CN113124884A - Vehicle positioning method and device based on LSTM neural network model - Google Patents

Vehicle positioning method and device based on LSTM neural network model Download PDF

Info

Publication number
CN113124884A
CN113124884A CN202110411903.9A CN202110411903A CN113124884A CN 113124884 A CN113124884 A CN 113124884A CN 202110411903 A CN202110411903 A CN 202110411903A CN 113124884 A CN113124884 A CN 113124884A
Authority
CN
China
Prior art keywords
sequence data
data
angular velocity
module
acceleration
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110411903.9A
Other languages
Chinese (zh)
Other versions
CN113124884B (en
Inventor
费再慧
李晓宵
单国航
朱磊
贾双成
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhidao Network Technology Beijing Co Ltd
Original Assignee
Zhidao Network Technology Beijing Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Zhidao Network Technology Beijing Co Ltd filed Critical Zhidao Network Technology Beijing Co Ltd
Priority to CN202110411903.9A priority Critical patent/CN113124884B/en
Publication of CN113124884A publication Critical patent/CN113124884A/en
Application granted granted Critical
Publication of CN113124884B publication Critical patent/CN113124884B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
    • G01C21/28—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network with correlation of data from several navigational instruments
    • 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
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Abstract

The application relates to a vehicle positioning method and device based on an LSTM neural network model. The method comprises the following steps: inputting first acceleration sequence data and first angular velocity sequence data of a positioning module and second acceleration sequence data and second angular velocity sequence data of an inertial measurement unit into a preset LSTM neural network model; converging the second acceleration sequence data to the first acceleration sequence data, and converging the second angular velocity sequence data to the first angular velocity sequence data to obtain a trained data model; inputting the measurement data of the inertia measurement unit into the trained data model; enabling the data model to carry out time correction on the measurement data of the inertial measurement unit and output corrected measurement data; and correcting the measurement data according to the initial pose of the vehicle to obtain the positioning data of the vehicle. The technical scheme provided by the application can reduce the accumulated error of the inertia measurement unit based on the LSTM neural network model, and improve the precision of vehicle positioning.

