CN112595313A - Vehicle-mounted navigation method and device based on machine learning and computer equipment - Google Patents
Vehicle-mounted navigation method and device based on machine learning and computer equipment Download PDFInfo
- Publication number
- CN112595313A CN112595313A CN202011339096.6A CN202011339096A CN112595313A CN 112595313 A CN112595313 A CN 112595313A CN 202011339096 A CN202011339096 A CN 202011339096A CN 112595313 A CN112595313 A CN 112595313A
- Authority
- CN
- China
- Prior art keywords
- imu
- model
- bds
- vdm
- machine learning
- 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
Images
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
- G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
-
- 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
- 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
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Automation & Control Theory (AREA)
- Computer Networks & Wireless Communication (AREA)
- Navigation (AREA)
Abstract
The application relates to a vehicle navigation method, a device and computer equipment based on machine learning, wherein the method comprises the following steps: establishing a VDM-assisted IMU error correction model, establishing a BDS/IMU/VDM combined navigation model, and establishing a machine learning training model when the BDS is in a normal navigation state, wherein the position and the speed obtained by the IMU through mechanical arrangement and VDM correction are used as training input samples, the BDS/IMU/VDM combined filtering output is used as an output sample, and when the BDS is in an unlocked state, the machine learning model with mature training is started to predict a filtering output value, and the position and speed information solved by the IMU/VDM is corrected. By testing in a tunnel, an underground garage and a satellite signal shielding area, the simplified machine learning network provided by the invention can be used for establishing auxiliary navigation information, realizing real-time auxiliary inertial navigation/wheel speed updating, avoiding long-time positioning error divergence and greatly improving the navigation positioning precision of the system.
Description
Technical Field
The application relates to the technical field of mapping and navigation, in particular to a vehicle-mounted navigation method and device based on machine learning and computer equipment.
Background
The GNSS/INS integrated navigation technology is a navigation positioning technology widely applied at present, and is particularly applied to the fields of vehicle navigation and automatic driving. In an open environment, the GNSS can provide continuous high-precision navigation positioning information, and once entering a shielded area, the navigation performance is sharply reduced; the inertial navigation can provide short-time high-precision positioning information, and effectively makes up the vehicle navigation positioning performance of the GNSS unlocking state. However, when the satellite is unlocked for a long time, such as entering a long tunnel, a long and narrow tree hole, and an urban canyon region, the inertial navigation positioning error is accumulated over time, and when the satellite enters an open area again, because the positioning error is too large, the GNSS cannot perform normal satellite search, which results in positioning failure, the GNSS/INS integrated navigation system still cannot solve the problem of poor positioning accuracy of the unlocked satellite for a long time.
The Vehicle Dynamics Model (VDM) is a method commonly used for assisting an inertial navigation system in recent years, namely, partial motion parameters such as speed and position information are calculated by utilizing a vehicle mass self dynamics model and input quantity, the information is combined with output information of the inertial navigation system, and the inertial navigation positioning precision is improved by utilizing a Kalman filtering method. Although the vehicle dynamics model improves the positioning performance of the inertial navigation system to a certain extent, the method still has the problem of long-time positioning error divergence. In recent years, with the rapid development of artificial intelligence, an intelligent algorithm is gradually applied from a theory, and by utilizing the strong self-learning and predicting capabilities of the intelligent algorithm, the missing information of the GNSS can be made up, and the virtual observed value of the lock losing position of the GNSS can be obtained. The constraint conditions in the vehicle dynamics model can be used for simplifying parameters, the function of the simple navigation model is realized, and conditions are opened up for the real-time application of the intelligent algorithm. The vehicle dynamics model and the intelligent algorithm are assisted by double, so that the problem of continuous high-precision navigation and positioning of the GNSS/INS integrated navigation system in a long-time out-of-lock state can be solved.
Disclosure of Invention
In view of the above, it is necessary to provide a vehicle navigation method, device and computer device based on machine learning to solve the above technical problems.
In a first aspect, a machine learning-based vehicle navigation method is provided, the method comprising:
(1) establishing a VDM-assisted IMU error correction model;
(2) establishing a BDS/IMU/VDM combined navigation model;
(3) when the BDS is in a normal navigation state, establishing a machine learning training model to obtain a trained machine learning model; the position and the speed obtained by mechanical arrangement and VDM correction of the IMU are used as training input samples, and BDS/IMU/VDM combined filtering output is used as output samples;
(4) and when the BDS is in an unlocking state, predicting a filtering output value based on the trained machine learning model, and correcting the position and speed information solved by the IMU/VDM.
