CN109059909A - Satellite based on neural network aiding/inertial navigation train locating method and system - Google Patents
Satellite based on neural network aiding/inertial navigation train locating method and system Download PDFInfo
- Publication number
- CN109059909A CN109059909A CN201810813630.9A CN201810813630A CN109059909A CN 109059909 A CN109059909 A CN 109059909A CN 201810813630 A CN201810813630 A CN 201810813630A CN 109059909 A CN109059909 A CN 109059909A
- Authority
- CN
- China
- Prior art keywords
- data
- satellite
- train
- neural network
- speed
- 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
Links
Classifications
-
- 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
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/20—Instruments for performing navigational calculations
-
- G—PHYSICS
- 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
- 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
- G01S19/50—Determining position whereby the position solution is constrained to lie upon a particular curve or surface, e.g. for locomotives on railway tracks
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Automation & Control Theory (AREA)
- Navigation (AREA)
Abstract
The present invention provides a kind of satellite based on neural network aiding/inertial navigation train locating method and system, method therein includes: to detect to the satellite data of acquisition, if positioning accuracy is good, it is then transferred to data fusion module and carries out data fusion, if positioning accuracy is unsatisfactory for requiring, give up this group of satellite data；Then, inertial reference calculation module resolves the data of collected Inertial Measurement Unit, and the positioning result of calculation result and satellite is carried out fusion treatment.Finally, the result after fusion treatment is exported, incorporate in car-mounted computer.Using the present invention, it is able to solve the problems such as data reliability is low, precision is low in existing train positioning system.
Description
Technical field
The present invention relates to vehicle positioning technology fields, more specifically, are related to a kind of defending based on neural network aiding
Star/inertial navigation train locating method and system.
Background technique
Current existing train locating method is mainly include the following types: odometer, inquiry response devices, Doppler radar, rail
Road circuit.Wherein odometer is at low cost, and position is obtained by rate integrating, and there are the accumulation of error；Doppler radar is mainly used for
It tests the speed, precision is higher, but equally exists the accumulation of error, and cost is also higher；It is largely laid with transponder on ground and track circuit is used
Accumulated error is positioned to eliminate, while needing to increase ON TRAINS corresponding reception device, but needs a large amount of cost, and ground
Face equipment needs periodic maintenance, and maintenance engineering amount is big, low efficiency.Train speed is mainly grinding for train control system with location information
Study carefully object, measurement error will directly influence train security protection distance, tracking interval, and occlusion control mode etc. is missed
Difference crosses senior general and directly touches train braking, influences train efficiency and comfort of passenger, even jeopardizes train driving peace when serious
Entirely.
In recent years, Technique of Satellite Navigation and Positioning is gradually in the research of train positioning field and application, Technique of Satellite Navigation and Positioning
With round-the-clock, continuous, real-time, long-term accuracy is high advantage, the accurate positioning of train can be provided, however satellite-signal holds
Vulnerable to environment interference and cause positioning accuracy to decline rapidly, therefore only using satellite as positioning information source, safety with
Reliability is insufficient.Inertial navigation, which has, does not depend on the entirely autonomous sex work of external environment, navigation information it is complete (posture, speed,
Position), dynamic and good etc. the feature of continuity, can be to provide high-precision location information in the short time, but inertial navigation needs
The initial information of train is wanted, position error dissipates at any time, and two kinds of navigation modes, which combine, can achieve excellent scarce complementation, therefore defend
Star/inertial navigation is one of the integrated navigation mode being most widely used at present.
Satellite/inertial navigation train positioning system precision depends on satellite-signal, when satellite is good, when system can be long
Between the high-precision location data of continual output, provide reliable location information for the safe driving etc. of train, and work as satellite
Poor signal, such as the place that train has high building or trees to block by city, forest etc., have multipath effect to lead to satellite
Locating effect influences system output there are large error, or even under cavern, the environment such as tunnel when the complete losing lock of satellite, at this time
System output calculates and outputs after acquiring data by Inertial Measurement Unit, due to the accumulation of error of Inertial Measurement Unit, so that system
Output accuracy be gradually reduced, and be exactly the area of Frequent Accidents in the environment of the satellites losing lock such as cavern, tunnel, therefore at this
Under a little environment, reliable accurately train position information is even more important for the safe driving of train.This improves satellite mistakes
The positioning accuracy for locking situation Train, also means that the reliability and security for improving train positioning system.
To solve the problems, such as that data reliability is low in existing train positioning system, precision is low, the present invention provides a kind of bases
In the satellite/inertial navigation train locating method and system of neural network aiding.
Summary of the invention
In view of the above problems, the object of the present invention is to provide a kind of, and the satellite based on neural network aiding/inertial navigation train is fixed
Position method and system, to solve the problems such as data reliability is low, precision is low in existing train positioning system.