Description

Vehicle positioning method and device based on LSTM neural network model
Technical Field
The application relates to the technical field of navigation, in particular to a vehicle positioning method and device based on an LSTM neural network model.
Background
The satellite Positioning module, such as a GPS (Global Positioning System) Positioning module, has the characteristics of good performance, high accuracy and wide application. However, in some situations, such as under a bridge, a culvert, a tunnel, and a dense building, the satellite positioning module in the related art has a large positioning deviation and cannot provide a positioning result. The Inertial navigation system including an Inertial Measurement Unit (IMU) can calculate the motion trajectory of the carrier by using the Measurement data of the Inertial Measurement Unit.
The inertial navigation system uses the measurement data of an accelerometer and a gyroscope of the inertial measurement unit to calculate the motion trail of the carrier. However, the accelerometer and the gyroscope in the inertial measurement unit are affected by various factors, and after the inertial measurement unit is used for a period of time, internal parameters and performance of the inertial measurement unit change, so that measured data have deviation, and large errors are accumulated along with the lapse of time, so that a motion track obtained by the inertial navigation system through calculation of the measured data of the inertial measurement unit has large errors.
Disclosure of Invention
In order to solve the problems in the related art, the application provides a vehicle positioning method and device based on an LSTM neural network model, which can reduce the accumulated error of an inertial measurement unit based on the LSTM neural network model and improve the precision of vehicle positioning according to the measurement data of the inertial measurement unit.
The application provides a vehicle positioning method based on an LSTM neural network model in a first aspect, and the method comprises the following steps:
respectively carrying out spline fitting processing to construct a first spline curve group according to the position information and the posture information of the positioning module, wherein the first spline curve group comprises a first position spline curve and a first posture spline curve;
according to the first spline curve group, obtaining first acceleration sequence data of the positioning module and first angular speed sequence data of the positioning module;
inputting the first acceleration sequence data and the first angular velocity sequence data, and second acceleration sequence data and second angular velocity sequence data of an inertial measurement unit to a preset LSTM neural network model, wherein the first acceleration sequence data and the second acceleration sequence data are aligned in sampling time, and the first angular velocity sequence data and the second angular velocity sequence data are aligned in sampling time;
converging the second acceleration sequence data to the first acceleration sequence data, and converging the second angular velocity sequence data to the first angular velocity sequence data to obtain a trained data model;
inputting the measurement data of an inertia measurement unit into the trained data model;
enabling the trained data model to carry out time correction on the measurement data of the inertial measurement unit and output corrected measurement data;
and obtaining positioning data of the vehicle according to the initial pose of the vehicle and the corrected measurement data.
Preferably, the spline fitting processing is respectively performed according to the position information and the posture information of the positioning module to construct a first spline curve group, where the first spline curve group includes a first position spline curve and a first posture spline curve, and the method includes:
b spline fitting processing is carried out according to the position information of the positioning module to construct the first position spline curve;
and B-spline fitting is carried out according to the attitude information of the positioning module to construct the first attitude spline curve.
Preferably, the obtaining the first acceleration sequence data of the positioning module and the first angular velocity sequence data of the positioning module according to the first spline set includes:
and performing second-order derivation on the first position spline curve to obtain the first acceleration sequence data, and performing first-order derivation on the first posture spline curve to obtain the first angular velocity sequence data.
Preferably, the obtaining a trained data model by converging the second acceleration sequence data to the first acceleration sequence data and converging the second angular velocity sequence data to the first angular velocity sequence data includes:
calculating to obtain mean square errors of the second acceleration sequence data and the first acceleration sequence data, and the second angular velocity sequence data and the first angular velocity sequence data by adopting a mean square error algorithm in a preset time step;
and obtaining a trained data model by enabling the mean square error convergence to be smaller than a preset threshold value.
Preferably, the value of the time step is a multiple of the output frequency of the positioning module.
A second aspect of the present application provides a vehicle localization apparatus based on an LSTM neural network model, the apparatus comprising:
the curve construction module is used for respectively carrying out spline fitting processing according to the position information and the posture information of the positioning module to construct a first spline curve group, and the first spline curve group comprises a first position spline curve and a first posture spline curve;
the sequence data acquisition module is used for acquiring first acceleration sequence data of the positioning module and first angular velocity sequence data of the positioning module according to the first curve group constructed by the curve construction module;
a first input module, configured to input, to a preset LSTM neural network model, the first acceleration sequence data and the first angular velocity sequence data obtained by the sequence data acquisition module, and second acceleration sequence data and second angular velocity sequence data of an inertial measurement unit, where the first acceleration sequence data and the second acceleration sequence data are aligned at a sampling time, and the first angular velocity sequence data and the second angular velocity sequence data are aligned at the sampling time;
a training module, configured to converge second acceleration sequence data input by the first input module to first acceleration sequence data input by the first input module, and converge second angular velocity sequence data input by the first input module to first angular velocity sequence data input by the first input module, so as to obtain a trained data model;
the second input module is used for inputting the measurement data of the inertial measurement unit into the trained data model;
the correction data acquisition module is used for enabling the trained data model to carry out time correction on the measurement data of the inertial measurement unit input by the second input module and output correction measurement data;
and the positioning module is used for obtaining positioning data of the vehicle according to the initial pose of the vehicle and the corrected measurement data.
Preferably, the curve construction module is specifically configured to:
b spline fitting processing is carried out according to the position information of the positioning module to construct the first position spline curve;
and B-spline fitting is carried out according to the attitude information of the positioning module to construct the first attitude spline curve.
Preferably, the sequence data acquiring module is specifically configured to perform second-order derivation on the first position spline curve constructed by the curve constructing module to obtain the first acceleration sequence data, and perform first-order derivation on the first posture spline curve constructed by the curve constructing module to obtain the first angular velocity sequence data.
Preferably, the training module is further configured to:
calculating by using a mean square error algorithm with a preset time step to obtain mean square errors of second acceleration sequence data and first acceleration sequence data input by the first input module and second angular velocity sequence data and first angular velocity sequence data input by the first input module;
and obtaining the trained data model by enabling the mean square error convergence to be smaller than a preset threshold value.
A third aspect of the present application provides an electronic device comprising:
a processor; and
a memory having executable code stored thereon, which when executed by the processor, causes the processor to perform the method as described above.
A fourth aspect of the present application provides a non-transitory machine-readable storage medium having stored thereon executable code, which when executed by a processor of an electronic device, causes the processor to perform a method as described above.
The technical scheme provided by the application can comprise the following beneficial effects:
according to the technical scheme, the first sample curve group is constructed by spline fitting processing of the positioning information of the positioning module, derivation is conducted on the first sample curve group, and first acceleration sequence data and first angular velocity sequence data of the positioning module are obtained. And aligning the positioning information of the low-frequency positioning module with the measurement data of the high-frequency inertial measurement unit in a spline curve mode through spline fitting processing. Compared with the related art, which takes position and velocity as input samples, the embodiment of the application takes the first acceleration sequence data and the first angular velocity sequence data of the positioning module, and the second acceleration series data and the second angular velocity series data of the inertial measurement unit are samples, taking the first acceleration sequence data and the first angular speed sequence data of the positioning module as output expected values, obtaining a trained data model by converging the second acceleration sequence data toward the first acceleration sequence data and converging the second angular velocity sequence data toward the first angular velocity sequence data, under the condition of the positioning information of the positioning modules with the same quantity, the generalization effect is improved, the efficiency of data model training can be improved, the training precision of the data model is improved, the accumulative error of the inertia measurement unit is better reduced, and the precision of vehicle positioning according to the measurement data of the inertia measurement unit is improved. The trained data model based on the LSTM neural network model is used for carrying out time correction on the measurement data input into the inertial measurement unit and outputting corrected measurement data; and acquiring positioning data of the vehicle according to the initial pose of the vehicle and the corrected measurement data, so that the accumulated error of the inertial measurement unit can be reduced based on the LSTM neural network model, and the accuracy of positioning the vehicle according to the measurement data of the inertial measurement unit is improved.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the application.