In one embodiment, the VDM-assisted IMU error correction model includes:
using a state equation of 15-dimensional state parameters:
in the formula: delta rN,δrE,δrD,δvN,δvE,δvD,δψN,δψE,δψDErrors of the position, the speed and the posture in the north N direction, the east E direction and the ground D direction are shown;zero offset of the accelerometer in the x direction, the y direction and the z direction; epsilonbx,εby,εbzZero offset of the gyroscope in the x direction, the y direction and the z direction;
the corresponding discretization expression of the state equation at the k-th moment is as follows:
xk=Φk,k+1xk-1+wk,wk~N(0,Qk)
in which N is wkObey (0, Q)k) Special sign of distribution,. phik,k+1Is a state transition matrix of error quantities, QkA noise variance matrix which is an equation of state;
the vehicle dynamics model provides the vehicle position and the speed in the direction of travel, and when the tire is not slipping, the relationship between the speed in the direction perpendicular to the direction of travel of the vehicle and the noise is represented by a non-integrity constrained model as:
in the formula, Vby,VbzFor the velocity component of the vehicle perpendicular to the direction of travel of the vehicle in the carrier coordinate system, vy,vzAnd if the white gaussian noise is corresponding to the white gaussian noise, the velocity observed quantity obtained according to the vehicle dynamics model is represented as:
Vbxis the speed component in the X direction under the carrier coordinate system;a direction cosine matrix from the carrier system b to the navigation system n;
the observations of the dynamics-assisted IMU error correction model are represented as follows:
in the formula, PINSAnd VINSRespectively representing position and velocity information, P, calculated by the IMUVDM,VVDMRespectively representing the position and velocity information calculated by the VDM;
the discretization expression of the corresponding observation equation is as follows:
zk=Hkzk+vk
in the formula, HkFor observing the coefficient matrix, vkTo observe the noise figure.
In one embodiment, the method further comprises:
constructing Kalman filtering based on a state equation and an observation equation, wherein the prediction process is expressed as follows:
in the formula (I), the compound is shown in the specification,representing the state prior estimate at time k,representing the state estimate at time k-1,andare respectively asAndvariance matrix of phik,k+1State transition matrix, Q, being the amount of error at time kk-1A noise variance matrix which is a state equation at the moment of k-1; the corresponding measurement update procedure is represented as:
wherein, KkFor filtering the gain matrix, RkFor measuring noise covariance matrix, HkAnd obtaining the optimal estimated value of the error state for observing the coefficient matrix through the Kalman filtering process, and correcting various state parameters of the IMU.
In one embodiment, the BDS/IMU/VDM combined navigation model comprises:
the BDS observed quantity is introduced into an observation equation, and the expression is as follows:
of formula (II) to (III)'INSAnd V'INSRespectively representing position and velocity information, P, acquired by the IMU after vehicle dynamics correctionBDSAnd VBDSRespectively representing the position and speed information acquired by the BDS, and the discretization expression of the corresponding observation equation is as follows:
z′k=Hkxk+vk
in the formula, HkFor observing the coefficient matrix, vkTo observe the noise coefficient, z'kA position velocity error matrix for introducing BDS observed quantity;
the corresponding discretization expression of the state equation is as follows:
xk=Φk,k+1xk-1+wk,wk~N(0,Qk)
in which N is wkObey (0, Q)k) Special sign of distribution,. phik,k+1Is a state transition matrix of error quantities, QkA noise variance matrix which is an equation of state;
and constructing Kalman filtering by introducing an observation equation and a state equation of the BDS observed quantity, and acquiring updated position, speed and attitude information.
In one embodiment, the machine learning training model established in step (3) includes:
(301) determining training sample data; the input data of the model training is position and speed data (P) obtained by navigation solution of IMUINS,VINS) (ii) a The output data of the model training is BDS and IMU combined navigation filtering output data;
(302) randomly grouping a sample training set, dividing the sample training set into P sub-training sets, and training for T times; p is the number of samples, which is less than the total number of samples; t is the maximum training frequency;
(303) carrying out normalization processing on the sample data;
(304) initializing a weight value;
(305) inputting the p sample data, and calculating the output and reverse transmission error of the current layer;
(306) if P < P, P is P +1, go to step 305, otherwise go to step 307;
(307) adjusting the weight according to the back transmission error, and adjusting the connection weight of each layer;
(308) and calculating the output, the reverse transmission error and the network total error E (T) of each layer according to the adjusted connection weight, if E (T) < epsilon or T > T, terminating the training process, otherwise, executing the step 305 and performing a new round of training.