The present invention provides a kind of satellite based on neural network aiding/inertial navigation train locating method, comprising:
Obtain sensor real time data, wherein the sensor real time data includes Inertial Measurement Unit data and satellite
Data；
The Inertial Measurement Unit data are resolved by inertial reference calculation module, obtain the posture of train, speed and
Position；
The satellite data is detected, wherein detection satellite ephemeris observation data and satellite positioning dilution of precision value
Whether the condition of satisfaction setting, if meeting condition, this satellite data is available, if being unsatisfactory for condition, this satellite data
It is unavailable；
When satellite data can be used, by the satellite data and the column by being obtained after inertial reference calculation module resolving
Posture, speed and the location information of vehicle carry out fusion treatment, obtain the speed of train and position after correcting；
Meanwhile posture, the speed of the train obtained after being resolved by the satellite data and by the inertial reference calculation module
It is inputted with location information as neural network, is exported system-computed error amount as neural network, neural network is instructed
Practice；
When satellite data is unavailable, by the output valve after neural metwork training and pass through the inertial reference calculation module solution
Posture, speed and the location information of the train obtained after calculation carry out fusion treatment, obtain the speed of train and position after correcting.
The present invention also provides a kind of satellite based on neural network aiding/inertial navigation train positioning system, comprising:
Inertial reference calculation module obtains the posture, speed and position of train for obtaining Inertial Measurement Unit data in real time；
Satellite signal reception module for obtaining satellite data in real time, and detects the satellite data, wherein
Whether detection satellite ephemeris observation data and satellite positioning dilution of precision value meet the condition of setting, if meeting condition, this
Satellite data is available, if being unsatisfactory for condition, this satellite data is unavailable；
Data fusion module, for by the satellite data and passing through the inertial reference calculation mould when satellite data can be used
Posture, speed and the location information for the train that block obtains after resolving carry out fusion treatment, obtain the speed of train and position after correcting
It sets；
Meanwhile posture, the speed of the train obtained after being resolved by the satellite data and by the inertial reference calculation module
It is inputted with location information as neural network, is exported system-computed error amount as neural network, neural network is instructed
Practice；
When satellite data is unavailable, by the output valve after neural metwork training and pass through the inertial reference calculation module solution
Posture, speed and the location information of the train obtained after calculation carry out fusion treatment, obtain the speed of train and position after correcting.
It can be seen from the above technical scheme that the satellite provided by the invention based on neural network aiding/inertial navigation train positioning
Method and system, can obtain it is following the utility model has the advantages that
1, train locating method proposed by the present invention can not only improve the essence of train in the good situation of satellite-signal
Degree, and the train positioning result of degree of precision can be provided under conditions of satellite-signal exception.
2, train locating method proposed by the present invention uses the data fusion of neural network aiding, can effectively inhibit to defend
The diverging of system position error in the case where star exception, additionally it is possible to improve satellite/inertial navigation train positioning system reliability and peace
Quan Xing.
3, it works independently between each module of train positioning system proposed by the present invention, is independent of each other, can be improved system
Stability and anti-interference.
To the accomplishment of the foregoing and related purposes, one or more aspects of the present invention includes being particularly described below
Feature.Certain illustrative aspects of the invention is described in detail in the following description and the annexed drawings.However, what these aspects indicated
Some of the various ways in the principles of the present invention only can be used.In addition, the present invention is intended to include all these sides
Face and their equivalent.
Detailed description of the invention
By reference to the explanation below in conjunction with attached drawing, and with a more complete understanding of the present invention, of the invention is other
Purpose and result will be more clearly understood and understood.In the accompanying drawings:
Fig. 1 is to be shown according to the satellite based on neural network aiding/inertial navigation train locating method process of the embodiment of the present invention
It is intended to；
Fig. 2 is to be shown according to the satellite based on neural network aiding/inertial navigation train positioning system structure of the embodiment of the present invention
It is intended to；
Fig. 3 is according to the embodiment of the present invention based on neural network aiding satellite/inertial navigation train positioning calculation algorithm frame
Figure；
Fig. 4 is the inertial navigation more new algorithm schematic diagram according to the embodiment of the present invention；
Fig. 5 is the data anastomosing algorithm flow diagram according to the embodiment of the present invention；
Fig. 6 is the neural network model schematic diagram according to the embodiment of the present invention；
Fig. 7 is the neural metwork training logical schematic according to the embodiment of the present invention；
Fig. 8 is according to the neural network of embodiment of the present invention predictive filtering measuring value logical schematic in systems.
Identical label indicates similar or corresponding feature or function in all the appended drawings.
Specific embodiment
In the following description, for purposes of illustration, in order to provide the comprehensive understanding to one or more embodiments,
Numerous specific details are set forth.It may be evident, however, that these implementations can also be realized without these specific details
Example.
For the problems such as data reliability is low, precision is low in the existing train positioning system of aforementioned proposition, the present invention is mentioned
A kind of satellite based on neural network aiding/inertial navigation train locating method and system are supplied.
Hereinafter, specific embodiments of the present invention will be described in detail with reference to the accompanying drawings.
In order to illustrate the satellite based on neural network aiding/inertial navigation train locating method provided by the invention, Fig. 1 is shown
Satellite according to an embodiment of the present invention based on neural network aiding/inertial navigation train locating method process.