Drawings
The foregoing and other objects, features and advantages of the application will be apparent from the following more particular descriptions of exemplary embodiments of the application, as illustrated in the accompanying drawings wherein like reference numbers generally represent like parts throughout the exemplary embodiments of the application.
FIG. 1 is a schematic flow chart diagram illustrating a method for vehicle localization based on an LSTM neural network model according to an embodiment of the present application;
FIG. 2 is a schematic structural diagram of a vehicle locating device based on an LSTM neural network model according to an embodiment of the present application;
fig. 3 is a schematic structural diagram of an electronic device shown in an embodiment of the present application.
Detailed Description
Preferred embodiments of the present application will be described in more detail below with reference to the accompanying drawings. While the preferred embodiments of the present application are shown in the drawings, it should be understood that the present application may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in this application and the appended claims, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.
It should be understood that although the terms "first," "second," "third," etc. may be used herein to describe various information, these information should not be limited to these terms. These terms are only used to distinguish one type of information from another. For example, first information may also be referred to as second information, and similarly, second information may also be referred to as first information, without departing from the scope of the present application. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the present application, "a plurality" means two or more unless specifically limited otherwise.
The embodiment of the application provides a vehicle positioning method based on an LSTM neural network model, which can reduce the accumulated error of the measurement data of an inertia measurement unit based on the LSTM neural network model and improve the precision of vehicle positioning according to the measurement data of the inertia measurement unit.
The technical solutions of the embodiments of the present application are described in detail below with reference to the accompanying drawings.
The first embodiment is as follows:
fig. 1 is a flowchart illustrating a vehicle localization method based on an LSTM neural network model according to an embodiment of the present application.
Referring to fig. 1, a vehicle localization method based on an LSTM neural network model includes:
in step 101, positioning information of the positioning module and measurement data of the inertial measurement unit are acquired.
In one specific implementation mode, the vehicle is provided with an inertia measurement unit and a positioning module. The inertial measurement unit comprises an accelerometer and a gyroscope, and the measurement data of the inertial measurement unit comprises the acceleration of the accelerometer of the inertial measurement unit and the angular velocity of the gyroscope. The acceleration of the vehicle, which may be obtained by an accelerometer of the inertial measurement unit, and the angular velocity of the vehicle, which may be obtained by a gyroscope of the inertial measurement unit. The positioning module may include, but is not limited to, at least one of a GPS, a beidou satellite positioning module, an RTK (Real Time Kinematic) positioning module, and the like. The positioning module may be utilized to obtain positioning information of the vehicle, which may include, but is not limited to, position information, attitude information, speed information.
In one embodiment, the positioning information of the RTK positioning module, the measurement data of the inertial measurement unit, is acquired in the event that the RTK positioning module signal is available on the vehicle. And obtaining continuous-time positioning information of the vehicle according to the continuous-time positioning signal of the RTK positioning module, wherein the positioning information can comprise position information, attitude information and speed information, the position information comprises latitude and longitude coordinate information describing a position, and the attitude information comprises but is not limited to course angle information describing a course. When the RTK positioning module is used for acquiring the positioning information of the vehicle in continuous time, the accelerometer of the inertial measurement unit is used for acquiring the acceleration of the vehicle in continuous time, and the gyroscope of the inertial measurement unit is used for acquiring the angular velocity of the vehicle in continuous time.
In step 102, spline fitting is performed to construct a first spline curve group according to the position information and the attitude information of the positioning module, wherein the first spline curve group comprises a first position spline curve and a first attitude spline curve.
In one embodiment, according to the position information of the positioning module, B-spline (B-spline) fitting processing is carried out to construct a first position spline curve; and B-spline fitting is carried out according to the attitude information of the positioning module to construct a first attitude spline curve.
In a specific embodiment, according to at least two pieces of position information of the continuous time of the vehicle, B-spline fitting processing is carried out to construct a continuous function curve between the position information and the time of the positioning module, and a first position spline curve is obtained; and B-spline fitting is carried out according to at least two pieces of attitude information of the vehicle continuous time to construct a continuous function curve between the attitude information of the positioning module and the time, and a first attitude spline curve is obtained.
It should be noted that, in the embodiment of the present application, the first position spline curve and the first pose spline curve may be constructed through B-spline fitting processing, or the first position spline curve and the first pose spline curve may also be constructed through other spline fitting algorithms, for example, Cubic spline interpolation (Cubic spline interpolation) fitting processing, which is not limited in the embodiment of the present application.
In step 103, first acceleration sequence data of the localization module in the same time period and first angular velocity sequence data of the localization module in a time period are obtained according to the first spline set.
In a specific embodiment, the second order derivation is performed on the first position spline curve of the first spline curve group to obtain a continuous function curve between the acceleration and the time of the positioning module, that is, the first acceleration spline curve of the positioning module, and the first acceleration sequence data is obtained according to the first acceleration spline curve. And performing first-order derivation on the first posture spline curve of the first spline curve group to obtain a continuous function curve between the angular speed and the time of the positioning module, namely the first angular speed spline curve of the positioning module, and obtaining first angular speed sequence data according to the first angular speed spline curve.
In step 104, first acceleration sequence data and first angular velocity sequence data, which are aligned in sampling time, and second acceleration sequence data and second angular velocity sequence data of the inertial measurement unit, which are aligned in sampling time, are input to a preset LSTM neural network model.
In one embodiment, the first acceleration sequence data, the first angular velocity sequence data, the second acceleration sequence data, and the second angular velocity sequence data aligned with the sampling time of the same time period are used as training samples and input into a preset LSTM neural network model. It is to be understood that the first acceleration series data and the second acceleration series data are aligned in time, and the first angular velocity series data and the second angular velocity series data are also aligned in time, so as to avoid inaccurate results due to the difference in reference time between the two.
The LSTM neural Network in the embodiment of the present application is a Short name of a Long Short Term Memory Network (Long Short Term Memory Network), and is a recurrent neural Network.
In step 105, the trained data model is obtained by converging the second acceleration sequence data to the first acceleration sequence data and converging the second angular velocity sequence data to the first angular velocity sequence data.
In one embodiment, the first parameters of the LSTM neural network model are obtained by converging the second acceleration sequence data toward the first acceleration sequence data.
In one specific embodiment, the method comprises the steps of acquiring the acceleration error of each same time point of first acceleration sequence data and second acceleration sequence data respectively in time sequence, analyzing the acceleration error of the first acceleration sequence data and the second acceleration sequence data of each same time point, taking the acceleration of the first acceleration sequence data as an output expected value, and losing the acceleration error of a certain time point if the acceleration error of the certain time point of the first acceleration sequence data and the second acceleration sequence data is larger than or equal to a first set error threshold value; and if the error of the acceleration of a certain time point of the first acceleration sequence data and the second acceleration sequence data is smaller than a first set error threshold value, updating a first set of parameters of the LSTM neural network model, and converging the second acceleration sequence data to the first acceleration sequence data to enable the error of the acceleration of the first acceleration sequence data and the second acceleration sequence data to be smaller than the first set threshold value, so as to obtain the first set of parameters of the LSTM neural network model.
