINS navigation system error model through a longitude error formula, a latitude error formula and an altitude error formula, wherein the latitude error formula is as follows:
Figure BDA00023943708600000526
wherein,
Figure BDA00023943708600000527
for latitude error, RMIs the radius of the meridian principal curvature, h is the height, δ h is the height error, vNIs the north velocity, δ vNIs the north velocity error;
the longitude error formula is:
Figure BDA0002394370860000061
wherein,
Figure BDA0002394370860000062
is the differential of the longitude, and is,
Figure BDA0002394370860000063
as a derivative of the longitude error, RNhIs the curvature radius of the mortise and unitary ring, L is the latitude value,
Figure BDA0002394370860000064
in order to be the east-direction speed,
Figure BDA0002394370860000065
is east speed error;
the height error formula is:
Figure BDA0002394370860000066
wherein,
Figure BDA00023943708600000611
in order to be a differential of the height,
Figure BDA0002394370860000067
in order to differentiate the high-level error,
Figure BDA0002394370860000068
in order to obtain the speed in the direction of the sky,
Figure BDA0002394370860000069
is the error in the speed in the direction of the day.
An inertial measurement element in an SINS navigation system error model consists of a gyroscope and an accelerometer, the two sensors have corresponding device errors, and the zero offset measured by the accelerometer is recorded as
Figure BDA00023943708600000610
The random constant drift of the gyroscope is recorded as epsilonbAnd preparing for the next step of system equation establishment of Kalman filtering through the establishment of an error model.
In the above embodiment, a kalman filter system equation and a measurement equation of the BDS/SINS integrated navigation model are constructed, the loose integration is adopted in the integration mode, the BDS outputs the original speed and position of the receiver, the SINS outputs the speed and position calculated by the inertial navigation algorithm, and the difference between the two is used as the measurement input of the kalman filter. And the SINS navigation system error model error is used as the initial state input of the system.
Optionally, as an embodiment of the present invention, as shown in fig. 3 to 4, a BDS receiver model in the BDS navigation model includes an antenna, a radio frequency front end, and the radio frequency front end performs signal capture and signal tracking and outputs data.
The SINS navigation system error model comprises a three-axis gyroscope and a three-axis accelerometer, SINS resolving is carried out, data fusion is carried out on resolved data and data in the BDS navigation model, and the SINS navigation system error model outputs a resolved value. And outputting a fusion value by the BDS/SINS combined navigation model.
Specifically, the process of obtaining the original value and the estimated value of the target receiver by using the BDS/SINS combined navigation model includes:
calculating the system state of the BDS/SINS combined navigation model according to a state equation, wherein the state equation is as follows:
Figure BDA0002394370860000071
wherein F (t) · X (t) is original value, G (t) · W (t) is calculated value, X (t) is state quantity equation of combined navigation model, F (t) is state transition matrix, G (t) is noise distribution matrix, W (t) is system noise vector,
Figure BDA0002394370860000072
Fwa state transition matrix for the error values of the SINS navigation system error model,
Figure BDA0002394370860000073
Figure BDA0002394370860000074
in the form of a matrix of poses,
Figure BDA0002394370860000075
ωgx,ωgy,ωgzis a corresponding random drift of the X, Y, Z axis, ωax,ωay,ωazAcceleration random error corresponding to axis X, Y, Z;
the process of obtaining the state quantity is as follows:
selecting a northeast coordinate system as a navigation coordinate system, and establishing a state quantity equation according to 15-dimensional state parameters, wherein the state quantity equation is as follows:
Figure BDA0002394370860000076
wherein phi isE,φN,φUFor the three-axis misalignment angle error in the northeast direction,
Figure BDA0002394370860000077
in order to combine the velocity error values of the navigation model,
Figure BDA0002394370860000078
respectively, a longitude error value, a latitude error value and an altitude error value of the combined navigation model, epsilonbx,εby,εbzGyro random constant drift values of three-axis coordinates of x, y and z respectively,
Figure BDA0002394370860000079
the accelerometers, which are x, y and z triaxial coordinates respectively, are randomly constant with zero drift.
Optionally, as an embodiment of the present invention, the calculating a difference between the original value and the estimated value, and the process of using the difference as a measurement vector includes:
calculating a difference between the original value and the extrapolated value using a metrology equation:
Figure BDA0002394370860000081
wherein Z (t) is a measurement vector of the integrated navigation model, H (t) is a measurement matrix of the integrated navigation model, X (t) is a state quantity equation of the integrated navigation model, V (t) is a measurement noise vector of the integrated navigation model,
Figure BDA0002394370860000082
Hpand HvRespectively, a measurement matrix of the position and the velocity,
Figure BDA0002394370860000083