As shown in Figure 1, the satellite based on neural network aiding/inertial navigation train locating method provided by the invention includes:
S110: sensor real time data is obtained, wherein sensor real time data includes Inertial Measurement Unit data and satellite data；
S120: Inertial Measurement Unit data are resolved by inertial reference calculation module, obtain posture, the speed of train
The position and；
S130: satellite data is detected, wherein detection satellite ephemeris observation data and satellite positioning dilution of precision
Whether value meets the condition of setting, if meeting condition, this satellite data is available, if being unsatisfactory for condition, this satellite number
According to unavailable；
S140: when satellite data can be used, by satellite data and the train by being obtained after the resolving of inertial reference calculation module
Posture, speed and location information carry out fusion treatment, obtain the speed of train and position after correcting；
S150: meanwhile, by posture, the speed of the satellite data and the train by being obtained after the resolving of inertial reference calculation module
Degree and location information are inputted as neural network, are exported system-computed error amount as neural network, are carried out to neural network
Training；
S160: when satellite data is unavailable, by the output valve after neural metwork training and pass through inertial reference calculation module solution
Posture, speed and the location information of the train obtained after calculation carry out fusion treatment, obtain the speed of train and position after correcting.
Fig. 1 combination Fig. 3, the satellite based on neural network aiding/inertial navigation train locating method provided by the invention it is detailed
Steps are as follows: step 1:IMU data acquire S101, satellite data receives S102；
Step 2: satellite-signal detection, satellite-signal detection algorithm refer to detection satellite ephemeris observation data and satellite
Whether position dilution of precision value meets the condition of setting, if meeting condition, this satellite data is available, if being unsatisfactory for condition,
Then this satellite data is unavailable；If being unsatisfactory for condition, the measuring value S106 as data fusion is exported using neural network,
If meeting condition, this group of satellite data is exported into S105.
Wherein, it should be noted that be according to the condition that the actual demand of application is set in the present invention: number of satellite >
6 or HDOP < 10, that is, when meet the condition that one of both or both all meets the case where, satellite data is available
S105, if the condition of one of both is all unsatisfactory for, this satellite data cannot use S106.It, then can be with when number of satellite is 6
Meet the precision of satellite data detection, when being greater than 6, then the precision of satellite data detection can be better met, so in the present invention
The middle one of condition set is number of satellite > 6.
Step 3: inertial reference calculation S103, the inertial reference calculation refer to the data using collected Inertial Measurement Unit into
Row resolves, and obtains the posture, speed and position of carrier, and the result of inertial reference calculation is exported S104；
Step 4: data fusion S107.The information (when satellite data can be used) of satellite receiver output, neural network is pre-
The result of measured value (when satellite data is unavailable) and inertial reference calculation carries out fusion treatment, obtains modified train position estimation；
Step 5: the result of step 4 is output to S108 in car-mounted computer；
Step 6: neural metwork training S109.Sensor information in system and the information of resolving is defeated as neural network
Enter, using system-computed error amount as desired output, training neural network reaches up to the precision met the requirements.
Wherein, obtaining sensor real time data further comprises: three axis (x, y, z) angular speed of gyroscope output accelerates
Three axis (x, y, z) acceleration of degree meter output, longitude, the latitude, height, east orientation speed, north orientation speed of satellite receiver output
And sky orientation speed.
As shown in figure 3, a kind of satellite based on neural network aiding/inertial navigation train locating method provided by the invention, is used to
It is as follows to lead computation detailed process:
(1) speed updates
Speed renewal equation are as follows:
Wherein,
ΔvmIt is accelerometer in period [tm-1,tm] in export specific force increment, in practice using specific force output multiplied by
Sampling interval carries out approximate.ΔθmIt is gyro in period [tm-1,tm] in export angle increment, by it multiplied by sampling interval Ts
Can approximate transform be angle increment.
(2) earth parameter calculates
The meridian circle principal radius of curvature and the prime vertical principal radius of curvature are calculated, calculation formula is as follows:
Wherein, e is oval eccentric rate, and f is ellipticity of ellipse, f=1/298.257, ReFor earth radius, RMFor meridian circle master
Radius of curvature, RNFor the prime vertical principal radius of curvature.
(3) location updating
Location updating equation are as follows:
Wherein, T is the sampling period, and λ is the longitude of train position, and L is the latitude of train position, and h is train
The height of position.
(4) posture renewal
It is updated using the angular speed of gyroscope measurement as quaternary number, it is real using acceleration of gravity as the observation of quaternary number
When resolve attitude angle.
The matrix form of the attitude quaternion differential equation are as follows:
Wherein, T is sampling interval, q0、q1、q2、q3Indicate the quaternary number of posture；
ωx、ωy、ωzIndicate three axis (x, y, the z) output of gyroscope.
Updated attitude quaternion is normalized:
Wherein, subscript i=0,1,2,3, respectively represent each value in attitude quaternion.
The attitude angle of carrier can be found out according to the quaternary number after normalization.