In one embodiment, the second set of parameters of the LSTM neural network model is obtained by converging the second angular velocity sequence data toward the first angular velocity sequence data.
In one embodiment, the error of the angular velocity at each same time point of the first angular velocity sequence data and the second angular velocity sequence data is acquired in time sequence, the error of the angular velocity at each same time point of the first angular velocity sequence data and the second angular velocity sequence data is analyzed, the angular velocity at a certain time point of the first angular velocity sequence data and the second angular velocity sequence data is taken as an output expected value, and if the error of the angular velocity at the certain time point is greater than or equal to a second set error threshold value, the error of the angular velocity at the certain time point is lost; and updating a second set of parameters of the LSTM neural network model if an error of the angular velocity at a certain time point of the first angular velocity sequence data and the second angular velocity sequence data is less than a second set error threshold, and obtaining the second set of parameters of the LSTM neural network model by converging the second angular velocity sequence data to the first angular velocity sequence data so that the error of the angular velocity of the first angular velocity sequence data and the second angular velocity sequence data is less than the second set threshold.
In one embodiment, a trained data model is obtained based on the first set of parameters and the second set of parameters.
In one embodiment, a trained data model is obtained based on the LSTM neural network model based on the first set of parameters and the second set of parameters.
In one embodiment, a mean square error algorithm is adopted to calculate the mean square error of the second acceleration sequence data and the first acceleration sequence data, and the second angular velocity sequence data and the first angular velocity sequence data in a preset time step; and obtaining the trained data model by enabling the mean square error convergence to be smaller than a preset threshold value.
In one embodiment, the time step has a value that is a multiple of the output frequency of the positioning module. In some embodiments, the value of the time step may be 10.
In one embodiment, the output frequency of the RTK positioning module is 1HZ, the output frequency of the inertial measurement unit is 100HZ, and the mean square error algorithm can be used to calculate, but is not limited to, the mean square error of the second acceleration sequence data and the first acceleration sequence data, and the mean square error of the second angular velocity sequence data and the first angular velocity sequence data in 1 second time step. For example, the acceleration of the second acceleration sequence data and the first acceleration sequence data at time point 1 second, time point 2 second, …, and time point n second is obtained, the angular velocities of the second angular velocity sequence data and the first angular velocity sequence data at time point 1 second, time point 2 second, …, and time point n second are obtained, n accelerations are calculated, and the mean square error MSE of n angular velocities is:
Figure BDA0003024530210000091
in the formula, a2, i is the acceleration of the second acceleration series data time point i, a1, i is the acceleration of the first acceleration series data time point i, ω 2, i is the angular velocity of the second angular velocity series data time point i, and ω 1, i is the angular velocity of the first angular velocity series data time point i.
The LSTM neural network model can be continuously optimized by adjusting one or more parameters in the LSTM neural network model and/or adjusting the structure of the LSTM neural network model, so that the MSE value is the minimum as possible, and the convergence is smaller than a preset threshold value, namely, the second acceleration sequence data is converged to the first acceleration sequence data and the second angular velocity sequence data is converged to the first angular velocity sequence data, so that the trained data model based on the LSTM neural network model is obtained.
In step 106, the measurement data of the inertial measurement unit is input to the trained data model.
In one embodiment, in the case that the signal of the vehicle's positioning module is unavailable, the acceleration of the vehicle may be obtained by an accelerometer of the inertial measurement unit, the angular velocity of the vehicle may be obtained by a gyroscope of the inertial measurement unit, and the acceleration and the angular velocity of the vehicle are input to a trained data model based on the LSTM neural network model.
In step 107, the trained data model is time-corrected for the measurement data of the inertial measurement unit and the corrected measurement data is output.
In one embodiment, the corrected acceleration and the corrected angular velocity output by the data model are obtained by causing the trained data model to time-correct the input acceleration and angular velocity of the vehicle, respectively. The data model based on the LSTM neural network model corrects the input acceleration and angular velocity of the vehicle, and outputs a corrected acceleration and a corrected angular velocity with reduced error.
In step 108, the measurement data is corrected according to the initial pose of the vehicle, and positioning data of the vehicle is obtained.
In one embodiment, according to the positioning information at the time before the signal of the positioning module of the vehicle is unavailable (the time when the signal of the positioning module is available), the time before the signal of the positioning module of the vehicle is unavailable (the time when the signal of the positioning module is available) is taken as an initial time, and according to the positioning information at the initial time of the vehicle, the initial pose of the vehicle is the initial pose of the vehicle, and the initial pose of the vehicle comprises the position information, the course angle information and the speed information of the vehicle at the initial time. And obtaining the positioning data of the vehicle according to the initial pose of the vehicle, the corrected acceleration and the corrected angular speed which are output after the data model is corrected.
According to the vehicle positioning method based on the LSTM neural network model, the first spline curve group is constructed by spline fitting processing of positioning information of the positioning module, derivation is conducted on the first spline curve group, and first acceleration sequence data and first angular velocity sequence data of the positioning module are obtained. And aligning the positioning information of the low-frequency positioning module with the measurement data of the high-frequency inertial measurement unit in a spline curve mode through spline fitting processing. Compared with the related art, which takes position and velocity as input samples, the embodiment of the application takes the first acceleration sequence data and the first angular velocity sequence data of the positioning module, and the second acceleration series data and the second angular velocity series data of the inertial measurement unit are samples, taking the first acceleration sequence data and the first angular speed sequence data of the positioning module as output expected values, obtaining a trained data model by converging the second acceleration sequence data toward the first acceleration sequence data and converging the second angular velocity sequence data toward the first angular velocity sequence data, under the condition of the positioning information of the positioning modules with the same quantity, the generalization effect is improved, the efficiency of data model training can be improved, the training precision of the data model is improved, the accumulative error of the inertia measurement unit is better reduced, and the precision of vehicle positioning according to the measurement data of the inertia measurement unit is improved. The trained data model based on the LSTM neural network model is used for carrying out time correction on the measurement data input into the inertial measurement unit and outputting corrected measurement data; the positioning data of the vehicle is obtained by correcting the measurement data according to the initial pose of the vehicle, the accumulated error of the inertia measurement unit can be reduced based on the LSTM neural network model, and the precision of vehicle positioning according to the measurement data of the inertia measurement unit is improved.
Further, the vehicle localization method based on the LSTM neural network model shown in the embodiment of the present application adopts a mean square error algorithm to calculate a mean square error between the second acceleration sequence data and the first acceleration sequence data, and between the second angular velocity sequence data and the first angular velocity sequence data at a preset time step; and obtaining the trained data model by enabling the mean square error convergence to be smaller than a preset threshold value. The numerical value of the time step is a multiple of the output frequency of the positioning module, so that the errors of the first acceleration sequence data and the first angular velocity sequence data can be reduced, and the training precision of the data model is improved.
Example two:
corresponding to the embodiment of the application function implementation method, the application also provides a vehicle positioning device based on the LSTM neural network model, electronic equipment and a corresponding embodiment.
Fig. 2 is a schematic structural diagram of a vehicle locating device based on an LSTM neural network model according to an embodiment of the present application.
Referring to fig. 2, the vehicle positioning device based on the LSTM neural network model includes a curve building module 201, a sequence data obtaining module 202, a first input module 203, a training module 204, a second input module 205, a correction data obtaining module 206, and a positioning module 207.
And the curve construction module 201 is configured to perform spline fitting processing respectively according to the position information and the posture information of the positioning module to construct a first spline curve group, where the first spline curve group includes a first position spline curve and a first posture spline curve.