Vpand VvPosition measurement white noise and velocity measurement white noise are separately measured.
Optionally, as an embodiment of the present invention, the process of performing filtering processing on the BDS/SINS combined navigation model by using the unscented kalman filter includes:
setting an initial state of the integrated navigation model according to a third formula, wherein the third formula is as follows:
Figure BDA0002394370860000084
wherein,
Figure BDA0002394370860000085
is an initial state value, E (x)0) T is a transposed matrix, P, to expectation0Is the variance;
sampling the BDS/SINS combined navigation model through a Sigma qualitative sampling point and a sampling formula, wherein the sampling formula is as follows:
Figure BDA0002394370860000086
wherein, ξi,k-1(i ═ 0, 1.. 2w) is a Sigma qualitative sample point setW is the dimension of the random vector x;
calculating a first-order weight value and a second-order weight value of a Sigma qualitative sampling point by using a first-order weight value calculation formula and a second-order weight value calculation formula, wherein the first-order weight value calculation formula is as follows:
Figure BDA0002394370860000087
wherein,
Figure BDA0002394370860000088
as a first order weight, λ is α2(w + K) -w, w is the dimension of the random vector x, and K is a proportional parameter;
the second-order weight calculation formula is as follows:
Figure BDA0002394370860000091
wherein,
Figure BDA0002394370860000092
as a second-order weight, λ is α2(w + K) -w, α is a positive scaling factor, α is used to adjust the Sigma Point and the Sigma Point after sampling
Figure BDA0002394370860000093
α is in the range of [0,1 ]]W is the dimension of the random vector x and β is a non-negative coefficient.
Optionally, as an embodiment of the present invention, the process of performing filtering processing on the BDS/SINS combined navigation model by using the unscented kalman filter further includes an updating step:
updating the Sigma qualitative sampling points according to a first updating formula, wherein the first updating formula is as follows:
γi，k/k-1＝f(ξi,k-1,uk-1)+qk-1，
wherein, γi，k/k-1For the updated set of points, f (ξ)i,k-1,uk-1) For non-linear transformation, qk-1Is the mean vector of the system process noise; updating the prediction mean value according to the updated Sigma qualitative sampling points and a second updating formula, wherein the second updating formula is as follows:
Figure BDA0002394370860000094
wherein,
Figure BDA0002394370860000095
in order to predict the mean value of the mean,
Figure BDA0002394370860000096
is a first order weight, gammai，k/k-1K and k-1 are respectively time for the updated point set;
updating the covariance according to the updated Sigma qualitative sampling points and a third updating formula, wherein the third updating formula is as follows:
Figure BDA0002394370860000097
wherein, Pk/k-1In order to be the covariance,
Figure BDA0002394370860000098
is a second order weight, gammai，k/k-1For updated point sets, Qk-1Is a non-negative-definite variance matrix;
updating the measurement equation according to a fourth updating formula, wherein the fourth updating formula is as follows:
Figure BDA0002394370860000099
wherein, Pz,kIn order to measure the data, the measurement data,
Figure BDA0002394370860000101
the second-order weight is a second-order weight,
Figure BDA0002394370860000102
Figure BDA0002394370860000103
is a first order weight, xi,k/k-1K and k-1 are respectively time, R for the updated prediction mean valuekIs a positive covariance matrix of the measured noise.
In the above embodiment, unscented kalman filtering is used as the kalman filtering, because in actual target tracking, the state model and the measurement model of the tracking system are mostly nonlinear, unscented kalman filtering is used to improve the data accuracy.
Optionally, as an embodiment of the present invention, as shown in fig. 5, the process of performing optimization training on the filtered BDS/SINS combined navigation model by using Back Propagation neural network includes:
carrying out iterative processing on the Back Propagation neural network according to a conjugate gradient method;
correcting the weight of the Back Propagation neural network after the iterative processing;
and predicting and correcting the Back Propagation neural network by using the corrected Back Propagation neural network.
When the BDS signal is locked, the BDS and the SINS use unscented Kalman filtering to carry out combined navigation, and carry out online training based on the BP neural network after conjugate gradient optimization, when the BDS signal is unlocked, a trained model in the signal locking is called to carry out online prediction, and the problem that the SINS precision is rapidly reduced under the condition that the BDS signal is unlocked is solved.
Specifically, the bp (back propagation) neural network, i.e., the learning process of the error back propagation algorithm, is composed of two processes of forward propagation of information and back propagation of errors. The traditional BP neural network algorithm adopts the steepest descent direction, namely the negative gradient direction, but only reflects the local property of an error function at a certain point, not necessarily the global steepest descent direction. The conjugate gradient method is an algorithm for improving the search direction, and the conjugate direction is obtained by adding a correction term to the original negative gradient direction of the BP standard algorithm, namely the steepest descent method has conjugation. In short, the search direction of the conjugate gradient method is a direction of conjugation. The algorithm mainly utilizesModifying the weight w in the direction of the conjugate gradientkLet wkThe determination is faster, the training speed of the BP neural network is accelerated, and the network is prevented from falling into local minimum.
The Back Propagation neural network has three layers, namely an input layer, a hidden layer and an output layer,
the target function of the Back Propagation neural network is set as follows:
Figure BDA0002394370860000111
wherein, the hidden layer is a nonlinear layer, and a sigmoid function is adopted:
Figure BDA0002394370860000112
the output of the hidden node is:
Figure BDA0002394370860000113
the output of the output node is:
Figure BDA0002394370860000114
in the formula, xi、yiZl are input, hidden and output nodes, wjiIs xiAnd yiInter network weight, vijIs yjAnd ziThe network weight of (2). ThetajAnd thetalRespectively xiAnd yiM and yjAnd ziTo a threshold value in between.
Specifically, the Back Propagation neural network is subjected to prediction correction by using the corrected Back Propagation neural network
In the above embodiment, the neural network adopts a BP neural network algorithm improved based on a conjugate gradient method, which is an algorithm for improving a search direction, and multiplies the gradient of a previous point by an appropriate coefficient, and adds the result to the gradient of a changed point to obtain a new search direction. The training speed of the BP neural network is accelerated, and the network is prevented from falling into local minimum;
the Back Propagation neural network is subjected to iterative processing according to a conjugate gradient method, and an iterative equation is as follows:
Pk+1＝-gk+1+βkPk，
in the formula:
gkis E to wkGradient of (2)
Figure BDA0002394370860000115
PkIn order to search for the direction(s),
βkis a conjugation factor having a value of
Figure BDA0002394370860000116
Correcting the weight of the BP neural network through the iterative equation:
Δ w (k +1) ═ w (k) + λ (k) p (k), where λ (k) is the optimal step size.
FIG. 2 is a block diagram of a navigation system according to an embodiment of the present invention;
optionally, as another embodiment of the present invention, as shown in fig. 2, a BDS/SINS combined navigation system based on neural network assistance includes:
the error value calculation module is used for establishing an SINS navigation system error model and obtaining error data by using the SINS navigation system error model;
the integrated navigation model establishing module is used for combining a BDS navigation model and an SINS navigation system error model in a loose combination mode to obtain a BDS/SINS integrated navigation model, using the error data as an initial value of the BDS/SINS integrated navigation model, respectively obtaining an original value and an estimated value of a target receiver by using the BDS/SINS integrated navigation model, calculating a difference value between the original value and the estimated value, and using the difference value as a measurement vector;
the filtering processing module is used for establishing an unscented Kalman filter, inputting the measurement vector into the unscented Kalman filter, and performing filtering processing on the BDS/SINS integrated navigation model by using the unscented Kalman filter;
the optimization module is used for carrying out optimization training on the BDS/SINS combined navigation model after filtering processing by using a Back Propagation neural network to obtain an optimized BDS/SINS combined navigation model;
and the tracking module is used for tracking and navigating the target receiver by using the optimized BDS/SINS combined navigation model.
Optionally, as another embodiment of the present invention, a BDS/SINS combined navigation system based on neural network assistance includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the computer program, the BDS/SINS combined navigation method based on neural network assistance as described above is implemented.
Alternatively, as another embodiment of the present invention, a computer-readable storage medium stores a computer program which, when executed by a processor, implements the BDS/SINS combined navigation method based on neural network assistance as described above.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the several embodiments provided in the present application, it should be understood that the disclosed system and method may be implemented in other ways. For example, the above-described system embodiments are merely illustrative, and for example, a division of a unit is merely a logical division, and an actual implementation may have another division, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed.
Units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment of the present invention.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. It will be understood that the technical solution of the present invention essentially contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
While the invention has been described with reference to specific embodiments, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims (9)