In one embodiment, the structure of the machine learning training model is a single layer neural network.
In one embodiment, when the BDS is in the unlocked state, the step (3) of training the mature machine learning model prediction filter output values is initiated, and includes:
determining input values of a predictive model as position and velocity data (P) of the IMU as a function of navigational resolutionINS,VINS);
Inputting the input value of the prediction model into the established training model;
outputting a prediction filtering value of BDS and IMU combined navigation from the training model;
the IMU/VDM calculated position and velocity information is corrected using the predicted filtered value.
In a second aspect, a vehicle navigation device based on machine learning is provided, the device comprising:
the device comprises a first establishing module, a second establishing module and a third establishing module, wherein the first establishing module is used for establishing a VDM (virtual machine model) -assisted IMU (inertial measurement unit) error correction model;
the second establishing module is used for establishing a BDS/IMU/VDM combined navigation model;
the third establishing module is used for establishing a machine learning training model when the BDS is in a normal navigation state to obtain the trained machine learning model; the position and the speed obtained by mechanical arrangement and VDM correction of the IMU are used as training input samples, and BDS/IMU/VDM combined filtering output is used as output samples;
and the correction module is used for correcting the position and speed information resolved by the IMU/VDM based on the trained machine learning model prediction filtering output value when the BDS is in the unlocking state.
In a third aspect, a computer device is provided, which includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the machine learning-based vehicle navigation method according to any one of the first aspect when executing the computer program.
In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the machine learning-based vehicle navigation method according to any one of the first aspect.
According to the vehicle navigation method based on machine learning, a VDM-assisted IMU error correction model is established, a BDS/IMU/VDM combined navigation model is established, a machine learning training model is established when the BDS is in a normal navigation state to obtain a trained machine learning model, and when the BDS is in an unlocked state, a filtering output value is predicted based on the trained machine learning model to correct position and speed information resolved by the IMU/VDM. By testing in a tunnel, an underground garage and a satellite signal shielding area, the simplified machine learning network provided by the invention can be used for establishing auxiliary navigation information, realizing real-time auxiliary inertial navigation/wheel speed updating, avoiding long-time positioning error divergence and greatly improving the navigation positioning precision of the system.
Drawings
FIG. 1 is a diagram of an exemplary application environment of a vehicle navigation method based on machine learning;
FIG. 2 is a flow chart illustrating a method for vehicle navigation based on machine learning according to an embodiment;
FIG. 3 is a flow chart illustrating a method for vehicle navigation based on machine learning according to an embodiment;
FIG. 4 is a flow chart illustrating a method for vehicle navigation based on machine learning according to an embodiment;
FIG. 5 is a block diagram of a vehicle navigation device based on machine learning according to an embodiment;
FIG. 6 is a diagram illustrating an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The vehicle-mounted navigation method based on machine learning can be applied to the application environment shown in fig. 1. Referring to fig. 1, fig. 1 shows a BDS/IMU/VDM vehicle navigation algorithm framework diagram based on machine learning. The server realizes the vehicle navigation method based on machine learning based on the algorithm frame diagram.
The following describes in detail the technical solutions of the present application and how the technical solutions of the present application solve the above technical problems by embodiments and with reference to the drawings. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. In the car navigation method based on machine learning according to the embodiments of fig. 2 to 5 of the present application, the execution subject is the server, or may be a car navigation device based on machine learning, and the car navigation device based on machine learning may be a part or all of the server by software, hardware, or a combination of software and hardware. In the following method embodiments, the following method embodiments are all described by taking the example where the execution subject is a server.
In one embodiment, as shown in fig. 2, there is provided a machine learning-based vehicle navigation method, including the steps of:
s201, establishing a VDM-assisted IMU error correction model;
s202, establishing a BDS/IMU/VDM combined navigation model;
s203, when the BDS is in a normal navigation state, establishing a machine learning training model to obtain a trained machine learning model; the position and the speed obtained by mechanical arrangement and VDM correction of the IMU are used as training input samples, and BDS/IMU/VDM combined filtering output is used as output samples;
and S204, when the BDS is in the unlocking state, predicting a filtering output value based on the trained machine learning model, and correcting the position and speed information solved by the IMU/VDM.