In one specific implementation mode, the vehicle is provided with an inertia measurement unit and a positioning module. The inertial measurement unit comprises an accelerometer and a gyroscope, and the measurement data of the inertial measurement unit comprises the acceleration of the accelerometer of the inertial measurement unit and the angular velocity of the gyroscope. The acceleration of the vehicle, which may be obtained by an accelerometer of the inertial measurement unit, and the angular velocity of the vehicle, which may be obtained by a gyroscope of the inertial measurement unit. The positioning module may include, but is not limited to, at least one of a GPS, a beidou satellite positioning module, an RTK (Real Time Kinematic) positioning module, and the like. The positioning module may be utilized to obtain positioning information of the vehicle, which may include, but is not limited to, position information, attitude information, speed information.
In one embodiment, in the case that the RTK positioning module signal is available on the vehicle, the curve construction module 201 obtains the positioning information of the RTK positioning module, and obtains the positioning information of the vehicle in continuous time according to the positioning signal of the RTK positioning module in continuous time, where the positioning information may include position information, attitude information, and speed information, the position information includes, but is not limited to, longitude and latitude coordinate information describing a position, and the attitude information includes, but is not limited to, heading angle information describing a heading. When the curve construction module 201 acquires the positioning information of the vehicle in continuous time through the RTK positioning module, the first input module 203 acquires the acceleration of the vehicle in continuous time through an accelerometer of the inertial measurement unit, and acquires the angular velocity of the vehicle in continuous time through a gyroscope of the inertial measurement unit.
In one embodiment, the curve construction module 201 is specifically configured to perform B-spline fitting processing to construct a first position spline curve according to the position information of the positioning module; and B-spline fitting is carried out according to the attitude information of the positioning module to construct a first attitude spline curve.
In a specific embodiment, the curve construction module 201 performs B-spline fitting processing to construct a continuous function curve between the position information of the positioning module and time according to at least two position information of the vehicle in continuous time, so as to obtain a first position sample curve; and B-spline fitting is carried out according to at least two pieces of attitude information of the vehicle continuous time to construct a continuous function curve between the attitude information of the positioning module and the time, and a first attitude spline curve is obtained.
It should be noted that, the curve constructing module 201 in the embodiment of the present application may construct the first position spline curve and the first pose spline curve through B-spline fitting processing, or may construct the first position spline curve and the first pose spline curve through fitting processing of other spline fitting algorithms, for example, Cubic spline interpolation (Cubic spline interpolation), which is not limited in the embodiment of the present application.
The sequence data obtaining module 202 is configured to obtain first acceleration sequence data of the positioning module and first angular velocity sequence data of the positioning module according to the first curve group constructed by the curve constructing module 201.
In a specific embodiment, the sequence data obtaining module 202 performs second-order derivation on the first position spline curve of the first spline curve group to obtain a continuous function curve between the acceleration and the time of the positioning module, which is the first acceleration spline curve of the positioning module, and obtains the first acceleration sequence data according to the first acceleration spline curve. The sequence data obtaining module 202 performs first-order derivation on the first posture spline curve of the first spline curve group to obtain a continuous function curve between the angular speed and time of the positioning module, that is, the first angular speed spline curve of the positioning module, and obtains first angular speed sequence data according to the first angular speed spline curve.
The first input module 203 is configured to input, to a preset LSTM neural network model, first acceleration sequence data and first angular velocity sequence data obtained by the sequence data obtaining module 202, and second acceleration sequence data and second angular velocity sequence data of an inertial measurement unit, where the first acceleration sequence data and the second acceleration sequence data are aligned at a sampling time, and the first angular velocity sequence data and the second angular velocity sequence data are aligned at the sampling time.
In one embodiment, the first input module 203 inputs the preset LSTM neural network model with the first acceleration sequence data, the first angular velocity sequence data obtained by the sampling time-aligned sequence data acquisition module 202 of the same time period, and the second acceleration sequence data and the second angular velocity sequence data of the inertial measurement unit as training samples. It is to be understood that the first acceleration series data and the second acceleration series data are aligned in time, and the first angular velocity series data and the second angular velocity series data are also aligned in time, so as to avoid inaccurate results due to the difference in reference time between the two.
A training module 204, configured to converge the second acceleration sequence data input by the first input module 203 to the first acceleration sequence data input by the first input module 203, and converge the second angular velocity sequence data input by the first input module 203 to the first angular velocity sequence data input by the first input module 203, so as to obtain a trained data model.
In one embodiment, the training module 204 obtains the first parameters of the LSTM neural network model by converging the second acceleration sequence data input by the first input module 203 to the first acceleration sequence data input by the first input module 203.
In a specific embodiment, the training module 204 respectively obtains the acceleration error of the first acceleration sequence data and the second acceleration sequence data input by the first input module 203 at each same time point in time sequence, analyzes the acceleration error of the first acceleration sequence data and the second acceleration sequence data at each same time point, takes the acceleration of the first acceleration sequence data input by the first input module 203 as an output expected value, and loses the acceleration error at the time point if the acceleration error at a certain time point of the first acceleration sequence data and the second acceleration sequence data is greater than or equal to a first set error threshold; and if the error of the acceleration of a certain time point of the first acceleration sequence data and the second acceleration sequence data is smaller than a first set error threshold value, updating a first set of parameters of the LSTM neural network model, and converging the second acceleration sequence data to the first acceleration sequence data to enable the error of the acceleration of the first acceleration sequence data and the second acceleration sequence data to be smaller than the first set threshold value, so as to obtain the first set of parameters of the LSTM neural network model.
In one embodiment, the training module 204 obtains the second set of parameters of the LSTM neural network model by converging the second angular velocity sequence data input by the first input module 203 toward the first angular velocity sequence data input by the first input module 203.
In one embodiment, the training module 204 obtains the error of the angular velocity of the first angular velocity sequence data and the second angular velocity sequence data input by the first input module 203 at each same time point in time sequence, analyzes the error of the angular velocity of the first angular velocity sequence data and the second angular velocity sequence data at each same time point, takes the angular velocity of the first angular velocity sequence data input by the first input module 203 as an output expected value, and loses the error of the angular velocity at a time point if the error of the angular velocity at the time point of the first angular velocity sequence data and the second angular velocity sequence data is greater than or equal to a second set error threshold; and updating a second set of parameters of the LSTM neural network model if an error of the angular velocity at a certain time point of the first angular velocity sequence data and the second angular velocity sequence data is less than a second set error threshold, and obtaining the second set of parameters of the LSTM neural network model by converging the second angular velocity sequence data to the first angular velocity sequence data so that the error of the angular velocity of the first angular velocity sequence data and the second angular velocity sequence data is less than the second set threshold.
In one embodiment, the training module 204 obtains a trained data model based on the first set of parameters and the second set of parameters.
In one embodiment, the training module 204 obtains a trained data model based on the LSTM neural network model based on the first set of parameters and the second set of parameters.
In one embodiment, the training module 204 is further configured to calculate, by using a mean square error algorithm, a mean square error between the second acceleration sequence data and the first acceleration sequence data, and between the second angular velocity sequence data and the first angular velocity sequence data, which are input by the first input module 203, in a preset time step; and obtaining the trained data model by enabling the mean square error convergence to be smaller than a preset threshold value.
In one embodiment, the training module 204 uses a time step value that is a multiple of the output frequency of the positioning module.