As shown in figure 5, a kind of satellite based on neural network aiding/inertial navigation train locating method provided by the invention, number
It is as follows according to blending algorithm detailed process:
(1) satellite/inertial navigation train positioning system mathematical model is established:
System state equation are as follows:
Xk=Φk/k-1Xk-1+Γk/k-1Wk-1
Wherein, XkIndicate the state vector at k moment, Φk/k-1Indicate the state Matrix of shifting of a step of 15 × 15 dimensions, Γk/k-1
Indicate the system noise allocation matrix of 15 × 15 dimensions, Wk-1Indicate the system noise vector of 15 × 1 dimensions, it is the Gauss of zero-mean
White noise sequence vector.
System measurements equation are as follows:
Zk=HkXk+Vk
Wherein, ZkIndicate system measurements vector, HkIndicate the measurement matrix of 6 × 15 dimensions, VkIndicate that the measurement of 6 × 1 dimensions is made an uproar
Sound vector, it is the white Gaussian noise sequence vector of zero-mean.
(2) systematic error equation is established
The equation of simplified systematic error is established according to systematic observation matrix are as follows:
Wherein,WithAngular speed and its error of the n system relative to i system are respectively indicated,WithAngular speed and its error of the n system relative to e system are respectively indicated,WithRespectively indicate angular speed of the e system relative to i system
And its error,Direction cosine matrix for b system relative to n system.State Matrix of shifting of a step Φ is arrived as available from the above equationk/k-1。
(3) data fusion:
Data anastomosing algorithm is divided into two parts, and forecast updating and measurement update, and forecast updating refers to is by foundation
The system mode of system model and previous moment updates the system mode at this moment, measures update and refers to after forecast updating, benefit
System mode is updated with measuring value, obtains final system mode value.After forecast updating, detect whether new
Measuring value, if so, then carry out measurement update, if it is not, directly with resolving of the value of forecast updating to inertial navigation be worth into
Row amendment.
(4) it is corrected using velocity location value of the system mode vector after data fusion to system:
Wherein,The respectively velocity error estimated value in three direction of system (east-north-day),Respectively system latitude, longitude, height error estimated value.
(5) revised data are exported.
As shown in fig. 6, a kind of satellite based on neural network aiding/inertial navigation train locating method provided by the invention, mind
It is three layers, i.e. input layer, hidden layer, output layer through network topology structure.In Fig. 6, φ1、 φ2…φkIndicate that hidden layer is each
The radial basis function value of node, ω1、ω2、ω3…ωkValue information of the expression each node of hidden layer to output layer.X is network
Input, y be neural network output.
As illustrated in figs. 7 and 8, a kind of satellite based on neural network aiding/inertial navigation train positioning side provided by the invention
The scheme of method, the positioning of neural network aiding train is as follows: when satellite-signal is normal, being exported using satellite and inertial reference calculation
Speed position information estimates the Position And Velocity information of train after being merged, while using the neural network input chosen
Neural network is trained with exporting；Measurement when satellite-signal is unavailable, using the output valve of neural network as data fusion
Value obtains satisfied location data so that filter continues as inertial navigation system and provides correction value.
In conclusion a kind of satellite based on neural network aiding/inertial navigation train locating method provided by the invention, mind
The scheme positioned through network assistance train are as follows: when satellite-signal is normal, using the velocity location of satellite and inertial reference calculation output
Information estimates the Position And Velocity information of train after being merged, while being instructed using the neural network input chosen with output
Practice neural network；When satellite data is unavailable, measuring value using the output valve of neural network as data fusion, so that
Inertial navigation system continues to obtain correction value.
It corresponds to the above method, the satellite based on neural network aiding/inertial navigation train positioning that the present invention also provides a kind of
System, the satellite based on neural network aiding/inertial navigation train positioning system logic that Fig. 2 shows according to an embodiment of the present invention
Structure.
As shown in Fig. 2, a kind of satellite based on neural network aiding/inertial navigation train positioning system packet provided by the invention
Include safety power supply module 1, inertial reference calculation module 2, satellite signal reception module 3, data fusion module 4, data outputting module 5,
Car-mounted computer 6.
Wherein, inertial reference calculation module 2 21 is made of the acquisition of IMU data with 22 two submodules of data calculation, be responsible for by
Collected IMU data calculation obtains speed, position and the posture information of carrier, establishes carrier coordinate system to navigational coordinate system
Posture transfer matrix.
Satellite signal reception module 3 receives 31 by satellite data and forms with 32 two submodules of Data Detection, is responsible for connecing
The data of satellite are received, and the data of satellite are detected, output meets the satellite data of testing conditions.
Data fusion module 4 is by 43 3 data fusion 41, data correction 42, neural metwork training and prediction submodules
Composition is responsible for carrying out fusion treatment to the data of inertial reference calculation and the data of satellite, and the correction value that fusion is obtained is to inertial navigation
The data of resolving are modified, and obtain the position estimation value of train.
Car-mounted computer includes that data show that 61 save 62 with data.Data display sub-module 61 is responsible for will be transmitted to vehicle
The data for carrying computer carry out real-time display, and the data of display include: satellite/inertial navigation train location data, train travel track
Curve, satellite ephemeris observation data and dilution of precision, train attitude angle real-time resolving curve, three axle speed real-time resolving of train
Curve.The data that data preservation submodule will be transmitted to car-mounted computer are stored in car-mounted computer, for later reference.