1. A BDS/SINS combined navigation method based on neural network assistance is characterized by comprising the following steps:
establishing an SINS navigation system error model, and obtaining error data by using the SINS navigation system error model;
combining a BDS navigation model and an SINS navigation system error model in a loose combination mode to obtain a BDS/SINS combined navigation model, taking the error data as an initial value of the BDS/SINS combined navigation model, obtaining an original value and an estimated value of a target receiver by using the BDS/SINS combined navigation model, calculating a difference value of the original value and the estimated value, and taking the difference value as a measurement vector;
establishing an unscented Kalman filter, inputting the measurement vector into the unscented Kalman filter, and performing filtering processing on the BDS/SINS combined navigation model by using the unscented Kalman filter;
carrying out optimization training on the BDS/SINS combined navigation model after filtering by using a Back Propagation neural network to obtain an optimized BDS/SINS combined navigation model;
and tracking and navigating the target receiver by using the optimized BDS/SINS combined navigation model.
2. The method of claim 1, wherein the obtaining error data using the SINS error model comprises:
obtaining a misalignment angle error value of an SINS navigation system error model through a first formula, wherein the first formula is as follows:
Figure FDA0002394370850000011
wherein n is a navigation reference coordinate system, b is a carrier coordinate system, i is an inertia coordinate system,
Figure FDA0002394370850000012
is the rotation of n relative to i,
Figure FDA0002394370850000013
is the angular velocity of b relative to i,
Figure FDA0002394370850000014
for the misalignment angle error value to be the value,
Figure FDA0002394370850000015
in the form of a differential of the misalignment angle error, let delta be the error sign,
Figure FDA0002394370850000016
for n is the error of the calculation,
Figure FDA0002394370850000017
the value of the error is measured by the gyroscope,
Figure FDA0002394370850000018
is an attitude matrix;
obtaining a speed error of an SINS navigation system error model through a second formula, wherein the second formula is as follows:
Figure FDA0002394370850000021
wherein v isnIs a velocity value, δ is an error sign, δ vnIn order to be able to determine the speed error,
Figure FDA0002394370850000022
in order to differentiate the speed error,
Figure FDA0002394370850000023
in the form of a matrix of poses,
Figure FDA0002394370850000024
in the form of an accelerometer measurement,
Figure FDA0002394370850000025
in order for the accelerometer to measure the error,
Figure FDA0002394370850000026
is the angular velocity of the earth's rotation,
Figure FDA0002394370850000027
n is the angular velocity of rotation,
Figure FDA0002394370850000028
an error is calculated for the rotational angular velocity of the earth,
Figure FDA0002394370850000029
calculating an error for the navigation system rotation;
calculating the longitude error, the latitude error and the altitude error of an SINS navigation system error model through a longitude error formula, a latitude error formula and an altitude error formula, wherein the latitude error formula is as follows:
Figure FDA00023943708500000210
wherein,
Figure FDA00023943708500000211
for latitude error, RMIs the radius of the meridian principal curvature, h is the height, δ h is the height error, vNIs the north velocity, δ vNIs the north velocity error;
the longitude error formula is:
Figure FDA00023943708500000212
wherein,
Figure FDA00023943708500000213
is the differential of the longitude, and is,
Figure FDA00023943708500000214
as a derivative of the longitude error, RNhIs the curvature radius of the mortise and unitary ring, L is the latitude value,
Figure FDA00023943708500000215
is east ofThe speed of the moving-direction is controlled,
Figure FDA00023943708500000216
is east speed error;
the height error formula is:
Figure FDA00023943708500000217
wherein,
Figure FDA00023943708500000218
in order to be a differential of the height,
Figure FDA00023943708500000219
in order to differentiate the high-level error,
Figure FDA00023943708500000220
in order to obtain the speed in the direction of the sky,
Figure FDA00023943708500000221
is the error in the speed in the direction of the day.
3. The navigation method of claim 1, wherein the step of obtaining the original value and the estimated value of the target receiver by using the BDS/SINS combined navigation model comprises:
calculating the system state of the BDS/SINS combined navigation model according to a state equation, wherein the state equation is as follows:
Figure FDA00023943708500000222
wherein F (t) · X (t) is original value, G (t) · W (t) is calculated value, X (t) is state quantity equation of combined navigation model, F (t) is state transition matrix, G (t) is noise distribution matrix, W (t) is system noise vector,
Figure FDA0002394370850000031
Fwa state transition matrix for the error values of the SINS navigation system error model,
Figure FDA0002394370850000032
Figure FDA0002394370850000033
in the form of a matrix of poses,
Figure FDA0002394370850000034
ωgx,ωgy,ωgzis a corresponding random drift of the X, Y, Z axis, ωax,ωay,ωazAcceleration random error corresponding to axis X, Y, Z;
the process of obtaining the state quantity is as follows:
selecting a northeast coordinate system as a navigation coordinate system, and establishing a state quantity equation according to 15-dimensional state parameters, wherein the state quantity equation is as follows:
Figure FDA0002394370850000035
wherein phi isE,φN,φUFor the three-axis misalignment angle error in the northeast direction,
Figure FDA0002394370850000036
in order to combine the velocity error values of the navigation model,
Figure FDA0002394370850000037
respectively, a longitude error value, a latitude error value and an altitude error value of the combined navigation model, epsilonbx,εby,εbzGyro random constant drift values for the three-axis x, y and z coordinates, ▽x,▽y,▽zThe accelerometers, which are x, y and z triaxial coordinates respectively, are randomly constant with zero drift.
4. The navigation method according to claim 3, wherein the step of calculating a difference between the original value and the estimated value as a measurement vector comprises:
calculating a difference between the original value and the extrapolated value using a metrology equation:
Z(t)＝H(t)·X(t)+V(t)，
wherein Z (t) is a measurement vector of the integrated navigation model, H (t) is a measurement matrix of the integrated navigation model, X (t) is a state quantity equation of the integrated navigation model, V (t) is a measurement noise vector of the integrated navigation model,
Figure FDA0002394370850000038
Hpand HvRespectively, a measurement matrix of the position and the velocity,
Figure FDA0002394370850000039
Vpand VvPosition measurement white noise and velocity measurement white noise are separately measured.
5. The navigation method according to claim 4, wherein the filtering the BDS/SINS combined navigation model by using the unscented Kalman filter comprises:
setting an initial state of the integrated navigation model according to a third formula, wherein the third formula is as follows:
Figure FDA0002394370850000041
wherein,
Figure FDA0002394370850000042
is an initial state value, E (x)0) T is a transposed matrix, P, to expectation0Is the variance;
sampling the BDS/SINS combined navigation model through a Sigma qualitative sampling point and a sampling formula, wherein the sampling formula is as follows:
Figure FDA0002394370850000043
wherein, ξi,k-1(i ═ 0,1,. 2w) is a set of Sigma qualitative sample points, w is the dimension of the random vector x;
calculating a first-order weight value and a second-order weight value of a Sigma qualitative sampling point by using a first-order weight value calculation formula and a second-order weight value calculation formula, wherein the first-order weight value calculation formula is as follows:
Figure FDA0002394370850000044
wherein, Wi mAs a first order weight, λ is α2(w + K) -w, w is the dimension of the random vector x, and K is a proportional parameter;
the second-order weight calculation formula is as follows:
Figure FDA0002394370850000045
wherein, Wi cAs a second-order weight, λ is α2(w + K) -w, α is a positive scaling factor, α is used to adjust the Sigma Point and the Sigma Point after sampling
Figure FDA0002394370850000046
α is in the range of [0,1 ]]W is the dimension of the random vector x and β is a non-negative coefficient.
6. The navigation method according to claim 5, wherein the process of filtering the BDS/SINS combined navigation model by using the unscented Kalman filter further comprises the step of updating:
updating the Sigma qualitative sampling points according to a first updating formula, wherein the first updating formula is as follows:
γi，k/k-1＝f(ξi,k-1,uk-1)+qk-1，
wherein, γi，k/k-1For the updated set of points, f (ξ)i,k-1,uk-1) Is notLinear transformation, qk-1Is the mean vector of the system process noise; updating the prediction mean value according to the updated Sigma qualitative sampling points and a second updating formula, wherein the second updating formula is as follows:
Figure FDA0002394370850000051
wherein,
Figure FDA0002394370850000052
to predict the mean, Wi mIs a first order weight, gammai，k/k-1K and k-1 are respectively time for the updated point set;
updating the covariance according to the updated Sigma qualitative sampling points and a third updating formula, wherein the third updating formula is as follows:
Figure FDA0002394370850000053
wherein, Pk/k-1Is covariance, Wi cIs a second order weight, gammai，k/k-1For updated point sets, Qk-1Is a non-negative-definite variance matrix;
updating the measurement equation according to a fourth updating formula, wherein the fourth updating formula is as follows:
Figure FDA0002394370850000054
wherein, Pz,kFor measuring data, Wi cThe second-order weight is a second-order weight,
Figure FDA0002394370850000055
Wi mis a first order weight, xi,k/k-1K and k-1 are respectively time, R for the updated prediction mean valuekIs a positive covariance matrix of the measured noise.
7. The navigation method according to claim 5, wherein the process of optimally training the filtered BDS/SINS combined navigation model by using Back Propagation neural network comprises:
carrying out iterative processing on the Back Propagation neural network according to a conjugate gradient method;
correcting the weight of the Back Propagation neural network after the iterative processing;
and predicting and correcting the Back Propagation neural network by using the corrected Back Propagation neural network.
8. A BDS/SINS combined navigation system based on neural network assistance is characterized by comprising:
the error value calculation module is used for establishing an SINS navigation system error model and obtaining error data by using the SINS navigation system error model;
the integrated navigation model establishing module is used for combining a BDS navigation model and an SINS navigation system error model in a loose combination mode to obtain a BDS/SINS integrated navigation model, using the error data as an initial value of the BDS/SINS integrated navigation model, obtaining an original value and an estimated value of a target receiver by using the BDS/SINS integrated navigation model, calculating a difference value between the original value and the estimated value, and using the difference value as a measurement vector;
the filtering processing module is used for establishing an unscented Kalman filter, inputting the measurement vector into the unscented Kalman filter, and performing filtering processing on the BDS/SINS integrated navigation model by using the unscented Kalman filter;
the optimization module is used for carrying out optimization training on the BDS/SINS combined navigation model after filtering processing by using a Back Propagation neural network to obtain an optimized BDS/SINS combined navigation model;
and the tracking module is used for tracking and navigating the target receiver by using the optimized BDS/SINS combined navigation model.
9. A BDS/SINS combined navigation system based on neural network assistance, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that when the computer program is executed by the processor, the BDS/SINS combined navigation method based on neural network assistance according to any one of claims 1 to 7 is realized.
CN202010125926.9A 2020-02-27 2020-02-27 BDS/SINS combined navigation method and system based on neural network assistance Pending CN111290007A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010125926.9A CN111290007A (en) 2020-02-27 2020-02-27 BDS/SINS combined navigation method and system based on neural network assistance

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010125926.9A CN111290007A (en) 2020-02-27 2020-02-27 BDS/SINS combined navigation method and system based on neural network assistance

Publications (1)

Publication Number Publication Date
CN111290007A true CN111290007A (en) 2020-06-16

Family

ID=71025687

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010125926.9A Pending CN111290007A (en) 2020-02-27 2020-02-27 BDS/SINS combined navigation method and system based on neural network assistance

Country Status (1)

Country Link
CN (1) CN111290007A (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112665581A (en) * 2020-12-04 2021-04-16 山东省计算中心（国家超级计算济南中心） Combined navigation method based on BP neural network assisted Kalman filtering
CN112762932A (en) * 2021-04-07 2021-05-07 智道网联科技(北京)有限公司 Trajectory calculation method and device based on neural network model
CN113465628A (en) * 2021-06-17 2021-10-01 杭州鸿泉物联网技术股份有限公司 Inertial measurement unit data compensation method and system
CN116224407A (en) * 2023-05-06 2023-06-06 山东科技大学 GNSS and INS integrated navigation positioning method and system
CN117310773A (en) * 2023-11-30 2023-12-29 山东省科学院海洋仪器仪表研究所 Autonomous positioning method and system for underwater robot based on binocular stereoscopic vision

Citations (6)

* 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
CN108761512A (en) * 2018-07-28 2018-11-06 南京理工大学 A kind of adaptive CKF filtering methods of missile-borne BDS/SINS deep combinations
CN109000642A (en) * 2018-05-25 2018-12-14 哈尔滨工程大学 A kind of improved strong tracking volume Kalman filtering Combinated navigation method
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
CN109324330A (en) * 2018-09-18 2019-02-12 东南大学 Based on USBL/SINS tight integration navigation locating method of the mixing without derivative Extended Kalman filter
CN110398257A (en) * 2019-07-17 2019-11-01 哈尔滨工程大学 The quick initial alignment on moving base method of SINS system of GPS auxiliary