According to the vehicle navigation method based on machine learning, a VDM-assisted IMU error correction model is established, a BDS/IMU/VDM combined navigation model is established, a machine learning training model is established when the BDS is in a normal navigation state to obtain a trained machine learning model, and when the BDS is in an unlocked state, a filtering output value is predicted based on the trained machine learning model to correct position and speed information resolved by the IMU/VDM. By testing in a tunnel, an underground garage and a satellite signal shielding area, the simplified machine learning network provided by the invention can be used for establishing auxiliary navigation information, realizing real-time auxiliary inertial navigation/wheel speed updating, avoiding long-time positioning error divergence and greatly improving the navigation positioning precision of the system.
In one embodiment, the VDM-assisted IMU error correction model includes:
using a state equation of 15-dimensional state parameters:
in the formula: delta rN,δrE,δrD,δvN,δvE,δvD,δψN,δψE,δψDErrors of the position, the speed and the posture in the north N direction, the east E direction and the ground D direction are shown;zero offset of the accelerometer in the x direction, the y direction and the z direction; epsilonbx,εby,εbzZero offset of the gyroscope in the x direction, the y direction and the z direction;
the corresponding discretization expression of the state equation at the k-th moment is as follows:
xk=Φk,k+1xk-1+wk,wk~N(0,Qk)
in which N is wkObey (0, Q)k) Special sign of distribution,. phik,k+1Is a state transition matrix of error quantities, QkA noise variance matrix which is an equation of state;
the vehicle dynamics model provides the vehicle position and the speed in the direction of travel, and when the tire is not slipping, the relationship between the speed in the direction perpendicular to the direction of travel of the vehicle and the noise is represented by a non-integrity constrained model as:
in the formula, Vby,VbzFor the velocity component of the vehicle perpendicular to the direction of travel of the vehicle in the carrier coordinate system, vy,vzAnd if the white gaussian noise is corresponding to the white gaussian noise, the velocity observed quantity obtained according to the vehicle dynamics model is represented as:
Vbxis the speed component in the X direction under the carrier coordinate system;a direction cosine matrix from the carrier system b to the navigation system n;
the observations of the dynamics-assisted IMU error correction model are represented as follows:
in the formula, PINSAnd VINSRespectively representing position and velocity information, P, calculated by the IMUVDM,VVDMRespectively representing the position and velocity information calculated by the VDM;
the discretization expression of the corresponding observation equation is as follows:
zk=Hkzk+vk
in the formula, HkFor observing the coefficient matrix, vkTo observe the noise figure.
In this embodiment, the simplified machine learning network of the present invention can establish auxiliary navigation information, implement real-time auxiliary inertial navigation/wheel speed update, avoid long-time positioning error divergence, and greatly improve the navigation positioning accuracy of the system.
In one embodiment, the method further comprises:
constructing Kalman filtering based on a state equation and an observation equation, wherein the prediction process is expressed as follows:
in the formula (I), the compound is shown in the specification,representing the state prior estimate at time k,representing the state estimate at time k-1,andare respectively asAndvariance matrix of phik,k+1State transition matrix, Q, being the amount of error at time kk-1A noise variance matrix which is a state equation at the moment of k-1; the corresponding measurement update procedure is represented as:
wherein, KkFor filtering the gain matrix, RkFor measuring noise covariance matrix, HkAnd obtaining the optimal estimated value of the error state for observing the coefficient matrix through the Kalman filtering process, and correcting various state parameters of the IMU.
In this embodiment, the simplified machine learning network of the present invention can establish auxiliary navigation information, implement real-time auxiliary inertial navigation/wheel speed update, avoid long-time positioning error divergence, and greatly improve the navigation positioning accuracy of the system.
In one embodiment, the BDS/IMU/VDM combined navigation model comprises:
the BDS observed quantity is introduced into an observation equation, and the expression is as follows:
of formula (II) to (III)'INSAnd V'INSRespectively representing position and velocity information, P, acquired by the IMU after vehicle dynamics correctionBDSAnd VBDSRespectively representing the position and speed information acquired by the BDS, and the discretization expression of the corresponding observation equation is as follows:
z′k=Hkxk+vk
in the formula, HkFor observing the coefficient matrix, vkTo observe the noise coefficient, z'kA position velocity error matrix for introducing BDS observed quantity;
the corresponding discretization expression of the state equation is as follows:
xk=Φk,k+1xk-1+wk,wk~N(0,Qk)
in which N is wkObey (0, Q)k) Special sign of distribution,. phik,k+1Is a state transition matrix of error quantities, QkA noise variance matrix which is an equation of state;
and constructing Kalman filtering by introducing an observation equation and a state equation of the BDS observed quantity, and acquiring updated position, speed and attitude information.
In this embodiment, the simplified machine learning network of the present invention can establish auxiliary navigation information, implement real-time auxiliary inertial navigation/wheel speed update, avoid long-time positioning error divergence, and greatly improve the navigation positioning accuracy of the system.
In one embodiment, as shown in fig. 3, the machine learning training model established in step S203 includes:
s301, determining training sample data; the input data of the model training is position and speed data (P) obtained by navigation solution of IMUINS,VINS) (ii) a The output data of the model training is BDS and IMU combined navigation filtering output data;
s302, randomly grouping a sample training set, dividing the sample training set into P sub-training sets, and training for T times; p is the number of samples, which is less than the total number of samples; t is the maximum training frequency;
s303, normalizing the sample data;
s304, initializing a weight w, and enabling t to be 1; p is 1;
s305, inputting the p sample data, and calculating the output and reverse transmission error of the current layer;
s306, if P < P, making P equal to P +1, and executing step S305, otherwise, executing step S307;
s307, adjusting the weight according to the reverse transmission error, and adjusting the connection weight of each layer;
s308, calculating the output and the back transmission error of each layer and the total network error E (t) according to the adjusted connection weight;
s309, if E (T) < epsilon or T > T, executing the step S310; otherwise, let t be t +1, and execute returning to step S305 to perform a new round of training;
and S310, terminating the training process and ending.
In this embodiment, the simplified machine learning network of the present invention can establish auxiliary navigation information, implement real-time auxiliary inertial navigation/wheel speed update, avoid long-time positioning error divergence, and greatly improve the navigation positioning accuracy of the system.
In one embodiment, the structure of the machine learning training model is a single layer neural network.
In one embodiment, as shown in fig. 4, when the BDS is in the unlocked state, the starting step (3) of training the mature machine learning model prediction filter output values includes:
s401, determining the input value of the prediction model as position and speed data (P) obtained by navigation solution of the IMUINS,VINS);
S402, inputting an input value of the prediction model into the established training model;
s403, outputting a prediction filtering value of BDS and IMU combined navigation from the training model;
s404, correcting the position and speed information calculated by the IMU/VDM by using the predicted filtering value.
In this embodiment, the simplified machine learning network of the present invention can establish auxiliary navigation information, implement real-time auxiliary inertial navigation/wheel speed update, avoid long-time positioning error divergence, and greatly improve the navigation positioning accuracy of the system.
It should be understood that although the various steps in the flow charts of fig. 2-4 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2-4 may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed in turn or alternately with other steps or at least some of the other steps.
In one embodiment, as shown in fig. 5, there is provided a machine learning-based in-vehicle navigation apparatus including: a first establishing module 01, a second establishing module 02, a third establishing module 03 and a correcting module 04, wherein:
the device comprises a first establishing module, a second establishing module and a third establishing module, wherein the first establishing module is used for establishing a VDM (virtual machine model) -assisted IMU (inertial measurement unit) error correction model;
the second establishing module is used for establishing a BDS/IMU/VDM combined navigation model;
the third establishing module is used for establishing a machine learning training model when the BDS is in a normal navigation state to obtain the trained machine learning model; the position and the speed obtained by mechanical arrangement and VDM correction of the IMU are used as training input samples, and BDS/IMU/VDM combined filtering output is used as output samples;
and the correction module is used for correcting the position and speed information resolved by the IMU/VDM based on the trained machine learning model prediction filtering output value when the BDS is in the unlocking state.
For specific definition of the vehicle-mounted navigation device based on machine learning, the above definition of the vehicle-mounted navigation method based on machine learning can be referred to, and details are not repeated here. The modules in the vehicle navigation device based on machine learning can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in fig. 6. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement a machine learning based vehicle navigation method. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
Those skilled in the art will appreciate that the architecture shown in fig. 6 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program:
(1) establishing a VDM-assisted IMU error correction model;
(2) establishing a BDS/IMU/VDM combined navigation model;
(3) when the BDS is in a normal navigation state, establishing a machine learning training model to obtain a trained machine learning model; the position and the speed obtained by mechanical arrangement and VDM correction of the IMU are used as training input samples, and BDS/IMU/VDM combined filtering output is used as output samples;
(4) and when the BDS is in an unlocking state, predicting a filtering output value based on the trained machine learning model, and correcting the position and speed information solved by the IMU/VDM.
The implementation principle and technical effect of the computer device provided by the above embodiment are similar to those of the above method embodiment, and are not described herein again.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:
(1) establishing a VDM-assisted IMU error correction model;
(2) establishing a BDS/IMU/VDM combined navigation model;
(3) when the BDS is in a normal navigation state, establishing a machine learning training model to obtain a trained machine learning model; the position and the speed obtained by mechanical arrangement and VDM correction of the IMU are used as training input samples, and BDS/IMU/VDM combined filtering output is used as output samples;
(4) and when the BDS is in an unlocking state, predicting a filtering output value based on the trained machine learning model, and correcting the position and speed information solved by the IMU/VDM.