In a specific embodiment, the output frequency of the RTK positioning module is 1HZ, the output frequency of the inertial measurement unit is 100HZ, and the training module 204 may calculate, but not limited to, a mean square error of the second acceleration sequence data and the first acceleration sequence data, the second angular velocity sequence data and the first angular velocity sequence data input by the first input module 203 in 1 second time step by using a mean square error algorithm. For example, the training module 204 obtains accelerations of the second acceleration sequence data and the first acceleration sequence data from time point 1 second, time point 2 seconds, time point … to time point n seconds; obtaining the angular velocities of the second angular velocity sequence data and the first angular velocity sequence data from the time point of 1 second to the time point of 2 seconds to … to the time point of n seconds, and calculating n accelerations, wherein the mean square error MSE of the n angular velocities is as follows:
Figure BDA0003024530210000151
in the formula, a2, i is the acceleration of the second acceleration series data time point i, a1, i is the acceleration of the first acceleration series data time point i, ω 2, i is the angular velocity of the second angular velocity series data time point i, and ω 1, i is the angular velocity of the first angular velocity series data time point i.
The training module 204 may continuously optimize the LSTM neural network model by adjusting one or more parameters in the LSTM neural network model and/or adjusting a structure of the LSTM neural network model such that the MSE value is the smallest as possible and the convergence is less than a preset threshold, that is, the trained data model based on the LSTM neural network model is obtained by converging the second acceleration sequence data to the first acceleration sequence data and converging the second angular velocity sequence data to the first angular velocity sequence data.
And a second input module 205, configured to input the measurement data of the inertial measurement unit to the trained data model.
In one embodiment, in the case that the signal of the vehicle positioning module is not available, the second input module 205 may obtain the acceleration of the vehicle through an accelerometer of the inertial measurement unit, obtain the angular velocity of the vehicle through a gyroscope of the inertial measurement unit, and input the acceleration and the angular velocity of the vehicle to a trained data model based on the LSTM neural network model.
And a correction data obtaining module 206, configured to enable the trained data model to perform time correction on the measurement data of the inertial measurement unit input by the second input module 205, and output corrected measurement data.
In one embodiment, the correction data obtaining module 206 obtains the corrected acceleration and the corrected angular velocity output by the data model by making the trained data model time-correct the acceleration and the angular velocity of the vehicle input by the second input module 205, respectively. The data model based on the LSTM neural network model corrects the acceleration and angular velocity of the vehicle input by the second input module 205, respectively, and outputs a corrected acceleration and a corrected angular velocity that reduce the error.
And the positioning module 207 is used for correcting the measurement data according to the initial pose of the vehicle to obtain the positioning data of the vehicle.
In one embodiment, the positioning module 207 uses the time (the time when the signal of the positioning module is available) before the signal of the positioning module of the vehicle is unavailable as the initial time according to the positioning information of the time before the signal of the positioning module of the vehicle is unavailable (the time when the signal of the positioning module is available), and uses the positioning information of the initial time of the vehicle as the initial pose of the vehicle, wherein the initial pose of the vehicle comprises the position information, the course angle information and the speed information of the vehicle at the initial time. The positioning module 207 obtains positioning data of the vehicle according to the initial pose of the vehicle, the corrected acceleration and the corrected angular velocity output after the data model is corrected.
According to the technical scheme, the first spline curve group is constructed by spline fitting processing of the positioning information of the positioning module, derivation is conducted on the first spline curve group, and first acceleration sequence data and first angular velocity sequence data of the positioning module are obtained. And aligning the positioning information of the low-frequency positioning module with the measurement data of the high-frequency inertial measurement unit in a spline curve mode through spline fitting processing. Compared with the related art, which takes position and velocity as input samples, the embodiment of the application takes the first acceleration sequence data and the first angular velocity sequence data of the positioning module, and the second acceleration series data and the second angular velocity series data of the inertial measurement unit are samples, taking the first acceleration sequence data and the first angular speed sequence data of the positioning module as output expected values, obtaining a trained data model by converging the second acceleration sequence data toward the first acceleration sequence data and converging the second angular velocity sequence data toward the first angular velocity sequence data, under the condition of the positioning information of the positioning modules with the same quantity, the generalization effect is improved, the efficiency of data model training can be improved, the training precision of the data model is improved, the accumulative error of the inertia measurement unit is better reduced, and the precision of vehicle positioning according to the measurement data of the inertia measurement unit is improved. The trained data model based on the LSTM neural network model is used for carrying out time correction on the measurement data input into the inertial measurement unit and outputting corrected measurement data; the positioning data of the vehicle is obtained by correcting the measurement data according to the initial pose of the vehicle, the accumulated error of the inertia measurement unit can be reduced based on the LSTM neural network model, and the precision of vehicle positioning according to the measurement data of the inertia measurement unit is improved.
With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.
Fig. 3 is a schematic structural diagram of an electronic device shown in an embodiment of the present application.
Referring to fig. 3, the electronic device 30 includes a memory 301 and a processor 302.
The Processor 302 may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic, discrete hardware components, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The memory 301 may include various types of storage units, such as system memory, Read Only Memory (ROM), and permanent storage. Wherein the ROM may store static data or instructions for the processor 302 or other modules of the computer. The persistent storage device may be a read-write storage device. The persistent storage may be a non-volatile storage device that does not lose stored instructions and data even after the computer is powered off. In some embodiments, the persistent storage device employs a mass storage device (e.g., magnetic or optical disk, flash memory) as the persistent storage device. In other embodiments, the permanent storage may be a removable storage device (e.g., floppy disk, optical drive). The system memory may be a read-write memory device or a volatile read-write memory device, such as a dynamic random access memory. The system memory may store instructions and data that some or all of the processors require at runtime. In addition, the memory 301 may comprise any combination of computer-readable storage media, including various types of semiconductor memory chips (DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), magnetic and/or optical disks, may also be employed. In some embodiments, memory 301 may include a removable storage device that is readable and/or writable, such as a Compact Disc (CD), a read-only digital versatile disc (e.g., DVD-ROM, dual layer DVD-ROM), a read-only Blu-ray disc, an ultra-density optical disc, a flash memory card (e.g., SD card, min SD card, Micro-SD card, etc.), a magnetic floppy disc, or the like. Computer-readable storage media do not contain carrier waves or transitory electronic signals transmitted by wireless or wired means.
The memory 301 has stored thereon executable code that, when processed by the processor 302, may cause the processor 302 to perform some or all of the methods described above.
The aspects of the present application have been described in detail hereinabove with reference to the accompanying drawings. In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments. Those skilled in the art should also appreciate that the acts and modules referred to in the specification are not necessarily required in the present application. In addition, it can be understood that the steps in the method of the embodiment of the present application may be sequentially adjusted, combined, and deleted according to actual needs, and the modules in the device of the embodiment of the present application may be combined, divided, and deleted according to actual needs.
Furthermore, the method according to the present application may also be implemented as a computer program or computer program product comprising computer program code instructions for performing some or all of the steps of the above-described method of the present application.
Alternatively, the present application may also be embodied as a non-transitory machine-readable storage medium (or computer-readable storage medium, or machine-readable storage medium) having stored thereon executable code (or a computer program, or computer instruction code) which, when executed by a processor of an electronic device (or electronic device, server, etc.), causes the processor to perform part or all of the various steps of the above-described method according to the present application.
Those of skill would further appreciate that the various illustrative logical blocks, modules, circuits, and algorithm steps described in connection with the applications disclosed herein may be implemented as electronic hardware, computer software, or combinations of both.
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems and methods according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
Having described embodiments of the present application, the foregoing description is intended to be exemplary, not exhaustive, and not limited to the disclosed embodiments. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen in order to best explain the principles of the embodiments, the practical application, or improvements made to the technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.