Patent Citations (6)

* 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
CN109000642A (en) * 2018-05-25 2018-12-14 哈尔滨工程大学 A kind of improved strong tracking volume Kalman filtering Combinated navigation method
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
CN108761512A (en) * 2018-07-28 2018-11-06 南京理工大学 A kind of adaptive CKF filtering methods of missile-borne BDS/SINS deep combinations
CN109324330A (en) * 2018-09-18 2019-02-12 东南大学 Based on USBL/SINS tight integration navigation locating method of the mixing without derivative Extended Kalman filter
CN110398257A (en) * 2019-07-17 2019-11-01 哈尔滨工程大学 The quick initial alignment on moving base method of SINS system of GPS auxiliary

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
杨建国: "基于GNSS/SINS雷达导引头组合导航系统研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112665581A (en) * 2020-12-04 2021-04-16 山东省计算中心（国家超级计算济南中心） Combined navigation method based on BP neural network assisted Kalman filtering
CN112762932A (en) * 2021-04-07 2021-05-07 智道网联科技(北京)有限公司 Trajectory calculation method and device based on neural network model
CN112762932B (en) * 2021-04-07 2021-07-13 智道网联科技(北京)有限公司 Trajectory calculation method and device based on neural network model
CN113465628A (en) * 2021-06-17 2021-10-01 杭州鸿泉物联网技术股份有限公司 Inertial measurement unit data compensation method and system
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
CN117310773A (en) * 2023-11-30 2023-12-29 山东省科学院海洋仪器仪表研究所 Autonomous positioning method and system for underwater robot based on binocular stereoscopic vision
CN117310773B (en) * 2023-11-30 2024-02-02 山东省科学院海洋仪器仪表研究所 Autonomous positioning method and system for underwater robot based on binocular stereoscopic vision

Similar Documents

Publication Publication Date Title
CN111290007A (en) BDS/SINS combined navigation method and system based on neural network assistance
CN110487301B (en) Initial alignment method of radar-assisted airborne strapdown inertial navigation system
CN101949703B (en) Strapdown inertial/satellite combined navigation filtering method
CN104181572B (en) Missile-borne inertia/ satellite tight combination navigation method
CN109459044B (en) GNSS dual-antenna assisted vehicle-mounted MEMS inertial navigation combined navigation method
CN101246011B (en) Multi-target multi-sensor information amalgamation method based on convex optimized algorithm
CN104764467B (en) Re-entry space vehicle inertial sensor errors online adaptive scaling method
CN111121766B (en) Astronomical and inertial integrated navigation method based on starlight vector
CN111982106A (en) Navigation method, navigation device, storage medium and electronic device
CN106405670A (en) Gravity anomaly data processing method applicable to strapdown marine gravimeter
CN103674030A (en) Dynamic measuring device and method for plumb line deviation kept on basis of astronomical attitude reference
CN107270893A (en) Lever arm, time in-synchronization error estimation and the compensation method measured towards real estate
CN111735474B (en) Moving base compass alignment method based on data backtracking
CN109708663B (en) Star sensor online calibration method based on aerospace plane SINS assistance
CN111722295B (en) Underwater strapdown gravity measurement data processing method
CN103900565A (en) Method for obtaining inertial navigation system attitude based on DGPS (differential global positioning system)
CN110006427B (en) BDS/INS tightly-combined navigation method in low-dynamic high-vibration environment
CN107677292B (en) Vertical line deviation compensation method based on gravity field model
CN103900608A (en) Low-precision inertial navigation initial alignment method based on quaternion CKF
CN112504275A (en) Water surface ship horizontal attitude measurement method based on cascade Kalman filtering algorithm
CN111156986B (en) Spectrum red shift autonomous integrated navigation method based on robust adaptive UKF
CN115878939A (en) High-precision dynamic measurement method based on aircraft control surface deflection
CN110514200B (en) Inertial navigation system and high-rotation-speed rotating body attitude measurement method
CN111964675A (en) Intelligent aircraft navigation method for blackout area
CN117053802A (en) Method for reducing positioning error of vehicle navigation system based on rotary MEMS IMU

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20200616

RJ01 Rejection of invention patent application after publication