The implementation principle and technical effect of the computer-readable storage medium provided by the above embodiments are similar to those of the above method embodiments, and are not described herein again.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A vehicle navigation method based on machine learning is characterized by comprising the following steps:
(1) establishing a VDM-assisted IMU error correction model;
(2) establishing a BDS/IMU/VDM combined navigation model;
(3) when the BDS is in a normal navigation state, establishing a machine learning training model to obtain a trained machine learning model; the position and the speed obtained by mechanical arrangement and VDM correction of the IMU are used as training input samples, and BDS/IMU/VDM combined filtering output is used as output samples;
(4) and when the BDS is in an unlocking state, correcting the position and speed information resolved by the IMU/VDM based on the trained machine learning model prediction filtering output value.
2. The method of claim 1, wherein the VDM-assisted IMU error correction model comprises:
using a state equation of 15-dimensional state parameters:
in the formula: delta rN,δrE,δrD,δvN,δvE,δvD,δψN,δψE,δψDErrors of position, speed and posture in three directions are obtained;zero bias for the accelerometer in three directions; epsilonbx,εby,εbzZero bias for the gyroscope in three directions;
the corresponding discretization expression of the state equation at the k-th moment is as follows:
xk=Φk,k+1xk-1+wk,wk~N(0,Qk)
in which N is wkObey (0, Q)k) Special sign of distribution,. phik,k+1Is a state transition matrix of error quantities, QkA noise variance matrix which is an equation of state;
the vehicle dynamics model provides the vehicle position and the speed in the direction of travel, and when the tire is not slipping, the relationship between the speed in the direction perpendicular to the direction of travel of the vehicle and the noise is represented by a non-integrity constrained model as:
in the formula, Vby,VbzFor the velocity component of the vehicle perpendicular to the direction of travel of the vehicle in the carrier coordinate system, vy,vzAnd if the white gaussian noise is corresponding to the white gaussian noise, the velocity observed quantity obtained according to the vehicle dynamics model is represented as:
Vbxis the speed component in the X direction under the carrier coordinate system;a direction cosine matrix from the carrier system b to the navigation system n;
the observations of the dynamics-assisted IMU error correction model are represented as follows:
in the formula, PINSAnd VINSRespectively representing position and velocity information, P, calculated by the IMUVDM,VVDMRespectively representing the position and velocity information calculated by the VDM;
the discretization expression of the corresponding observation equation is as follows:
zk=Hkzk+vk
in the formula, HkFor observing the coefficient matrix, vkTo observe the noise figure.
3. The method of claim 2, further comprising:
constructing Kalman filtering based on the state equation and the observation equation, wherein the prediction process is expressed as follows:
in the formula (I), the compound is shown in the specification,representing the state prior estimate at time k,representing the state estimate at time k-1,andare respectively asAndvariance matrix of phik,k+1State transition matrix, Q, being the amount of error at time kk-1A noise variance matrix which is a state equation at the moment of k-1; the corresponding measurement update procedure is represented as:
wherein, KkFor filtering the gain matrix, RkFor measuring noise covariance matrix, HkAnd obtaining the optimal estimated value of the error state for observing the coefficient matrix through the Kalman filtering process, and correcting various state parameters of the IMU.
4. The method of claim 1, wherein the BDS/IMU/VDM combined navigation model comprises:
the BDS observed quantity is introduced into an observation equation, and the expression is as follows:
of formula (II) to (III)'INSAnd V'INSRespectively representing position and velocity information, P, acquired by the IMU after vehicle dynamics correctionBDSAnd VBDSRespectively representing the position and speed information acquired by the BDS, and the discretization expression of the corresponding observation equation is as follows:
z′k=Hkxk+vk
in the formula, HkFor observing the coefficient matrix, vkTo observe the noise coefficient, z'kA position velocity error matrix for introducing BDS observed quantity;
the corresponding discretization expression of the state equation is as follows:
xk=Φk,k+1xk-1+wk,wk~N(0,Qk)
in which N is wkObey (0, Q)k) Special sign of distribution,. phik,k+1Is a state transition matrix of error quantities, QkA noise variance matrix which is an equation of state;
and constructing Kalman filtering by the observation equation introducing the BDS observed quantity and the state equation, and acquiring updated position, speed and attitude information.
5. The method according to any one of claims 1 to 4, wherein the machine learning training model established in the step (3) comprises:
(301) determining training sample data; the input data of the model training is position and speed data (P) obtained by navigation solution of IMUINS,VINS) (ii) a The output data of the model training is BDS and IMU combined navigation filtering output data;
(302) randomly grouping a sample training set, dividing the sample training set into P sub-training sets, and training for T times; p is the number of samples, which is less than the total number of samples; t is the maximum training frequency;
(303) carrying out normalization processing on the sample data;
(304) initializing a weight value;
(305) inputting the p sample data, and calculating the output and reverse transmission error of the current layer;
(306) if P < P, then P is P +1, perform step 305, otherwise perform step 307;
(307) adjusting the weight according to the back transmission error, and adjusting the connection weight of each layer;
(308) and calculating the output, the reverse transmission error and the network total error E (T) of each layer according to the adjusted connection weight, if E (T) < epsilon or T > T, terminating the training process, otherwise, executing the step 305 and performing a new round of training.
6. The method of claim 5, wherein the structure of the machine learning training model is a single layer neural network.
7. The method of claim 1, wherein the initiating step (3) of training the mature machine learning model prediction filter output values when the BDS is in the out-of-lock state comprises:
determining input values of a predictive model as position and velocity data (P) of the IMU as a function of navigational resolutionINS,VINS);
Inputting the input value of the prediction model into the established training model;
outputting a prediction filtering value of BDS and IMU combined navigation from the training model;
and correcting the IMU/VDM calculated position and velocity information using the predicted filtered value.
8. A vehicle-mounted navigation device based on machine learning, the device comprising:
the device comprises a first establishing module, a second establishing module and a third establishing module, wherein the first establishing module is used for establishing a VDM (virtual machine model) -assisted IMU (inertial measurement unit) error correction model;
the second establishing module is used for establishing a BDS/IMU/VDM combined navigation model;
the third establishing module is used for establishing a machine learning training model when the BDS is in a normal navigation state to obtain the trained machine learning model; the position and the speed obtained by mechanical arrangement and VDM correction of the IMU are used as training input samples, and BDS/IMU/VDM combined filtering output is used as output samples;
and the correction module is used for correcting the position and speed information resolved by the IMU/VDM based on the trained machine learning model prediction filtering output value when the BDS is in the unlocking state.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method of any of claims 1 to 7.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011339096.6A CN112595313A (en) | 2020-11-25 | 2020-11-25 | Vehicle-mounted navigation method and device based on machine learning and computer equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011339096.6A CN112595313A (en) | 2020-11-25 | 2020-11-25 | Vehicle-mounted navigation method and device based on machine learning and computer equipment |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112595313A true CN112595313A (en) | 2021-04-02 |
Family
ID=75184198
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011339096.6A Pending CN112595313A (en) | 2020-11-25 | 2020-11-25 | Vehicle-mounted navigation method and device based on machine learning and computer equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112595313A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113311851A (en) * | 2021-04-25 | 2021-08-27 | 北京控制工程研究所 | Spacecraft pursuit-escape intelligent orbit control method and device and storage medium |
CN113325704A (en) * | 2021-04-25 | 2021-08-31 | 北京控制工程研究所 | Spacecraft backlight approaching intelligent orbit control method and device and storage medium |
CN115585807A (en) * | 2022-12-12 | 2023-01-10 | 北京航天长城卫星导航科技有限公司 | GNSS/INS integrated navigation method based on machine learning |
CN116047567A (en) * | 2023-04-03 | 2023-05-02 | 长沙金维信息技术有限公司 | Deep learning assistance-based guard and inertial navigation combined positioning method and navigation method |
CN117706583A (en) * | 2023-12-29 | 2024-03-15 | 无锡物联网创新中心有限公司 | High-precision positioning method and system |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20100127343A (en) * | 2009-05-26 | 2010-12-06 | 위드로봇 주식회사 | Navigation system having enhanced reliability and method thereof |
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 |
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 |
CN110346824A (en) * | 2019-07-15 | 2019-10-18 | 广东工业大学 | A kind of automobile navigation method, system, device and readable storage medium storing program for executing |
-
2020
- 2020-11-25 CN CN202011339096.6A patent/CN112595313A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20100127343A (en) * | 2009-05-26 | 2010-12-06 | 위드로봇 주식회사 | Navigation system having enhanced reliability and method thereof |
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 |
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 |
CN110346824A (en) * | 2019-07-15 | 2019-10-18 | 广东工业大学 | A kind of automobile navigation method, system, device and readable storage medium storing program for executing |
Non-Patent Citations (3)
Title |
---|
XIUFENG HE 等: "MEMS IMU and two-antenna GPS integration navigation system using interval adaptive Kalman filter", 《IEEE AEROSPACE AND ELECTRONIC SYSTEMS MAGAZINE》, vol. 28, no. 10, pages 22 - 28, XP011531222, DOI: 10.1109/MAES.2013.6642828 * |
李烨;郭建国;赵斌;尤雨骅;卢晓东;周军;: "飞行器动力学信息辅助MEMS惯导***", ***工程与电子技术, no. 08, pages 1880 - 1885 * |
高为广 等: "神经网络辅助的GPS/INS组合导航故障检测算法", 《测绘学报》, vol. 37, no. 4, pages 403 - 409 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113311851A (en) * | 2021-04-25 | 2021-08-27 | 北京控制工程研究所 | Spacecraft pursuit-escape intelligent orbit control method and device and storage medium |
CN113325704A (en) * | 2021-04-25 | 2021-08-31 | 北京控制工程研究所 | Spacecraft backlight approaching intelligent orbit control method and device and storage medium |
CN113325704B (en) * | 2021-04-25 | 2023-11-10 | 北京控制工程研究所 | Spacecraft backlighting approaching intelligent orbit control method, device and storage medium |
CN115585807A (en) * | 2022-12-12 | 2023-01-10 | 北京航天长城卫星导航科技有限公司 | GNSS/INS integrated navigation method based on machine learning |
CN115585807B (en) * | 2022-12-12 | 2023-08-15 | 北京航天长城卫星导航科技有限公司 | GNSS/INS integrated navigation method based on machine learning |
CN116047567A (en) * | 2023-04-03 | 2023-05-02 | 长沙金维信息技术有限公司 | Deep learning assistance-based guard and inertial navigation combined positioning method and navigation method |
CN117706583A (en) * | 2023-12-29 | 2024-03-15 | 无锡物联网创新中心有限公司 | High-precision positioning method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112595313A (en) | Vehicle-mounted navigation method and device based on machine learning and computer equipment | |
CN111527382B (en) | System and method for controlling vehicle | |
CN112577521B (en) | Combined navigation error calibration method and electronic equipment | |
CN112798021B (en) | Inertial navigation system inter-travelling initial alignment method based on laser Doppler velocimeter | |
CN113465628B (en) | Inertial measurement unit data compensation method and system | |
CN111562603B (en) | Navigation positioning method, equipment and storage medium based on dead reckoning | |
CN114689047B (en) | Deep learning-based integrated navigation method, device, system and storage medium | |
CN111024124B (en) | Combined navigation fault diagnosis method for multi-sensor information fusion | |
CN110346821B (en) | SINS/GPS combined attitude-determining and positioning method and system for solving long-time GPS unlocking problem | |
US20230366680A1 (en) | Initialization method, device, medium and electronic equipment of integrated navigation system | |
CN113252048B (en) | Navigation positioning method, navigation positioning system and computer readable storage medium | |
EP4345421A2 (en) | Method for calibrating sensor parameters based on autonomous driving, apparatus, storage medium, and vehicle | |
CN114061570A (en) | Vehicle positioning method and device, computer equipment and storage medium | |
CN115143954A (en) | Unmanned vehicle navigation method based on multi-source information fusion | |
CN113341443A (en) | Processing method of positioning track information and vehicle-mounted navigation device | |
CN114935345A (en) | Vehicle-mounted inertial navigation mounting angle error compensation method based on pattern recognition | |
CN113566850B (en) | Method and device for calibrating installation angle of inertial measurement unit and computer equipment | |
CN109655057B (en) | Filtering optimization method and system for accelerator measurement value of six-push unmanned aerial vehicle | |
CN116224407A (en) | GNSS and INS integrated navigation positioning method and system | |
CN114019954B (en) | Course installation angle calibration method, device, computer equipment and storage medium | |
CN115950447A (en) | High-precision alignment method and system for underwater movable base based on magnetic compass and velocimeter | |
CN108957508B (en) | Vehicle-mounted POS (point of sale) offline combined estimation method and device | |
CN113137975B (en) | Inertial correction method and device for astronomical inertial integrated navigation and electronic equipment | |
CN115164886B (en) | Scale factor error compensation method of vehicle-mounted GNSS/INS integrated navigation system | |
CN113566849B (en) | Method and device for calibrating installation angle of inertial measurement unit and computer equipment |
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 |