Above-mentioned module connection relationship is as follows: inertial reference calculation module 2 is electrically connected data fusion module 4；Satellite signal receiving mould
Block 3 is electrically connected data fusion module 4；Data fusion module 4 is electrically connected data outputting module 5；Data outputting module 5 is electrically connected
Car-mounted computer 6；Safety power supply module 1 respectively with inertial reference calculation module 2, satellite signal reception module 3, data fusion module 4
And data outputting module 5 is electrically connected.
As shown in Fig. 2, a kind of satellite based on neural network aiding/inertial navigation train locating method provided by the invention
Core is the train integrated positioning algorithm of neural network aiding, which can be roughly divided into inertial reference calculation algorithm, defend
Three star signal detection algorithm, data anastomosing algorithm modules, these three modules are independent in structure, but in each calculating
It is to be mutually related in period, the input of system includes: 3-axis acceleration information and three axis from Inertial Measurement Unit output
Angular rate information and the satellite ephemeris of satellite receiver observe data, Horizontal Dilution of Precision, longitude, latitude, height, east orientation
Speed, north orientation speed, sky orientation speed, system output is the train speed estimated and position.
Wherein, inertial reference calculation module includes that Inertial Measurement Unit (IMU) signal acquisition and inertial guidance data resolve, and inertia is surveyed
Amount unit (IMU) sensor model number is 3DM-IMU200A, carries out microprocessor model used in inertial reference calculation and is
STM32F103C6T6。
Satellite signal reception module includes satellite data acquisition and Data Detection, what satellite data acquisition sensor used
It is GNSS receiver K700.
Data fusion module includes data fusion, neural metwork training and prediction module, data correction, data fusion mould
The microprocessor model that block uses is STM32F103C6T6.
Car-mounted computer includes that data are shown and data preservation, the data of display are as follows: satellite/inertial navigation train location data,
Train travel geometric locus, satellite ephemeris observation data and dilution of precision, train attitude angle real-time resolving curve, three axis of train
Speed (tri- axis of xyz) real-time resolving curve.The data of preservation have: the positional number that train position estimated value, satellite receiver export
According to, IMU original signal and corresponding temporal information.Data outputting module is serial using RS422 with vehicle computing mechatronics
Communication connection.
By above embodiment as can be seen that the satellite based on neural network aiding/inertial navigation train provided by the invention
Localization method and system can not only improve the precision of train in the good situation of satellite-signal, and can believe in satellite
The train positioning result of degree of precision is provided under conditions of number exception；It, can be effective using the data fusion of neural network aiding
The diverging of system position error in the case where inhibition satellite exception, additionally it is possible to it is reliable to improve satellite/inertial navigation train positioning system
Property and safety；It works independently, is independent of each other between each module of train positioning system, can be improved the stability of system and resist
Interference.
Describe the defending based on neural network aiding proposed according to the present invention in an illustrative manner above with reference to attached drawing
Star/inertial navigation train locating method and system.It will be understood by those skilled in the art, however, that aforementioned present invention is proposed
Satellite based on neural network aiding/inertial navigation train locating method and system, can also be on the basis for not departing from the content of present invention
On make various improvement.Therefore, protection scope of the present invention should be determined by the content of appended claims.
Claims (8)
1. a kind of satellite based on neural network aiding/inertial navigation train locating method, comprising:
Obtain sensor real time data, wherein the sensor real time data includes Inertial Measurement Unit data and satellite data；
The Inertial Measurement Unit data are resolved by inertial reference calculation module, obtain the posture, speed and position of train；
The satellite data is detected, wherein detection satellite ephemeris observation data and satellite positioning dilution of precision value whether
Meet the condition of setting, if meeting the condition of setting, this satellite data is available, if being unsatisfactory for the condition of setting, this is defended
Sing data is unavailable；
When satellite data can be used, the posture of the train obtained after being resolved by the satellite data and by inertial reference calculation module,
Speed and location information carry out fusion treatment, obtain the speed of train and position after correcting；
Meanwhile posture, speed and the position of the train obtained after being resolved by the satellite data and by the inertial reference calculation module
Confidence breath is inputted as neural network, is exported system-computed error amount as neural network, is trained to neural network；
When satellite data is unavailable, obtained after being resolved by the output valve after neural metwork training and by inertial reference calculation module
Posture, speed and the location information of train carry out fusion treatment, obtain the speed of train and position after correcting.
2. the satellite based on neural network aiding/inertial navigation train locating method as described in claim 1, wherein the inertia
Measuring unit data include three axis (x, y, z) angular speed of gyroscope output, and three axis (x, y, z) of accelerometer output accelerate
Degree；
The satellite data includes longitude, latitude, height, east orientation speed, north orientation speed and the day of satellite receiver output to speed
Degree.
3. the satellite based on neural network aiding/inertial navigation train locating method as described in claim 1, wherein by used
It leads and resolves during module resolves the Inertial Measurement Unit data,
(1) speed updates
Speed renewal equation are as follows:
Wherein,
Wherein, Δ vmIt is accelerometer in period [tm-1,tm] in export specific force increment；ΔθmIt is gyro in the period
[tm-1,tm] in export angle increment, by it multiplied by sampling interval TsCan approximate transform be angle increment；
(2) earth parameter calculates
The meridian circle principal radius of curvature and the prime vertical principal radius of curvature are calculated, calculation formula is as follows:
Wherein, e is oval eccentric rate, and f is ellipticity of ellipse, f=1/298.257, ReFor earth radius, RMFor meridian circle principal curvatures
Radius, RNFor the prime vertical principal radius of curvature；
(3) location updating
Location updating equation are as follows:
Wherein, T is the sampling period, and λ is the longitude of train position, and L is the latitude of train position, and h is train place
The height of position；
(4) posture renewal
It is updated using the angular speed of gyroscope measurement as quaternary number, using acceleration of gravity as the observation of quaternary number, is solved in real time
Calculate attitude angle；
Wherein, the matrix form of the attitude quaternion differential equation are as follows:
Wherein, T is the sampling interval；q0、q1、q2、q3Indicate the quaternary number of posture；
ωx、ωy、ωzIndicate three axis (x, y, the z) output of gyroscope；
Updated attitude quaternion is normalized:
Wherein, subscript i=0,1,2,3, respectively represent each value in attitude quaternion；
The attitude angle of carrier can be found out according to the quaternary number after normalization.
4. the satellite based on neural network aiding/inertial navigation train locating method as described in claim 1, wherein working as satellite
When data can be used, by the satellite data with by being obtained after inertial reference calculation module resolving the posture of train, speed and
Location information carry out fusion treatment, obtain amendment after train speed and position during,
(1) satellite/inertial navigation train positioning system mathematical model is established:
System state equation are as follows:
Xk=Φk/k-1Xk-1+Γk/k-1Wk-1
Wherein, XkIndicate the state vector at k moment, Φk/k-1Indicate the state Matrix of shifting of a step of 15 × 15 dimensions, Γk/k-1It indicates
The system noise allocation matrix of 15 × 15 dimensions, Wk-1The system noise vector for indicating 15 × 1 dimensions is the white Gaussian noise of zero-mean
Sequence vector；
System measurements equation are as follows:
Zk=HkXk+Vk
Wherein, ZkIndicate system measurements vector, HkIndicate the measurement matrix of 6 × 15 dimensions, VkIndicate 6 × 1 dimension measurement noise to
Amount is the white Gaussian noise sequence vector of zero-mean；
(2) systematic error equation is established
The equation of simplified systematic error is established according to systematic observation matrix are as follows:
Wherein, WithAngular speed and its error of the n system relative to i system are respectively indicated,WithRespectively
Indicate angular speed and its error of the n system relative to e system,WithAngular speed and its error of the e system relative to i system are respectively indicated,Direction cosine matrix for b system relative to n system；
State Matrix of shifting of a step Φ is arrived as available from the above equationk/k-1；
(3) data fusion
Data anastomosing algorithm is divided into two parts, and forecast updating and measurement update, and forecast updating refers to the system mould by foundation
The system mode of type and previous moment updates the system mode at this moment, measures update and refers to after forecast updating, utilization
Measured value is updated system mode, obtains final system mode value；After forecast updating, new measurement is detected whether
Value, if so, measurement update is then carried out, if it is not, being directly modified with resolving value of the value of forecast updating to inertial navigation；
(4) it is corrected using velocity location value of the system mode vector after data fusion to system:
Wherein,The respectively velocity error estimated value in three direction of system, Respectively system latitude, longitude, height error estimated value.
5. the satellite based on neural network aiding/inertial navigation train locating method as described in claim 1, wherein will be described
Posture, speed and the location information of satellite data and the train by obtaining after inertial reference calculation module resolving are as nerve net
Network input, exports system-computed error amount as neural network, during being trained to neural network,
The input value of neural network is respectively as follows: with desired output
Input:
Output: [δ v δ p]=[δ vE δvN δvU δl δλ δh]
The input of the neural network is respectively as follows: the speed of the output increment of accelerometer east orientation and north orientation, system east orientation and north orientation
Spend increment, pitch angle, roll angle, course angle error, pitch angle and roll angle；
The output of the neural network is respectively as follows: three direction velocity error of system, latitude error, longitude error and height accidentally
Difference.
6. a kind of satellite based on neural network aiding/inertial navigation train positioning system, comprising:
Inertial reference calculation module obtains the posture, speed and position of train for obtaining Inertial Measurement Unit data in real time；
Satellite signal reception module for obtaining satellite data in real time, and detects the satellite data, wherein detection
Whether satellite ephemeris observation data and satellite positioning dilution of precision value meet the condition of setting, if meeting condition, this satellite
Data are available, if being unsatisfactory for condition, this satellite data is unavailable；
Data fusion module, for by the satellite data and passing through the inertial reference calculation module solution when satellite data can be used
Posture, speed and the location information of the train obtained after calculation carry out fusion treatment, obtain the speed of train and position after correcting；
Meanwhile posture, speed and the position of the train obtained after being resolved by the satellite data and by the inertial reference calculation module
Confidence breath is inputted as neural network, is exported system-computed error amount as neural network, is trained to neural network；
When satellite data is unavailable, by after neural metwork training output valve with by the inertial reference calculation module resolve after obtain
Posture, speed and the location information of the train taken carry out fusion treatment, obtain the speed of train and position after correcting.
7. the satellite based on neural network aiding/inertial navigation train positioning system as claimed in claim 6, wherein further include peace
Full voltage module, data outputting module and car-mounted computer,
The safe voltage module, for respectively to the inertial reference calculation module, satellite signal reception module, data fusion mould
Block, data outputting module provide power supply；
The data outputting module, for the fused data of data fusion module to be transmitted to the car-mounted computer.
8. the satellite based on neural network aiding/inertial navigation train positioning system as claimed in claim 6, wherein described used
Property to resolve the Inertial Measurement Unit sensor model number of module be 3DM-IMU200A, microprocessor model used in inertial reference calculation is
STM32F103C6T6；
Acquire the GNSS receiver K700 that the sensor of satellite data uses；
The microprocessor model that data fusion module uses is STM32F103C6T6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810813630.9A CN109059909A (en) | 2018-07-23 | 2018-07-23 | Satellite based on neural network aiding/inertial navigation train locating method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810813630.9A CN109059909A (en) | 2018-07-23 | 2018-07-23 | Satellite based on neural network aiding/inertial navigation train locating method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109059909A true CN109059909A (en) | 2018-12-21 |
Family
ID=64835137
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810813630.9A Pending CN109059909A (en) | 2018-07-23 | 2018-07-23 | Satellite based on neural network aiding/inertial navigation train locating method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109059909A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109781099A (en) * | 2019-03-08 | 2019-05-21 | 兰州交通大学 | A kind of navigation methods and systems of adaptive UKF algorithm |
| CN109975851A (en) * | 2019-04-04 | 2019-07-05 | 河南思维轨道交通技术研究院有限公司 | A kind of train line fault point accurate positioning method and system |
| CN110395297A (en) * | 2019-07-29 | 2019-11-01 | 兰州交通大学 | Train locating method |
| CN110553644A (en) * | 2019-09-20 | 2019-12-10 | 鞍钢集团矿业有限公司 | Accurate positioning system and method for mining electric shovel |
| CN112083465A (en) * | 2020-09-18 | 2020-12-15 | 德明通讯（上海）有限责任公司 | Position information acquisition system and method |
| CN112113566A (en) * | 2020-09-24 | 2020-12-22 | 电子科技大学 | Inertial navigation data correction method based on neural network |
| CN112363196A (en) * | 2020-10-20 | 2021-02-12 | 北京嘀嘀无限科技发展有限公司 | Vehicle attribute determination method and device, storage medium and electronic equipment |
| CN113124884A (en) * | 2021-04-16 | 2021-07-16 | 智道网联科技(北京)有限公司 | Vehicle positioning method and device based on LSTM neural network model |
| CN113431550A (en) * | 2021-07-06 | 2021-09-24 | 中国矿业大学 | Anti-impact drilling robot drilling tool attitude determination method based on redundant inertial unit |
| CN113810877A (en) * | 2021-08-11 | 2021-12-17 | 上海商泰汽车信息系统有限公司 | Vehicle, vehicle speed determination method and device, storage medium and terminal |
| CN115855066A (en) * | 2023-02-22 | 2023-03-28 | 湖南迈克森伟电子科技有限公司 | High-speed rail coordinate positioning correction method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101871782A (en) * | 2010-05-19 | 2010-10-27 | 北京航空航天大学 | Position error forecasting method for GPS (Global Position System)/MEMS-INS (Micro-Electricomechanical Systems-Inertial Navigation System) integrated navigation system based on SET2FNN |
| CN102768361A (en) * | 2012-07-09 | 2012-11-07 | 东南大学 | GPS/INS combined positioning method based on genetic particle filtering and fuzzy neural network |
| CN106595652A (en) * | 2016-11-30 | 2017-04-26 | 西北工业大学 | Vehicle MCA (motion constraints aided) backtracking type aligning-on-the-move method |
| 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 |
-
2018
- 2018-07-23 CN CN201810813630.9A patent/CN109059909A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101871782A (en) * | 2010-05-19 | 2010-10-27 | 北京航空航天大学 | Position error forecasting method for GPS (Global Position System)/MEMS-INS (Micro-Electricomechanical Systems-Inertial Navigation System) integrated navigation system based on SET2FNN |
| CN102768361A (en) * | 2012-07-09 | 2012-11-07 | 东南大学 | GPS/INS combined positioning method based on genetic particle filtering and fuzzy neural network |
| CN106595652A (en) * | 2016-11-30 | 2017-04-26 | 西北工业大学 | Vehicle MCA (motion constraints aided) backtracking type aligning-on-the-move method |
| 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 |
Non-Patent Citations (2)
| Title |
|---|
| 崔留争: ""神经网络辅助卡尔曼滤波在组合导航中的应用"", 《光学精密工程》 * |