Claims (10)

1. A vehicle positioning method based on an LSTM neural network model is characterized by comprising the following steps:
respectively carrying out spline fitting processing to construct a first spline curve group according to the position information and the posture information of the positioning module, wherein the first spline curve group comprises a first position spline curve and a first posture spline curve;
according to the first spline curve group, obtaining first acceleration sequence data of the positioning module and first angular speed sequence data of the positioning module;
inputting the first acceleration sequence data and the first angular velocity sequence data, and second acceleration sequence data and second angular velocity sequence data of an inertial measurement unit to a preset LSTM neural network model, wherein the first acceleration sequence data and the second acceleration sequence data are aligned in sampling time, and the first angular velocity sequence data and the second angular velocity sequence data are aligned in sampling time;
converging the second acceleration sequence data to the first acceleration sequence data, and converging the second angular velocity sequence data to the first angular velocity sequence data to obtain a trained data model;
inputting the measurement data of an inertia measurement unit into the trained data model;
enabling the trained data model to carry out time correction on the measurement data of the inertial measurement unit and output corrected measurement data;
and obtaining positioning data of the vehicle according to the initial pose of the vehicle and the corrected measurement data.
2. The method of claim 1, wherein the spline fitting process is performed according to the position information and the pose information of the positioning module to construct a first spline curve group, and the first spline curve group comprises a first position spline curve and a first pose spline curve, and comprises:
b spline fitting processing is carried out according to the position information of the positioning module to construct the first position spline curve;
and B-spline fitting is carried out according to the attitude information of the positioning module to construct the first attitude spline curve.
3. The method of claim 1, wherein said obtaining first acceleration sequence data for the localization module and first angular velocity sequence data for the localization module from the first spline set comprises:
and performing second-order derivation on the first position spline curve to obtain the first acceleration sequence data, and performing first-order derivation on the first posture spline curve to obtain the first angular velocity sequence data.
4. The method of claim 1, wherein the obtaining a trained data model by converging the second acceleration sequence data to the first acceleration sequence data and converging the second angular velocity sequence data to the first angular velocity sequence data comprises:
calculating to obtain mean square errors of the second acceleration sequence data and the first acceleration sequence data, and the second angular velocity sequence data and the first angular velocity sequence data by adopting a mean square error algorithm in a preset time step;
and obtaining a trained data model by enabling the mean square error convergence to be smaller than a preset threshold value.
5. The method of claim 4, wherein: the numerical value of the time step is a multiple of the output frequency of the positioning module.
6. A vehicle localization apparatus based on an LSTM neural network model, comprising:
the curve construction module is used for respectively carrying out spline fitting processing according to the position information and the posture information of the positioning module to construct a first spline curve group, and the first spline curve group comprises a first position spline curve and a first posture spline curve;
the sequence data acquisition module is used for acquiring first acceleration sequence data of the positioning module and first angular velocity sequence data of the positioning module according to the first curve group constructed by the curve construction module;
a first input module, configured to input, to a preset LSTM neural network model, the first acceleration sequence data and the first angular velocity sequence data obtained by the sequence data acquisition module, and second acceleration sequence data and second angular velocity sequence data of an inertial measurement unit, where the first acceleration sequence data and the second acceleration sequence data are aligned at a sampling time, and the first angular velocity sequence data and the second angular velocity sequence data are aligned at the sampling time;
a training module, configured to converge second acceleration sequence data input by the first input module to first acceleration sequence data input by the first input module, and converge second angular velocity sequence data input by the first input module to first angular velocity sequence data input by the first input module, so as to obtain a trained data model;
the second input module is used for inputting the measurement data of the inertial measurement unit into the trained data model;
the correction data acquisition module is used for enabling the trained data model to carry out time correction on the measurement data of the inertial measurement unit input by the second input module and output correction measurement data;
and the positioning module is used for obtaining positioning data of the vehicle according to the initial pose of the vehicle and the corrected measurement data.
7. The apparatus of claim 6, wherein the curve construction module is specifically configured to:
b spline fitting processing is carried out according to the position information of the positioning module to construct the first position spline curve;
and B-spline fitting is carried out according to the attitude information of the positioning module to construct the first attitude spline curve.
8. The apparatus according to claim 6, wherein the sequence data obtaining module is specifically configured to perform a second-order derivation on the first position spline constructed by the curve construction module to obtain the first acceleration sequence data, and perform a first-order derivation on the first pose spline constructed by the curve construction module to obtain the first angular velocity sequence data.
9. The apparatus of claim 6, wherein the training module is further configured to:
calculating by using a mean square error algorithm with a preset time step to obtain mean square errors of second acceleration sequence data and first acceleration sequence data input by the first input module and second angular velocity sequence data and first angular velocity sequence data input by the first input module;
and obtaining the trained data model by enabling the mean square error convergence to be smaller than a preset threshold value.
10. A non-transitory machine-readable storage medium storing executable code, which when executed by a processor of an electronic device, causes the processor to perform the method of any one of claims 1-5.
CN202110411903.9A 2021-04-16 2021-04-16 Vehicle positioning method and device based on LSTM neural network model Active CN113124884B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110411903.9A CN113124884B (en) 2021-04-16 2021-04-16 Vehicle positioning method and device based on LSTM neural network model