| 戴邵武: ""北斗双星/SINS组合导航中的捷联惯导算法研究"", 《计算机与数字工程》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109781099A (en) * | 2019-03-08 | 2019-05-21 | 兰州交通大学 | A kind of navigation methods and systems of adaptive UKF algorithm |
| CN109975851A (en) * | 2019-04-04 | 2019-07-05 | 河南思维轨道交通技术研究院有限公司 | A kind of train line fault point accurate positioning method and system |
| CN110395297B (en) * | 2019-07-29 | 2021-08-10 | 兰州交通大学 | Train positioning method |
| CN110395297A (en) * | 2019-07-29 | 2019-11-01 | 兰州交通大学 | Train locating method |
| CN110553644A (en) * | 2019-09-20 | 2019-12-10 | 鞍钢集团矿业有限公司 | Accurate positioning system and method for mining electric shovel |
| CN112083465A (en) * | 2020-09-18 | 2020-12-15 | 德明通讯（上海）有限责任公司 | Position information acquisition system and method |
| CN112113566A (en) * | 2020-09-24 | 2020-12-22 | 电子科技大学 | Inertial navigation data correction method based on neural network |
| CN112363196A (en) * | 2020-10-20 | 2021-02-12 | 北京嘀嘀无限科技发展有限公司 | Vehicle attribute determination method and device, storage medium and electronic equipment |
| CN112363196B (en) * | 2020-10-20 | 2023-10-31 | 北京嘀嘀无限科技发展有限公司 | Vehicle attribute determining method, device, storage medium and electronic equipment |
| CN113124884A (en) * | 2021-04-16 | 2021-07-16 | 智道网联科技(北京)有限公司 | Vehicle positioning method and device based on LSTM neural network model |
| CN113431550A (en) * | 2021-07-06 | 2021-09-24 | 中国矿业大学 | Anti-impact drilling robot drilling tool attitude determination method based on redundant inertial unit |
| CN113810877A (en) * | 2021-08-11 | 2021-12-17 | 上海商泰汽车信息系统有限公司 | Vehicle, vehicle speed determination method and device, storage medium and terminal |
| CN113810877B (en) * | 2021-08-11 | 2023-08-18 | 上海商泰汽车信息系统有限公司 | Vehicle and vehicle speed determining method and device, storage medium and terminal |
| CN115855066A (en) * | 2023-02-22 | 2023-03-28 | 湖南迈克森伟电子科技有限公司 | High-speed rail coordinate positioning correction method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109059909A (en) | Satellite based on neural network aiding/inertial navigation train locating method and system | |
| Chen et al. | Estimate the pitch and heading mounting angles of the IMU for land vehicular GNSS/INS integrated system | |
| US9921065B2 (en) | Unit and method for improving positioning accuracy | |
| Skog et al. | In-car positioning and navigation technologies—A survey | |
| CN106595715B (en) | Based on inertial navigation and satellite combined guidance system mileage meter calibration method and device | |
| CN110779521A (en) | Multi-source fusion high-precision positioning method and device | |
| CN106643709B (en) | Combined navigation method and device for offshore carrier | |
| CN101382431B (en) | Positioning system and method thereof | |
| Iqbal et al. | An integrated reduced inertial sensor system—RISS/GPS for land vehicle | |
| CN109471144A (en) | Based on pseudorange/pseudorange rates multisensor tight integration train combined positioning method | |
| CN105606094B (en) | A kind of information condition matched filtering method of estimation based on MEMS/GPS combined systems | |
| CN107389064A (en) | A kind of unmanned vehicle based on inertial navigation becomes channel control method | |
| CN104729506A (en) | Unmanned aerial vehicle autonomous navigation positioning method with assistance of visual information | |
| CN107247275B (en) | Urban GNSS vulnerability monitoring system and method based on bus | |
| EP2219044A1 (en) | Navigation method, navigation system, navigation device, vehicle provided therewith and group of vehicles | |
| CN102621570B (en) | Automobile dynamic parameter measuring method based on double global positioning and inertia measurement | |
| Gao et al. | Development of precise GPS/INS/wheel speed sensor/yaw rate sensor integrated vehicular positioning system | |
| CN112432642B (en) | Gravity beacon and inertial navigation fusion positioning method and system | |
| Park et al. | MEMS 3D DR/GPS integrated system for land vehicle application robust to GPS outages | |
| CN109813306A (en) | A kind of unmanned vehicle planned trajectory satellite location data confidence level calculation method | |
| CN103968844B (en) | Big oval motor-driven Spacecraft Autonomous Navigation method based on low rail platform tracking measurement | |
| CN106093992A (en) | A kind of sub-meter grade combined positioning and navigating system based on CORS and air navigation aid | |
| CN105910623B (en) | The method for carrying out the correction of course using magnetometer assisted GNSS/MINS tight integration systems | |
| CN103575297A (en) | Estimation method of course angle of GNSS (Global Navigation Satellite System) and MIMU (MEMS based Inertial Measurement Units) integrated navigation based on satellite navigation receiver | |
| US20140249750A1 (en) | Navigational and location determination system |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181221 |