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110411903.9A CN113124884B (en) 2021-04-16 2021-04-16 Vehicle positioning method and device based on LSTM neural network model

Publications (2)

Publication Number Publication Date
CN113124884A true CN113124884A (en) 2021-07-16
CN113124884B CN113124884B (en) 2023-04-18

Family

ID=76776732

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110411903.9A Active CN113124884B (en) 2021-04-16 2021-04-16 Vehicle positioning method and device based on LSTM neural network model

Country Status (1)

Country Link
CN (1) CN113124884B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106980133A (en) * 2017-01-18 2017-07-25 中国南方电网有限责任公司超高压输电公司广州局 The GPS INS Combinated navigation methods and system for being compensated and being corrected using neural network algorithm
CN109059909A (en) * 2018-07-23 2018-12-21 兰州交通大学 Satellite based on neural network aiding/inertial navigation train locating method and system
CN109579853A (en) * 2019-01-24 2019-04-05 燕山大学 Inertial navigation indoor orientation method based on BP neural network
CN111007455A (en) * 2019-10-16 2020-04-14 张苏 Positioning system and method, database and neural network model training method
US20210095965A1 (en) * 2019-09-26 2021-04-01 Harbin Engineering University Elman neural network assisting tight-integrated navigation method without GNSS signals

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106980133A (en) * 2017-01-18 2017-07-25 中国南方电网有限责任公司超高压输电公司广州局 The GPS INS Combinated navigation methods and system for being compensated and being corrected using neural network algorithm
CN109059909A (en) * 2018-07-23 2018-12-21 兰州交通大学 Satellite based on neural network aiding/inertial navigation train locating method and system
CN109579853A (en) * 2019-01-24 2019-04-05 燕山大学 Inertial navigation indoor orientation method based on BP neural network
US20210095965A1 (en) * 2019-09-26 2021-04-01 Harbin Engineering University Elman neural network assisting tight-integrated navigation method without GNSS signals
CN111007455A (en) * 2019-10-16 2020-04-14 张苏 Positioning system and method, database and neural network model training method

Also Published As

Publication number Publication date
CN113124884B (en) 2023-04-18

Similar Documents

Publication Publication Date Title
CN112762933B (en) Vehicle positioning method and device based on neural network model
CN111928869B (en) Vehicle motion track estimation method and device and electronic equipment
CN101846734B (en) Agricultural machinery navigation and position method and system and agricultural machinery industrial personal computer
WO2021164341A1 (en) Heading mounting error determination
CN112146683B (en) Inertial measurement unit calibration parameter adjusting method and device and electronic equipment
CN112461269B (en) Inertial measurement unit calibration method, device and server
CN112985462B (en) Method and device for correcting inertial measurement data based on convolutional neural network model
CN111982158B (en) Inertial measurement unit calibration method and device
CN112197772B (en) Method and related device for determining vehicle movement track
CN112118537B (en) Method and related device for estimating movement track by using picture
CN113375669B (en) Attitude updating method and device based on neural network model
CN113252060B (en) Vehicle track calculation method and device based on neural network model
CN113607176B (en) Combined navigation system track output method and device
CN113175936B (en) Vehicle track calculation method and device based on neural network model
CN113124884B (en) Vehicle positioning method and device based on LSTM neural network model
CN112762932B (en) Trajectory calculation method and device based on neural network model
CN113917512B (en) Positioning method and device for automatic driving vehicle, electronic equipment and storage medium
CN112130188B (en) Vehicle positioning method and device and cloud server
CN112985463B (en) Calibration method and device for inertial measurement unit based on convolutional neural network model
CN112284421B (en) IMU internal reference adjusting method and related device
CN112348903B (en) Method and device for calibrating external parameters of automobile data recorder and electronic equipment
CN114018262B (en) Improved derivative volume Kalman filtering integrated navigation method
CN114543800A (en) Processing method and related device for vehicle movement track estimation
CN111928865B (en) Track calculation method and device, electronic equipment and storage medium
US20240118435A1 (en) Post-processing of global navigation satellite system (gnss) data

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant