WO2020048394A1 - 定位方法、装置、设备和计算机可读存储介质 - Google Patents

定位方法、装置、设备和计算机可读存储介质 Download PDF

Info

Publication number
WO2020048394A1
WO2020048394A1 PCT/CN2019/103662 CN2019103662W WO2020048394A1 WO 2020048394 A1 WO2020048394 A1 WO 2020048394A1 CN 2019103662 W CN2019103662 W CN 2019103662W WO 2020048394 A1 WO2020048394 A1 WO 2020048394A1
Authority
WO
WIPO (PCT)
Prior art keywords
information
fusion positioning
update information
positioning method
fusion
Prior art date
Application number
PCT/CN2019/103662
Other languages
English (en)
French (fr)
Inventor
梁乔忠
Original Assignee
腾讯科技（深圳）有限公司
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 腾讯科技（深圳）有限公司 filed Critical 腾讯科技（深圳）有限公司
Priority to EP19857105.1A priority Critical patent/EP3848730A4/en
Publication of WO2020048394A1 publication Critical patent/WO2020048394A1/zh
Priority to US17/075,880 priority patent/US11796320B2/en

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
    • G01C21/10—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
    • G01C21/18—Stabilised platforms, e.g. by gyroscope
    • 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
    • G01S19/01—Satellite radio beacon positioning systems transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
    • G01S19/13—Receivers
    • G01S19/24—Acquisition or tracking or demodulation of signals transmitted by the system
    • G01S19/30—Acquisition or tracking or demodulation of signals transmitted by the system code related
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
    • G01S19/38—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system
    • G01S19/39—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system the satellite radio beacon positioning system transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
    • G01S19/42—Determining position
    • G01S19/45—Determining position by combining measurements of signals from the satellite radio beacon positioning system with a supplementary measurement
    • G01S19/47—Determining position by combining measurements of signals from the satellite radio beacon positioning system with a supplementary measurement the supplementary measurement being an inertial measurement, e.g. tightly coupled inertial
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
    • G01S19/38—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system
    • G01S19/39—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system the satellite radio beacon positioning system transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
    • G01S19/42—Determining position
    • G01S19/48—Determining position by combining or switching between position solutions derived from the satellite radio beacon positioning system and position solutions derived from a further system
    • G01S19/49—Determining position by combining or switching between position solutions derived from the satellite radio beacon positioning system and position solutions derived from a further system whereby the further system is an inertial position system, e.g. loosely-coupled
    • 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
    • G01S19/53—Determining attitude
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • G06F18/251—Fusion techniques of input or preprocessed data

Definitions

  • the present application relates to the field of computer technology, and in particular, to a positioning method, an apparatus, a device, and a readable storage medium.
  • GPS Global Positioning System
  • inertial measurement units are generally used as auxiliary elements of GPS positioning.
  • GPS information is used as the basic element of positioning.
  • the information of the inertial measurement unit is used as the basic element of positioning.
  • the signal threshold of the GPS signal is set in general.
  • the information is switched to positioning based on the information of the inertial measurement unit, and the accuracy of the positioning accuracy is not high.
  • the embodiments of the present application provide a positioning method, a device, a device, and a readable storage medium, which can be used to solve the problem of low accuracy of positioning accuracy.
  • the technical solution is as follows:
  • a positioning method includes:
  • the current change update information includes GPS positioning information and dead reckoning position information, and the dead reckoning position information is determined based on terminal attitude change information;
  • the GPS positioning information and the dead reckoning position information in the current change update information are combined with the error information to perform fusion positioning to obtain the current positioning position.
  • a positioning device comprising:
  • An obtaining module configured to obtain current change update information, where the current change update information includes GPS positioning information and dead reckoning position information, and the dead reckoning position information is determined according to terminal attitude change information;
  • a determining module configured to determine a fusion positioning method that matches the current change update information, and the fusion positioning method is used to perform fusion positioning by using information in the current change update information;
  • the determining module is further configured to determine error information according to the current change update information, and the error information is used to indicate an error situation of the current change and more information;
  • a positioning module is configured to perform fusion positioning on the GPS positioning information and the dead reckoning position information in the current change update information in combination with the error information to obtain the current positioning position according to the integrated positioning method.
  • a computer device including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the steps of the method as described above are implemented.
  • a non-volatile computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method as described above are implemented.
  • the matching fusion positioning method and error information are determined, and then the fusion positioning is performed according to the matching capacity and positioning method and error information to obtain the current positioning position, which can dynamically classify the matches
  • the fusion positioning method for fusion positioning can effectively improve the accuracy of fusion positioning and can provide better robustness for the performance of different scenarios.
  • FIG. 1 is a flowchart of a positioning method provided by an exemplary embodiment of the present application
  • FIG. 2 is a schematic diagram of a positioning method provided by an exemplary embodiment of the present application.
  • FIG. 3 is a flowchart of a positioning method according to another exemplary embodiment of the present application.
  • FIG. 4 is a flowchart of a positioning method according to another exemplary embodiment of the present application.
  • FIG. 5 is a flowchart of a positioning method according to another exemplary embodiment of the present application.
  • FIG. 6 is a schematic structural diagram of a module of a positioning device according to an exemplary embodiment of the present application.
  • FIG. 7 is an internal structure diagram of a positioning device according to an exemplary embodiment of the present application.
  • the positioning method provided in the present application can be applied to any related device (such as car navigation, etc.) that can be used in combination with inertial sensing parameters measured by the inertial measurement unit to perform positioning.
  • the related equipment has an inertial measurement unit and a GPS positioning device, wherein the inertial measurement unit is a device that measures the three-axis attitude angle (or angular rate) and acceleration of an object.
  • the relevant device as an on-board navigation as an example, after the attitude change information is measured by the inertial measurement unit and the GPS positioning information is obtained by the GPS positioning device, the fusion positioning method is matched by combining the attitude change information and the GPS positioning information to perform the fusion positioning processing. Final positioning position.
  • the device After obtaining the posture change information and GPS positioning information sent by devices such as in-vehicle navigation, the device can match the fusion positioning method to perform fusion positioning. Processing, obtaining the final positioning position, and then returning the obtained final positioning position to a device such as a car navigation so that the device can be applied.
  • fusion positioning processing performed by the related device matching the fusion positioning method is performed to obtain the final positioning position as an example for description.
  • FIG. 1 is a flowchart of a positioning method provided by an exemplary embodiment of the present application. The method is applied to a terminal as an example for description. As shown in FIG. 1, the method includes:
  • Step 101 Obtain current change update information.
  • the current change update information includes GPS positioning information and dead reckoning position information, and the dead reckoning position information is determined according to terminal attitude change information.
  • the change update information is periodically obtained every other time period.
  • the above-mentioned GPS positioning information refers to GPS positioning information obtained through detection by a GPS positioning module at the current time, such as GPS angle, GPS position, and the like.
  • the above-mentioned current attitude change information may be determined through relevant parameters obtained by detection by the inertial measurement unit.
  • the inertial measurement unit is a device for measuring the three-axis attitude angle (or angular rate) and acceleration of an object.
  • the inertial measurement unit includes at least one of an accelerometer and a gyroscope.
  • the accelerometer is an instrument for measuring acceleration.
  • the gyroscope The instrument is also called an angular velocity sensor, which measures the angular velocity of the object when it is deflected or tilted, or is used to sense the direction.
  • the corresponding current attitude change information may include measured acceleration and angular velocity and other parameters.
  • the dead reckoning position information when the dead reckoning position information is determined according to the current attitude change information, the dead reckoning position information may be obtained by performing dead reckoning (DR) on the current attitude change information.
  • DR dead reckoning
  • the manner of determining the dead reckoning position information includes at least one of the following manners:
  • attitude change information includes speed information
  • the attitude change information is updated according to the gyro rotation speed and the angular velocity increments collected by the gyroscope for two consecutive times, and then dead reckoning is performed according to the updated current attitude change information to obtain the dead reckoning position information;
  • dead reckoning is performed according to the current attitude change information to obtain dead reckoning position information.
  • Step 102 Determine a fusion positioning mode that matches the current change update information.
  • the fusion positioning method may be a fusion positioning method determined from at least two fusion positioning methods.
  • the at least two fusion positioning methods include: a Kalman filter fusion positioning method and a particle filter fusion positioning method, and also a fusion positioning.
  • the method may also include other types of fusion positioning methods, and include a larger number of fusion positioning methods.
  • the last matching fusion positioning method may be determined as determining the matching fusion positioning method from at least two or more fusion positioning methods. Specifically, when the current time is a predetermined time period after the last matched fusion positioning method, the last matched fusion positioning method is determined as the matched fusion positioning method. Therefore, it is possible to directly use the last matching fusion positioning method for fusion positioning without dynamically determining an appropriate fusion positioning method from time to time, saving equipment performance, improving processing efficiency, and improving smoothness.
  • the update information based on historical changes before startup is used as the sample change update information, and the fusion positioning model is trained by using the sample change update information to obtain a trained fusion positioning model.
  • the fusion positioning model is trained by inputting the current change update information into the fusion positioning model, a corresponding matching fusion positioning method can be obtained.
  • the training process of the above-mentioned fusion positioning model may also be performed after the car navigation device is started, and within a certain period of time after the startup, using the change update information collected during the time period as the sample change update information. That is, within a certain period of time after the in-vehicle navigation device is started, the fusion positioning is performed using each of the fusion positioning methods in the initial fusion positioning model, so that the trained fusion positioning model is more in line with the actual application situation of the current car navigation device. To improve the positioning performance such as smoothness in the actual positioning process.
  • the current change update information refers to the change and new information obtained in the positioning process after the fusion positioning model has been trained to obtain it.
  • determining a matching fusion positioning mode may include the following steps 1021 and 1022.
  • Step 1021 Determine model input features based on the current change update information.
  • one embodiment may perform information processing based on the above-mentioned current change update information to obtain model input features corresponding to the input parameters of the fused positioning model.
  • the model input features determined based on the current change update information may include dead reckoning angle, dead reckoning position error, gyroscope drift, mileage error coefficient, GPS angle And GPS position error. It can be understood that, based on the different fusion positioning methods included in the fusion positioning model, the corresponding model input features (such as quantity and type) may be different. Therefore, the method for obtaining the model input features is not specifically limited in this embodiment. .
  • the determined input features of each model can be combined into a feature vector or a feature matrix and input into the fusion positioning model.
  • step 1022 the above-mentioned model is input into the fused localization model obtained through training to determine a fused localization mode that matches the current change update information.
  • a fusion localization model is obtained based on the training described above, the weights corresponding to the characteristics of each input parameter of the fusion localization model are determined, and the model input features are input into the fusion localization model obtained through training, so that each input corresponding to the information can be updated based on changes
  • the value corresponding to the parameter feature and the weight corresponding to each input parameter feature determine the probability that the change update information belongs to the classification of each fusion positioning method.
  • the fusion positioning model obtained by training the model input features input can obtain the probability of classifying the change update information into the Kalman filter fusion localization method. , And the probability of classification to the particle filter fusion positioning method2.
  • the fusion positioning method corresponding to the maximum probability may be determined as a matching fusion positioning method classification result.
  • the Kalman filter fusion positioning method is used as the classification result of the matching fusion positioning method, that is, the matched fusion positioning method is the Kalman filtering fusion positioning method.
  • the fusion positioning method corresponding to the probability with the largest value and greater than or equal to the probability threshold may be determined as the matching fusion positioning method classification result.
  • the Kalman filter fusion positioning method is used as the classification result of the matching fusion positioning method, that is, the matching fusion positioning method is Kalman filter fusion positioning method.
  • the probability threshold can be set in combination with the actual technical requirements, as long as it can match a significantly better fusion positioning method.
  • the fusion positioning method used at the last moment may be the fusion positioning method that was self-matched at the previous moment or may be before the previous moment
  • the matching fusion positioning method determines the classification result of the matching fusion positioning method.
  • the value of probability 1 is greater than the value of probability 2 and the value of probability 1 is less than the probability threshold, it means that the advantages of the fusion positioning using the Kalman filtering fusion positioning method and the particle filtering fusion positioning method are not obvious.
  • the fusion positioning method used at the previous moment can make the positioning process smooth.
  • Step 103 Determine the error information according to the current change update information.
  • the corresponding information to be used is different.
  • the particle filter fusion localization method estimates the fusion localization position based on the probability distribution of particles, which may not need to calculate error information.
  • the Kalman filter fusion positioning method needs to be combined with dead reckoning position information for fusion positioning, and in the process, it needs to be combined with error information.
  • a fusion positioning method including error information parameters is used as an example for description.
  • the error information may be determined based on the GPS positioning information and dead reckoning position information.
  • the error information may include at least one of a GPS positioning error, a dead reckoning error, and an error between the positioning information between the GPS positioning information and the dead reckoning position information.
  • the GPS positioning error includes at least one of a GPS angle error and a GPS position error
  • the dead reckoning error may be determined based on the GPS position increment of the GPS positioning information for a predetermined number of consecutive times, and may specifically include: DR position error, gyroscope drift At least one of the DR mileage error coefficient
  • errors between positioning information include at least one of an angular difference between GPS positioning information and dead reckoning position information, and a position difference between GPS positioning information and dead reckoning position information One.
  • Step 104 According to the fusion positioning method, the GPS positioning information and dead reckoning information in the current change update information are combined with the error information to perform fusion positioning to obtain the current positioning position.
  • the fusion positioning based on the matched fusion positioning method when performing the fusion positioning based on the matched fusion positioning method, it may be performed based on an existing fusion positioning process of the matched fusion positioning method, or based on an improved process of the fusion positioning method. This embodiment Not limited.
  • the fusion positioning is performed on the corresponding information in the current change update information to determine the current positioning position, as in one embodiment.
  • the initial positioning position obtained by the fusion positioning is updated to obtain the final current positioning position.
  • the matching fusion positioning method is a Kalman filtering fusion positioning method as an example.
  • the Kalman filtering positioning method may be used for fusion positioning to obtain a final positioning position.
  • the method may be existing and may appear in the future. Any Kalman filtering positioning method will not be described in detail here.
  • the predicted value weighting parameter Q may be a fixed amount
  • the measured value weighting parameter R may be an adjustment amount
  • the measured value weighting parameter R may be adjusted based on the relative error of the GPS positioning information. .
  • FIG. 2 shows a schematic diagram of a positioning method provided by an exemplary embodiment of the present application, in which the dead reckoning DR is determined by the inertial measurement unit 210 and the odometer 220 and is obtained by combining GPS and DR The positioning 230 is fused, and the positioning position is obtained by output.
  • the positioning method obtained in the embodiment of the present application obtains GPS positioning information and dead reckoning position information, determines a matching fusion positioning method and error information based on the information, and then according to the matching capacity and positioning method and error information
  • the fusion positioning is performed to obtain the current positioning position, so that the matching fusion positioning method can be dynamically classified to perform the fusion positioning, which can effectively improve the accuracy of the fusion positioning, and can provide better robustness for the performance of different scenarios.
  • a fusion positioning model is first obtained through training and learning, and then the fusion behavior model is applied to the actual positioning process to realize real-time fusion positioning.
  • the fusion positioning model can be obtained by training on an actual positioning device (such as a car navigation device) and applied to the device, or it can be obtained by training from a third-party device (such as a server), and the training fusion model can be sent to An actual positioning device (such as a car navigation device), which completes the actual positioning process.
  • the localization method in one embodiment includes the following steps 301 to 308.
  • Step 301 Obtain a predetermined number of sample change update information, and determine a sample input feature of each sample change update information.
  • all the sample change update information can be regarded as a set, or all the sample change update information can be divided into multiple sets, and each set contains a certain number of sample change updates.
  • Information the number of sample change update information included in each set can be set in combination with actual technical needs. In this embodiment, the number of sample change update information included in a set is recorded as a predetermined number.
  • any sample change update information may include: sample GPS positioning information and sample dead reckoning position information that are related to each other.
  • any sample change update information may include: interrelated sample GPS positioning information and sample pose change information, and then when training is needed or at the initial stage of training, based on the sample pose change information Perform dead reckoning to obtain the corresponding sample dead reckoning position information.
  • the sample GPS positioning information may be GPS information obtained through a GPS positioning module, or GPS positioning information obtained through other methods.
  • Dead reckoning refers to estimating the position at the next moment by measuring the distance and azimuth of the movement while knowing the position at the current moment.
  • the dead reckoning process may include two parts: vehicle attitude update and position update provided by the gyroscope.
  • dead reckoning may be performed after the speed is updated.
  • the speed update can be determined based on the mileage data provided by the odometer.
  • the speed update is performed based on the odometer data recorded by the odometer is subjected to odometer error processing, and the speed update is performed based on the odometer data processed by the odometer error.
  • the northward component representing the angular velocity of the N system relative to the E system (Earth coordinate system) and
  • r m indicates the radius of the long axis of the earth
  • h k-1 indicates the altitude at time k-1
  • tan represents the tangent trigonometric function
  • Lon k-1 represents the longitude at time k-1
  • r n represents the earth's short-axis radius.
  • a ib represents the rotation angle of the gyroscope from time k-1 to time k
  • delta (A 1 ) and delta (A 2 ) are the angular velocity increments of two consecutive gyro samples.
  • the position update can be performed by the following formula.
  • Lon k represents the longitude at time k
  • t represents time
  • sec is the tangent trigonometric function
  • r n represents the short axis radius of the earth
  • Lat refers to the latitude
  • Lat k represents the latitude at time k
  • r m represents the long axis radius of the earth.
  • the error equation of DR can be obtained:
  • the accelerometer information may not be needed during dead reckoning.
  • the acceleration measured by the accelerometer may also be used to perform attitude alignment, and then the above-mentioned dead reckoning process is performed.
  • the attitude alignment may mainly be performed by the pitch and roll angle pairs.
  • this embodiment is not specifically limited.
  • a car navigation system is used as an example.
  • the car navigation system uses a MEMS gyroscope
  • the azimuth of the GPS can be used to assist the alignment.
  • the acceleration measured by the accelerometer and the azimuth of the GPS can be used to perform the attitude. After alignment, dead reckoning is performed.
  • Step 302 Update information of any sample change, and use each fusion positioning method in the initial fusion positioning model to perform fusion positioning on the sample change update information, and obtain each fusion positioning method to perform fusion positioning on the sample change update information. Fusion positioning performance parameters.
  • the fusion positioning performance parameter may include any parameter that can be used to evaluate the performance of the fusion positioning, such as, but not limited to, parameters such as the time required for the fusion positioning, the resource consumption of the fusion positioning processing, and the positioning accuracy.
  • the input of the initial fusion positioning model may include GPS positioning information and current attitude change information, or may include GPS positioning information and dead reckoning position information obtained by dead reckoning based on the current attitude change information.
  • the initial fusion positioning model may include more than two fusion positioning methods.
  • One embodiment may include a particle filtering fusion positioning method and a Kalman filtering fusion positioning method.
  • it may also be other fusion positioning methods or include more fusion positioning methods, such as Bayesian estimation fusion positioning methods, Fuzzy logic fusion positioning method and so on.
  • Bayesian estimation fusion positioning methods Fuzzy logic fusion positioning method and so on.
  • the particle filter fusion positioning method is used to fuse and update the sample change update information
  • the particle filter fusion positioning method is used to change the sample. Update the information to perform fusion positioning, determine the sample fusion positioning position and fusion positioning performance parameters.
  • the specific fusion positioning process can be based on the existing and future possible methods of particle filter fusion algorithm.
  • the Kalman filter positioning method is used to perform fusion positioning on the sample change update information
  • the Kalman filter fusion positioning method is used to perform fusion positioning on the sample change update information to determine the sample fusion positioning position and fusion positioning performance parameters.
  • the Kalman filtering algorithm includes the following five basic equations, where equations (10) and (11) are time update equations, and equations (12), (13), and (14) are state update equations .
  • Kg (k) P (kk-1) H T (HP (kk-1) H T + R) -1 (13)
  • the parameter Q represents the statistical characteristics of the Kalman filter algorithm model and reflects the weighting of the model prediction value.
  • the parameter R characterizes the measurement noise characteristic in the measurement process and reflects the weighting of the measurement value.
  • the measurement value weighting parameter R is called a measurement value weighting parameter R. The larger the predicted value weighting parameter Q is, the more trusted the measured value is, and the larger the measured value weighting parameter R is, the more trusted the model predicted value is.
  • the first two equations (10) and (11) are used as prediction equations, and the above equations (12), (13), and (14) are used as update equations, and a Kalman filter algorithm is used for state modeling.
  • the modeled model may include relevant parameters for Kalman filter positioning, such as dead reckoning position information, dead reckoning error, and the like.
  • the modeled model in one embodiment may include the following related content.
  • Kalman filter fusion positioning can be performed in this application based on dead reckoning angle, dead reckoning position error, gyroscope drift, mileage error coefficient, GPS angle error, and GPS position error.
  • a dr indicates DR angle
  • delta (P dr ) indicates DR position error
  • E indicates gyroscope drift
  • delta (K) indicates DR mileage error coefficient
  • delta (A g ) indicates GPS angle error
  • delta (P g ) indicates GPS position error.
  • F1 3 ⁇ 3 represents a matrix of 3 rows and 3 columns
  • F1 3 ⁇ 3 represents a matrix of 3 rows and 3 columns
  • F2 3 ⁇ 3 represents a matrix of 3 rows and 3 columns
  • 0 3 ⁇ 3 represents a zero matrix with 3 rows and 3 columns, where the zero matrix means that all elements in the matrix are 0,
  • 1 3 ⁇ 3 represents the identity matrix with 3 rows and 3 columns, where the identity matrix refers to all diagonal elements in the matrix are 1 and other elements
  • 0 3 ⁇ 6 represents a zero matrix of 3 rows and 6 columns
  • 0 3 ⁇ 9 represents a zero matrix of 3 rows and 9 columns
  • 0 3 ⁇ 12 represents a zero matrix of 3 rows and 12 columns
  • 0 3 ⁇ 15 represents 3 rows 15-column zero matrix
  • the Kalman filter fusion positioning method is used to perform fusion positioning on the sample change update information, which may include the following steps A1 to A3.
  • Step A1 Estimating the position information based on the sample dead reckoning, and determining the corresponding sample dead reckoning error according to the error equations (8) and (9) established above.
  • Step A2 The Kalman filtering fusion positioning method in the initial fusion positioning model is used to perform fusion positioning based on the sample GPS positioning information and the sample dead reckoning position information to obtain the initial fusion positioning position and the initial fusion positioning corresponding to the Kalman filtering fusion positioning method. Correction values for performance parameters and sample dead reckoning errors.
  • the prediction value weighting parameter Q is a fixed amount
  • the measurement value weighting parameter R is an adjustment amount.
  • the weighting parameter R of the measurement value can be adjusted based on the relative error DOP of the GPS positioning information, so that the fusion result can be biased toward the DR information in the place where the GPS signal is not good.
  • Step A3 Correct the initial fusion positioning position based on the sample dead reckoning error correction value, and obtain the corresponding sample fusion positioning position and the corresponding fusion positioning performance parameter.
  • Step 303 Based on the fusion positioning performance parameters, determine a preferred fusion positioning method corresponding to each sample change update information from each fusion positioning method.
  • the information content specifically contained in the fusion positioning performance parameter may be used in any possible way.
  • the fusion positioning method corresponding to the smallest resource occupation amount may be used as the preferred fusion positioning method corresponding to the sample change update information.
  • these two parameters can be comprehensively considered (such as after weighting), and the weighted best value (such as the maximum value or the minimum value) corresponding to it can be selected.
  • the fusion positioning method is used as a preferred fusion positioning method corresponding to the sample change update information. In other embodiments, other methods may also be used to determine a preferred fusion positioning method corresponding to the sample change update information.
  • Step 304 Determine the correspondence between the change update information and the classification result of the fusion positioning method according to the preferred fusion positioning method corresponding to each sample change update information.
  • all sample change update information may be combined to determine the correspondence between the change update information and the classification result of the fusion positioning method, as shown in FIG. 4, which specifically includes the following steps 401 and 402.
  • step 401 the preferred fusion positioning method corresponding to each sample change update information is respectively determined as the classification result of the sample fusion positioning method corresponding to the sample change update information.
  • the corresponding preferred fusion positioning method is used as the classification of the sample change update information.
  • Step 402 Determine the correspondence between the change update information and the fusion positioning method based on the classification update information of each sample and the classification result of the corresponding sample fusion positioning method.
  • the correspondence between the change update information and the classification result of the fusion positioning method may be determined based on only the sample change update information.
  • Most of the sample change update information here corresponds to the same preferred fusion positioning method, and accounts for the largest number of the predetermined number of sample change update information. As shown in FIG. 5, it specifically includes the following steps 501 to 503.
  • Step 501 Determine the number of samples of the sample change update information corresponding to each preferred fusion positioning method.
  • the initial positioning model includes two fusion positioning methods: Kalman filtering fusion positioning method and particle filtering fusion positioning method, where among the N sample change update information, there is a preferred fusion corresponding to K1 sample change update information.
  • the positioning methods are all Kalman filter fusion positioning methods.
  • the number of samples of the sample change update information corresponding to the Kalman filter fusion positioning method is K1.
  • the preferred fusion positioning method corresponding to K2 sample change update information is the particle filter fusion positioning method.
  • the number of samples of the sample change update information corresponding to the particle filter fusion positioning method is K2.
  • Step 502 Determine the preferred fusion positioning method corresponding to the maximum number of samples as the fusion positioning method corresponding to a predetermined number of sample change update information.
  • the Kalman filter fusion positioning method is determined as the classification result of the sample fusion positioning method corresponding to the N sample change update information.
  • the preferred fusion positioning method corresponding to the maximum value and corresponding to a predetermined number of samples exceeding a predetermined ratio may be determined as the classification result of the fusion positioning method corresponding to the predetermined number of sample change update information.
  • the Kalman filter fusion positioning method is determined as the classification result of the sample fusion positioning method corresponding to the N sample change update information.
  • Step 503 Determine a correspondence between the change update information and the classification result of the fusion positioning method based on a predetermined number of sample variation update information or the maximum sample number of sample variation update information and the classification result of the fusion positioning method.
  • the fusion positioning model may be determined based on the correspondence between the determined change update information and the classification result of the fusion positioning method. Otherwise, update the initial fusion positioning model, such as updating the weights corresponding to the characteristics of each input parameter in the initial fusion positioning model, and enter the next round of training process.
  • the training end condition can be set in combination with actual needs. For example, it can be based on the sample GPS positioning information, the sample fusion positioning position corresponding to each fusion positioning method, and whether the fusion positioning performance parameter analysis converges as the training end condition. In this embodiment, training is not performed. End conditions are specifically limited.
  • the weights corresponding to the characteristics of each input parameter of the fusion positioning model are determined, that is, the weights corresponding to the characteristics of each input parameter may be different. Therefore, in the actual positioning based on the fusion positioning model, based on the value corresponding to each input parameter feature corresponding to the change update information and the weight corresponding to each input parameter feature, the probability that the change update information belongs to the classification of each fusion positioning method is determined. Thus, a matching fusion positioning method is obtained to perform actual fusion positioning.
  • the fusion positioning model trained as above can be used in the actual fusion positioning process.
  • the fusion positioning model can be sent to each positioning terminal (such as car navigation, mobile terminal, etc.), and each positioning terminal uses the fusion positioning model to complete the fusion positioning. process.
  • the fusion positioning model may also be stored to the third party.
  • the third party completes the fusion positioning process based on the GPS positioning information and attitude change information (or inertial positioning information determined based on the attitude change information) uploaded by the terminal. And return the obtained final positioning position to the terminal for use.
  • the positioning terminal can directly store the fusion positioning model and apply it in the subsequent positioning process.
  • Step 305 Obtain current change update information, where the current change update information includes GPS positioning information and dead reckoning position information, and the dead reckoning position information is determined according to the terminal attitude change information.
  • the above-mentioned GPS positioning information refers to GPS positioning information obtained through detection by a GPS positioning module at the current time, such as a GPS angle and a GPS position.
  • the above-mentioned current attitude change information may be determined through relevant parameters obtained through detection by the inertial measurement unit.
  • the inertial measurement unit is a device for measuring the three-axis attitude angle (or angular rate) and acceleration of an object.
  • the inertial measurement unit in an embodiment may include an accelerometer and a gyroscope.
  • the accelerometer is an instrument for measuring acceleration.
  • the gyroscope is also called an angular velocity sensor. , Is an instrument that measures the angular velocity of the object when it is deflected or tilted, or is used to sense the direction. At this time, the corresponding current attitude change information may include the measured acceleration and angular velocity and other parameters.
  • Step 306 Determine a fusion positioning mode matching the current change update information.
  • the current location update information is input into the fusion positioning model, so that the fusion location mode corresponding to the current change update information is output.
  • Step 307 Determine the error information according to the current change update information.
  • the error information is determined based on the GPS positioning information and dead reckoning position information.
  • the error information may include at least one of a GPS positioning error, a dead reckoning error, and an error between the positioning information between the GPS positioning information and the dead reckoning position information.
  • Step 308 According to the fusion positioning method, the GPS positioning information and dead reckoning information in the current change update information are combined with the error information to perform fusion positioning to obtain the current positioning position.
  • the matching fusion positioning method is the first fusion positioning method (such as the Kalman filter fusion positioning method), combined with the determined error information above, the fusion positioning is performed on the corresponding information in the current change update information to determine the current positioning position, such as a
  • the initial positioning position obtained by the fusion positioning may be updated in combination with the determined error information to obtain the final current positioning position.
  • the matching fusion positioning method is a Kalman filtering fusion positioning method as an example.
  • the Kalman filtering positioning method may be used for fusion positioning to obtain a final positioning position.
  • the method may be existing and may appear in the future. Any Kalman filtering positioning method will not be described in detail here.
  • the predicted value weighting parameter Q may be a fixed amount
  • the measured value weighting parameter R may be an adjustment amount
  • the measured value weighting parameter R may be adjusted based on the relative error of the GPS positioning information. .
  • the method may further include the step of using the first fusion positioning method at each positioning moment within a predetermined time period after the current time. Perform fusion positioning to determine the positioning position corresponding to each positioning moment. Therefore, after the matching fusion positioning method is obtained, the matched fusion positioning method can be used for fusion positioning at each positioning time in a subsequent predetermined time period, and the fusion positioning method matching is not dynamically performed from time to time, which improves processing efficiency, and Improved smoothness.
  • each positioning within a predetermined period of time from the current time.
  • the matching fusion positioning method is used for fusion positioning.
  • the fusion positioning method is rematched based on the fusion positioning model to verify whether the matched fusion positioning method still has a better fusion. Positioning performance.
  • the fusion positioning method can be rematched based on the fusion positioning model for a certain number of consecutive N2 times or for a specified period of time T2 until N3 or The fusion positioning methods that match T3 for a long period of time are consistent.
  • the values of N1, N2, and N3 may be the same or different, and the values of T1, T2, and T3 may be the same or different.
  • the training process may be directly re-executed to train a new fusion positioning model for application.
  • a matching fusion positioning method can also be obtained at the current time (in one embodiment, it can be limited to a fusion positioning method matched based on the fusion positioning model), and after a predetermined period of time from the current time, It is also possible to directly re-execute the training process so as to train a new fusion positioning model and a new fusion positioning model for application.
  • the positioning method obtained in the embodiment of the present application obtains GPS positioning information and dead reckoning position information, determines a matching fusion positioning method and error information based on the information, and then according to the matching capacity and positioning method and error information
  • the fusion positioning is performed to obtain the current positioning position, so that the matching fusion positioning method can be dynamically classified to perform the fusion positioning, which can effectively improve the accuracy of the fusion positioning, and can provide better robustness for the performance of different scenarios.
  • an embodiment provides a positioning device, which can be applied to any device that can obtain GPS information and inertial sensing parameters measured by the inertial measurement unit and quickly perform fusion positioning, such as Car navigation equipment, etc.
  • the positioning device includes an obtaining module 610, a determining module 620, and a positioning module 630.
  • An obtaining module 610 is configured to obtain current change update information, where the current change update information includes GPS positioning information and dead reckoning position information, and the dead reckoning position information is determined according to terminal attitude change information;
  • a determining module 620 configured to determine a fusion positioning method that matches the current change update information, and the fusion positioning method is used to perform fusion positioning by using information in the current change update information;
  • the determining module 620 is further configured to determine error information according to the current change update information, where the error information is used to indicate an error situation of the current change and more information;
  • a positioning module 630 is configured to perform fusion positioning on the GPS positioning information and the dead reckoning position information in the current change update information in combination with the error information to obtain the current positioning position according to the integrated positioning method.
  • the determining module 620 is further configured to perform odometer error processing on the odometer data recorded by the odometer, and after performing speed update based on the odometer data processed by the odometer error, according to the Performing dead reckoning on current attitude change information to obtain the dead reckoning position information, where the attitude change information includes speed information;
  • the determining module 620 is further configured to update attitude change information according to a gyro rotation speed and an angular velocity increment collected by the gyroscope twice in a row; and perform dead reckoning based on the updated current attitude change information to obtain the Dead reckoning position information;
  • the determining module 620 is further configured to perform dead reckoning based on the current attitude change information to obtain the dead reckoning position information after performing attitude alignment based on the acceleration acquired by the accelerometer;
  • the determining module 620 is further configured to perform dead reckoning based on the current attitude change information to obtain the dead reckoning position information after performing attitude alignment based on the GPS azimuth and the acceleration acquired by the accelerometer.
  • the determining module 620 is further configured to determine a model input feature based on the current change update information; input the model input feature to a fusion positioning model obtained through training, and determine update information with the current change Matching the fusion positioning mode.
  • the determining module 620 is further configured to use the fusion positioning method to perform the fusion positioning at each positioning time within a predetermined time period after the current time, and determine that each positioning time corresponds to Location.
  • the positioning module 630 is further configured to apply the fusion positioning method to perform the fusion positioning within a predetermined period of time after a fusion positioning method that is most recently matched before the current time.
  • the fusion positioning method includes any one of a Kalman filtering fusion positioning method and a particle filtering fusion positioning method.
  • the obtaining module 610 is further configured to obtain a predetermined number of sample change update information; for any one sample change update information, through each fusion positioning method in the initial fusion positioning model, respectively, Perform fusion positioning on the sample change update information, and obtain fusion positioning performance parameters of the fusion change positioning information on the sample change update information;
  • the determining module 620 is further configured to determine a preferred fusion positioning method corresponding to each sample change update information from each fusion positioning method based on the fusion positioning performance parameter; and to select a preferred fusion positioning method corresponding to each sample change update information, The correspondence between the change update information and each fusion positioning method is determined.
  • the determining module 620 is further configured to determine a preferred fusion positioning method corresponding to each of the sample change update information, and respectively determine a classification result of the sample fusion positioning method corresponding to the sample change update information. Determining a correspondence relationship between the change update information and each of the fusion positioning methods based on each of the sample change update information and a corresponding sample fusion positioning method classification result.
  • the determining module 620 is further configured to determine the number of samples of the sample change update information corresponding to each preferred fusion positioning method; and determine the preferred fusion positioning method corresponding to the largest number of samples as the desired fusion positioning method.
  • the fusion positioning method corresponding to the predetermined number of sample change update information is described; based on the predetermined number of sample change update information or the maximum sample number of sample change update information and the classification result of the fusion positioning method, the change update information and the fusion positioning method are determined. Correspondence between classification results.
  • the determining module 620 is further configured to determine the fusion positioning model based on the correspondence between the determined change update information and the classification result of the fusion positioning method when the training end condition is satisfied. , Otherwise, update the initial fusion positioning model and enter the next round of training process.
  • the positioning device obtained in the embodiment of the present application obtains GPS positioning information and dead reckoning position information, determines a matching fusion positioning method and error information based on the information, and then according to the matching capacity and positioning method and error information
  • the fusion positioning is performed to obtain the current positioning position, so that the matching fusion positioning method can be dynamically classified to perform the fusion positioning, which can effectively improve the accuracy of the fusion positioning, and can provide better robustness for the performance of different scenarios.
  • the positioning device 700 may be any device capable of performing positioning processing in combination with inertial sensing parameters measured by the inertial measurement unit 750, such as a terminal or a server.
  • the terminal here may be any device that can A terminal for performing inertial measurement, such as a car navigation device
  • the server may be any server that can obtain the inertial sensing parameters obtained by the inertial measurement unit 750 and perform fusion positioning according to the server.
  • An internal structure diagram of the positioning device 700 in one embodiment may be shown in FIG. 7.
  • the computer device includes a processor 710 and a memory 720 connected through a system bus, and may further include a network interface 730 connected through a system bus.
  • the computer device When the computer device is a terminal device, it may further include a display screen and an input device 740.
  • the positioning device 700 is a device capable of measuring inertial sensing parameters, such as a car navigation device, it may further include devices such as an inertial measurement unit 750, a GPS positioning device 770, an accelerometer, and an odometer 760 connected through a system bus.
  • the processor 710 of the computer device is used to provide computing and control capabilities.
  • the memory 720 of the computer device includes a non-volatile storage medium 721 and an internal memory 722.
  • the nonvolatile storage medium 721 stores an operating system 7211 and a computer program 7212.
  • the internal memory 722 provides an environment for running the operating system 7211, the computer program 7212, and the computer program 7221 in the nonvolatile storage medium 721.
  • the network interface 730 of the computer device is configured to communicate with an external terminal through a network connection.
  • the computer program is executed by the processor 710 to implement a positioning method.
  • the display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen
  • the input device 740 of the computer device may be a touch layer covered on the display screen, or a button, a trackball, or a touch button provided on the computer device casing.
  • the board can also be an external keyboard, trackpad, or mouse.
  • FIG. 7 is only a block diagram of a part of the structure related to the scheme of the present application, and does not constitute a limitation on the inertial navigation device to which the scheme of the present application is applied, and the specific inertial navigation
  • the device may include more or fewer components than shown in the figure, or some components may be combined, or have different component arrangements.
  • an inertial navigation device which includes a memory 720 and a processor 710.
  • a computer program is stored in the memory 720, and the processor 710 implements any of the embodiments described above when the computer program is executed Steps in the method.
  • Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory.
  • Volatile memory can include random access memory (RAM) or external cache memory.
  • RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain Synchlink DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
  • SRAM static RAM
  • DRAM dynamic RAM
  • SDRAM synchronous DRAM
  • DDRSDRAM dual data rate SDRAM
  • ESDRAM enhanced SDRAM
  • SLDRAM synchronous chain Synchlink DRAM
  • Rambus direct RAM
  • DRAM direct memory bus dynamic RAM
  • RDRAM memory bus dynamic RAM

Landscapes

  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Automation & Control Theory (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Navigation (AREA)

Abstract

提供了一种定位方法、定位装置、定位设备和计算机存储介质，涉及计算机技术领域，该方法包括：获取当前变化更新信息，当前变化更新信息中包括GPS定位信息和航位推算位置信息（101）；确定与当前变化更新信息匹配的融合定位方式（102）；根据当前变化更新信息确定误差信息（103）；根据融合定位方式，对当前变化更新信息中的GPS定位信息和航位推算位置信息结合误差信息进行融合定位，得到当前定位位置（104）。该定位方法可以动态分类出匹配的融合定位方式和误差信息来进行融合定位，从而可以有效提高融合定位的精度，能够为不同场景的性能提供较好的鲁棒性。

Description

定位方法、装置、设备和计算机可读存储介质
本申请要求于2018年09月04日提交的申请号为201811025588.0、发明名称为“定位方法、定位装置、定位设备和计算机存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及计算机技术领域，特别是涉及一种定位方法、装置、设备和可读存储介质。
背景技术
在车载导航领域，通过采用全球定位系统(Global Positioning System，GPS)作为定位的基本参考元素，GPS具有长时间精度较高的特性，但在某些情况下也会存在不准确的问题。例如，在高楼林立的城市道路等区域，容易受到地理遮蔽和天气等因素的影响而出现信号漂移的问题，在地下停车场、野外隧道等区域，GPS信号可能会出现不可用的情况。在这些情况下，基于GPS信号进行定位导航的车载导航也会受到较大程度的影响。
为了弥补由于GPS的这些缺陷造成的影响，一般使用惯性测量单元作为GPS定位的辅助元素，其在GPS信号好的时候，采用GPS信息作为定位的基本元素，在GPS信号不好或者没有GPS信号的时候，采用惯性测量单元的信息作为定位的基本元素。
然而，目前是笼统地设置GPS信号的信号阈值，在GPS信号的强度低于该信号阈值时切换到以惯性测量单元的信息为基本元素进行定位，定位精度的准确性不高。
发明内容
本申请实施例提供了一种定位方法、装置、设备和可读存储介质，可以用于解决定位精度的准确性不高的问题，所述技术方案如下：
一方面，提供了一种定位方法，所述方法包括：
获取当前变化更新信息，所述当前变化更新信息中包括GPS定位信息和航位推算位置信息，所述航位推算位置信息是根据终端姿态变化信息确定的；
确定与所述当前变化更新信息匹配的融合定位方式，所述融合定位方式用于通过所述当前变化更新信息中的信息进行融合定位；
根据所述当前变化更新信息确定误差信息，所述误差信息用于表示所述当前变化更新信息的误差情况；
根据所述融合定位方式，对当前变化更新信息中的所述GPS定位信息和所述航位推算位置信息结合所述误差信息进行融合定位，得到当前定位位置。
另一方面，提供了一种定位装置，所述装置包括：
获取模块，用于获取当前变化更新信息，所述当前变化更新信息中包括GPS定位信息和航位推算位置信息，所述航位推算位置信息是根据终端姿态变化信息确定的；
确定模块，用于确定与所述当前变化更新信息匹配的融合定位方式，所述融合定位方式用于通过所述当前变化更新信息中的信息进行融合定位；
所述确定模块，还用于根据所述当前变化更新信息确定误差信息，所述误差信息用于表示所述当前变化更信息的误差情况；
定位模块，用于根据所述融合定位方式，对当前变化更新信息中的所述GPS定位信息和所述航位推算位置信息结合所述误差信息进行融合定位，得到当前定位位置。
另一方面，提供了一种计算机设备，包括存储器和处理器，所述存储器存储有计算机程序，所述处理器执行所述计算机程序时，实现如上所述的方法的步骤。
另一方面，提供了一种非易失性计算机可读存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时，实现如上所述的方法的步骤。
本申请实施例提供的技术方案带来的有益效果至少包括：
通过获取GPS定位信息以及航位推算位置信息，基于这些信息确定匹配的融合定位方式以及误差信息，然后根据匹配的容和定位方式和误差信息进行融 合定位得到当前定位位置，从而可以动态分类出匹配的融合定位方式来进行融合定位，可以有效提高融合定位的精度，能够为不同场景的性能提供较好的鲁棒性。
附图说明
为了更清楚地说明本申请实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1为本申请一个示例性实施例提供的定位方法的流程图；
图2为本申请一个示例性实施例提供的定位方法的原理示意图；
图3为本申请另一个示例性实施例提供的的定位方法的流程图；
图4为本申请另一个示例性实施例提供的的定位方法的流程图；
图5为本申请另一个示例性实施例提供的的定位方法的流程图；
图6本申请一个示例性实施例提供的定位装置的模块结构示意图；
图7本申请一个示例性实施例提供的定位设备的内部结构图。
具体实施方式
为使本申请的目的、技术方案和优点更加清楚，下面将结合附图对本申请实施方式作进一步地详细描述。
本申请提供的定位方法，一个实施例中可以应用于任何可以结合惯性测量单元测得的惯性传感参数进行定位的相关设备(如车载导航等)中，以进行定位。该相关设备中具有惯性测量单元以及GPS定位装置，其中，惯性测量单元是测量物体三轴姿态角(或角速率)以及加速度的装置。以该相关设备为车载导航为例，可以在惯性测量单元测得姿态变化信息、且GPS定位装置获得GPS定位信息后，结合姿态变化信息和GPS定位信息匹配出融合定位方式进行融合定位处理，获得最终定位位置。在另一个实施例中，也可以是应用于其他的可以执行相关运算的设备，该设备可以在获得车载导航等设备发送过来的姿态变化信息和GPS定位信息之后，匹配出融合定位方式进行融合定位处理，获得最终定 位位置，然后再将获得的最终定位位置返回给车载导航等设备，以便该设备进行应用。本申请的下述实施例中，是以该相关设备匹配出融合定位方式进行融合定位处理，获得最终定位位置为例进行说明。
图1是本申请一个示例性实施例提供的定位方法的流程图，以该方法应用于终端中为例进行说明，如图1所示，该方法包括：
步骤101，获取当前变化更新信息，该当前变化更新信息中包括GPS定位信息和航位推算位置信息，该航位推算位置信息是根据终端姿态变化信息确定的。
可选地，以车载导航设备为例，在车载导航设备启动之后，每隔一个时间段定时获得变化更新信息。
其中，上述GPS定位信息指当前时刻通过GPS定位模块检测获得的GPS定位信息，例如GPS角度、GPS位置等。上述当前姿态变化信息可以通过惯性测量单元检测获得的相关参数确定。
惯性测量单元是测量物体三轴姿态角(或角速率)以及加速度的装置，一个实施例中，该惯性测量单元包括加速度计和陀螺仪中的至少一种，加速度计是测量加速度的仪表，陀螺仪又叫角速度传感器，是测量物体偏转、倾斜时的转动角速度或者说用来传感方向的仪表，此时，对应的当前姿态变化信息可包括测量获得的加速度和角速度等参数。
如上所述，在根据当前姿态变化信息确定航位推算位置信息时，可以是通过对当前姿态变化信息进行航位推算(Dead Reckoning，DR)获得该航位推算位置信息。
可选地，该航位推算位置信息的确定方式包括如下方式中的至少一种：
第一，对里程计记录的里程计数据进行里程计误差处理，基于里程计误差处理后的里程计数据进行速度更新之后，根据当前姿态变化信息进行航位推算获得上述航位推算位置信息，该姿态变化信息中包括速度信息；
第二，根据陀螺旋转速度以及陀螺仪连续两次采集的角速度增量进行姿态变化信息的更新，再根据更新后的当前姿态变化信息进行航位推算获得所述航位推算位置信息；
第三，基于加速度计采集获得的加速度进行姿态对准之后，根据当前姿态变化信息进行航位推算获得航位推算位置信息；
第四，基于GPS方位角、以及加速度计采集获得的加速度进行姿态对准之后，根据当前姿态变化信息进行航位推算获得航位推算位置信息。
步骤102，确定关于当前变化更新信息匹配的融合定位方式。
如上所述，该融合定位方式可以是从至少两种融合定位方式中确定的融合定位方式，该至少两种融合定位方式中包括：卡尔曼滤波融合定位方式以及粒子滤波融合定位方式，还融合定位方式也可以是包含其他类型的融合定位方式，以及包含更多数量的融合定位方式。
一个实施例中，可以是将上一次匹配的融合定位方式，确定为从至少两种以上的融合定位方式中确定匹配的融合定位方式。具体可以是在当前时刻位于上一次匹配的融合定位方式之后的预定时间段时，将上一次匹配的融合定位方式确定为匹配的融合定位方式。从而可以直接使用上次匹配的融合定位方式进行融合定位，无需时时动态确定合适的融合定位方式，节省设备性能，且提高了处理效率，且提高了平滑性。
一个实施例中，可以在车载导航设备启动之后，基于启动之前的历史的变化更新信息作为样本变化更新信息，并通过样本变化更新信息对融合定位模型进行训练后，得到训练后的融合定位模型，将当前变化更新信息输入该融合定位模型，即可得到对应匹配的融合定位方式。
在另一个实施例中，上述融合定位模型的训练过程，也可以在车载导航设备启动之后，在启动之后一定时长的时间段内，以该时间段内采集的各变化更新信息作为样本变化更新信息，即在车载导航设备启动之后的一定时长的时间段内，均采用初始融合定位模型中的各融合定位方式进行融合定位，以使得训练获得的融合定位模型更符合当前车载导航设备的实际应用情况，以提高实际定位过程中的平滑性等定位性能。本实施例中，该当前变化更新信息，指已经训练获得融合定位模型之后，在定位过程中获得的变化跟新信息。
可选地，确定匹配的融合定位方式，可以包括下述步骤1021和1022。
步骤1021，基于当前变化更新信息，确定模型输入特征。
在确定模型输入特征时，一个实施例中可以是基于上述当前变化更新信息进行信息处理，获得与融合定位模型的输入参数相对应的模型输入特征。
如上所述，以包含卡尔曼滤波融合定位方式为例，则基于当前变化更新信息确定的模型输入特征可以包括：航位推算角度、航位推算位置误差、陀螺仪漂移、里程误差系数、GPS角度误差以及GPS位置误差等参数。可以理解，基 于融合定位模型中包含的融合定位方式的不同，所对应的模型输入特征(如数量和类型)可以有所不同，因此，具体得到模型输入特征的方式，本实施例不做具体限定。
确定的各模型输入特征，可以组合为特征向量或者组合为特征矩阵，输入到融合定位模型中。
步骤1022，将上述模型输入特征输入训练获得的融合定位模型，确定与当前变化更新信息匹配的融合定位方式。
可选地，基于上述训练获得融合定位模型，确定了该融合定位模型的各个输入参数特征相对应的权重，将模型输入特征输入训练获得的融合定位模型，从而可以基于变化更新信息对应的各输入参数特征对应的值以及各输入参数特征对应的权重，确定出该变化更新信息属于各融合定位方式的分类的概率。以融合定位模型包含卡尔曼滤波融合定位方式和粒子滤波融合定位方式为例，将模型输入特征输入训练获得的融合定位模型，可获得将该变化更新信息分类至卡尔曼滤波融合定位方式的概率1，以及分类至粒子滤波融合定位方式的概率2。
一个实施例中，可以将最大的概率对应的融合定位方式确定为匹配的融合定位方式分类结果。示意性的，若概率1的值大于概率2的值，则将卡尔曼滤波融合定位方式作为匹配的融合定位方式分类结果，即匹配的融合定位方式为卡尔曼滤波融合定位方式。
一个实施例中，可以将值最大、且大于或者等于概率阈值的概率对应的融合定位方式，确定为匹配的融合定位方式分类结果。示意性的，若概率1的值大于概率2的值，且概率1的值大于或者等于概率阈值，则将卡尔曼滤波融合定位方式作为匹配的融合定位方式分类结果，即匹配的融合定位方式为卡尔曼滤波融合定位方式。概率阈值可以结合实际技术需要进行设定，只要能够从中匹配出明显较优的融合定位方式即可。
相应地，在一个实施例中，在值最大的概率小于概率阈值时，可以将上一时刻采用的融合定位方式(可能是上一时刻自行匹配出的融合定位方式，也可能是上一时刻之前匹配出的融合定位方式)，确定为匹配的融合定位方式分类结果。如上所述，若概率1的值大于概率2的值，且概率1的值小于概率阈值，则说明采用卡尔曼滤波融合定位方式和粒子滤波融合定位方式进行融合定位的优势差异并不明显，从而直接上一时刻采用的融合定位方式，可以使得定位过程具有较好的平滑性。
步骤103，根据当前变化更新信息确定误差信息。
可选地，对于不同的融合定位方式，其相应需要使用的信息有所不同。示意性的，粒子滤波融合定位方式是基于粒子的概率分布来估计出融合定位位置，其可以不需要计算误差信息。卡尔曼滤波融合定位方式需要结合航位推算位置信息进行融合定位，在此过程中需要结合误差信息来进行。本实施例中，以包含误差信息参数的融合定位方式为例进行说明。
一个实施例中，可以基于GPS定位信息以及航位推算位置信息确定误差信息。其中，误差信息可以包括：GPS定位误差，航位推算误差以及GPS定位信息与航位推算位置信息之间的定位信息间误差中的至少一种。
其中，GPS定位误差包括GPS角度误差、GPS位置误差中的至少一种；航位推算误差可以基于连续预定次数的GPS定位信息的GPS位置增量确定，具体可以包括：DR位置误差、陀螺仪漂移、DR里程误差系数中的至少一种；定位信息间误差包括：GPS定位信息与航位推算位置信息之间的角度差，以及GPS定位信息与航位推算位置信息之间的位置差中的至少一种。
步骤104，根据融合定位方式，对当前变化更新信息中的GPS定位信息和航位推算信息结合误差信息进行融合定位，得到当前定位位置。
可选地，基于匹配的融合定位方式进行融合定位时，可以基于该匹配的融合定位方式的已有的融合定位过程进行，或者是基于对该融合定位方式加以改进后的过程进行，本实施例不做进行限定。
可选地，在融合定位过程中，根据GPS定位信息和航位推算信息，结合上述确定的误差信息，对当前变化更新信息中对应的信息进行融合定位，确定当前定位位置，如一个实施例中，结合上述确定的误差信息对融合定位得到的初始定位位置进行更新，从而获得最终的当前定位位置。
一个实施例中，以匹配的融合定位方式为卡尔曼滤波融合定位方式为例，可以采用卡尔曼滤波定位方式进行融合定位，获得最终的定位位置的方式，可以是目前已有以及以后可能出现的任何的卡尔曼滤波定位方式，此处不再详加赘述。如上所述，在进行融合定位处理时，可以以预测值加权参数Q为固定量，以测量值加权参数R为调节量，且该测量值加权参数R，可以基于GPS定位信息的相对误差进行调节。
应该理解的是，尽管上述实施例的说明中，给出了具体的几种融合定位算法，但在实际技术实现过程中，还可以在融合定位模型中扩充或者增加其他的 融合定位方式，并不能加以限定。
示意性的，请参考图2，其示出了本申请一个示例性实施例提供的定位方法的原理示意图，其中，通过惯性测量单元210和里程计220确定航位推算DR，结合GPS和DR得到融合定位230，并输出得到定位位置。
综上所述，本申请实施例提供的定位方法，通过获取GPS定位信息以及航位推算位置信息，基于这些信息确定匹配的融合定位方式以及误差信息，然后根据匹配的容和定位方式和误差信息进行融合定位得到当前定位位置，从而可以动态分类出匹配的融合定位方式来进行融合定位，可以有效提高融合定位的精度，能够为不同场景的性能提供较好的鲁棒性。
一个实施例中，首先通过训练学习的方式获得融合定位模型，再将该融合行为模型应用到实际的定位过程中，实现实时的融合定位。该融合定位模型可以由实际进行定位的设备(如车载导航设备)训练获得并应用在该设备中，也可以是由第三方设备(如服务器)训练获得后，将训练获得的融合定位模型发送给实际进行定位的设备(如车载导航设备)，由该设备完成实际的定位过程。
以结合训练学习获得融合定位模型为例，参考图3所示，一个实施例中的定位方法，包括下述步骤301至步骤308。
步骤301：获取预定数目的样本变化更新信息，确定各样本变化更新信息的样本输入特征。
在模型训练过程中，得到样本变化更新信息后，可以将所有的样本变化更新信息作为一个集合，或者将所有的样本变化更新信息划分为多个集合，每个集合中包含一定数量的样本变化更新信息，每个集合中包含的样本变化更新信息的数量可以结合实际技术需要设定，本实施例中将一个集合中包含的样本变化更新信息的数目记为预定数目。
在一个实施例中，任何一个样本变化更新信息中，可以包括有：相互关联的样本GPS定位信息以及样本航位推算位置信息。在另一个实施例中，任何一个样本变化更新信息中，可以包括有：相互关联的样本GPS定位信息以及样本姿态变化信息，然后在需要进行训练时或者在训练的初始阶段，基于样本姿态变化信息进行航位推算获得对应的样本航位推算位置信息。其中，样本GPS定位信息可以是通过GPS定位模块获得的GPS信息，或者是通过其他的方式获得的GPS定位信息。
航位推算指在知道当前时刻位置的条件下，通过测量移动的距离和方位，推算下一时刻位置。在进行基于姿态变化信息获得对应的航位推算位置信息的航位推算时，航位推算的过程可包括陀螺仪提供的车辆姿态更新和位置更新两个部分。一个实施例中，可以是在速度更新之后再进行航位推算。速度更新基于里程计提供里程数据确定即可，在进行速度更新时，对里程计记录的里程计数据进行里程计误差处理，基于里程计误差处理后的里程计数据进行速度更新。
对于车辆姿态更新，可用如下的更新方程(1)进行：
Figure PCTCN2019103662-appb-000001
其中
Figure PCTCN2019103662-appb-000002
为k时刻的N系(导航坐标系)姿态四元数，
Figure PCTCN2019103662-appb-000003
为k-1时刻的N系姿态四元数，
Figure PCTCN2019103662-appb-000004
为k-1时刻到k时刻B系(机体坐标系)的姿态转换四元数。
其中，
Figure PCTCN2019103662-appb-000005
可基于B系相对N系的旋转角速度
Figure PCTCN2019103662-appb-000006
确定。tk-1时刻到tk时刻的对应的等效旋转矢量为：
Figure PCTCN2019103662-appb-000007
其中，
Figure PCTCN2019103662-appb-000008
表示B系相对I系(地心惯性坐标系)的角速度，
Figure PCTCN2019103662-appb-000009
表示B系相对N系的旋转角速度，
Figure PCTCN2019103662-appb-000010
表示B系到N系的坐标矩阵，且
Figure PCTCN2019103662-appb-000011
其中
Figure PCTCN2019103662-appb-000012
为地球自转角速度，为常值，
Figure PCTCN2019103662-appb-000013
表示N系相对E系(地球坐标系)的角速度的北向分量，且
Figure PCTCN2019103662-appb-000014
式中，
Figure PCTCN2019103662-appb-000015
表示惯导北向速度，r m表示地球长轴半径，h k-1表示k-1时刻的海拔高度，
Figure PCTCN2019103662-appb-000016
表示惯导东向速度，tan表示正切三角函数，Lon k-1表示k-1时刻的经度，r n表示地球短轴半径。
Figure PCTCN2019103662-appb-000017
可以利用等效旋转矢量的二子样算法进行计算：
Figure PCTCN2019103662-appb-000018
其中，A ib表示k-1时刻到k时刻的陀螺旋转角度，delta(A 1)和delta(A 2)是连续两次陀螺仪采样的角速度增量。
据此，可采用下式进行位置更新。
Figure PCTCN2019103662-appb-000019
Figure PCTCN2019103662-appb-000020
Figure PCTCN2019103662-appb-000021
其中，Lon k表示k时刻的经度，t表示时间，
Figure PCTCN2019103662-appb-000022
表示惯导东向速度，sec为正割三角函数，r n表示地球短轴半径，Lat指纬度，Lat k表示k时刻的纬度，r m表示地球长轴半径，
Figure PCTCN2019103662-appb-000023
表示惯导天线速度。以此为基础，可以得到DR的误差方程：
delta(P)＝p 1A+p 2delta(P)+p 3  (8)
delta(A)＝a 1A+a 2delta(P)+a 3  (9)
将上述DR的误差方程(8)、(9)带入上述各公式(1)至(7)，即可确定该误差方程中的各参数(包括p 1、p 2、p 3、a 1、a 2、a 3)的值，并应用在模型训练以及实际的定位过程中。基于上述DR的误差方程可见，当前时刻的位置误差delta(P)、惯导角度误差delta(A)，可以根据与当前时刻的惯导角度A以及上一时刻的惯导角度误差delta(P)确定。
如上所述，在航位推算过程中，可以不需要用到加速度计的信息。在本申请的一个实施例中，也可以使用加速度计测量得到的加速度进行姿态对准后，再执行上述航位推算的过程，这里的姿态对准，主要可以是进行俯仰角和横滚角对准，本实施例不做具体限定。一个实施例中，以车载导航系统为例，在车载导航系统使用MEMS陀螺仪时，则可以利用GPS的方位角来辅助对准，如可以结合加速度计测得的加速度和GPS的方位角进行姿态对准后，进行航位推算。
步骤302，对任意一个样本变化更新信息，通过初始融合定位模型中的各融合定位方式，分别对该样本变化更新信息进行融合定位，并获得各融合定位方式对该样本变化更新信息进行融合定位的融合定位性能参数。
可选地，该融合定位性能参数可以是包括任何可以用以评估融合定位的性能的参数，例如包括但不限于融合定位耗时、融合定位处理的资源占用量、定位精度等参数。
初始融合定位模型的输入，可以包括GPS定位信息和当前姿态变化信息，也可以是包括GPS定位信息、以及基于当前姿态变化信息进行航位推算得到的 航位推算位置信息。初始融合定位模型中，可以包含两个以上的融合定位方式。一个实施例中可包含粒子滤波融合定位方式和卡尔曼滤波融合定位方式，其他实施例中也可以是其他的融合定位方式或者是包含更多的融合定位方式，例如贝叶斯估计融合定位方式、模糊逻辑融合定位方式等等。出于简洁说明的目的，下述示例中是以同时包含粒子滤波融合定位方式和卡尔曼滤波融合定位方式进行举例说明。
以初始融合定位模型同时包含粒子滤波融合定位方式和卡尔曼滤波融合定位方式为例，一方面，采用粒子滤波融合定位方式对样本变化更新信息进行融合定位，获得采用粒子滤波融合定位方式对样本变化更新信息进行融合定位，确定的样本融合定位位置以及融合定位性能参数，具体的融合定位的过程可以基于粒子滤波融合算法的现有以及以后可能的任何方式进行。另一方面，同时采用卡尔曼滤波定位方式对样本变化更新信息进行融合定位，获得采用卡尔曼滤波融合定位方式对该样本变化更新信息进行融合定位确定的样本融合定位位置和融合定位性能参数。
在卡尔曼滤波算法中，包含下述五个基本方程中，其中，方程(10)、方程(11)为时间更新方程，方程(12)、方程(13)和方程(14)为状态更新方程。
X (k|k-1)＝FX (k-1|k-1)+BU (k)  (10)
P (k|k-1)＝FP (k-1|k-1)F T+Q  (11)
X (k|k)＝X (k|k-1)+Kg (k)(Z (k)-HX (k|k-1))  (12)
Kg (k)＝P (k|k-1)H T(HP (k|k-1)|H T+R) -1  (13)
P (k|k)＝(I-Kg (k)H)P (k-1|k-1)  (14)
如上述式(10)至(14)所述，在卡尔曼滤波算法中，存在两个参数Q、R，其中，参数Q表征了卡尔曼滤波算法模型的统计特性，体现了模型预测值的加权，在本申请实施例中称为预测值加权参数Q，参数R表征了测量过程中的测量噪声特性，体现了测量值的加权，本申请实施例中称为测量值加权参数R。预测值加权参数Q越大表示越信任测量值，测量值加权参数R越大表示越信任模型预测值。
其中，方程(14)中的I为1的矩阵，对于单模型单测量，I＝1。
在本实施例中，以上述前两个方程(10)、(11)作为预测方程，上述方程(12)、(13)、(14)作为更新方程，结合卡尔曼滤波算法的进行状态建模，建 模后的模型中可以包括用以进行卡尔曼滤波定位的相关参数，例如航位推算位置信息、航位推算误差等。一个实施例中建模后的模型可包含下述相关内容。
X＝[A dr,delta(P dr),E,delta(K),delta(A g),delta(P g)] T  (15)
Figure PCTCN2019103662-appb-000024
Z p＝P g-P dr
＝(P+delta(P g))-(P+delta(P dr))  (17)
＝delta(P g)-delta(P dr)
Z a＝delta(A dr)-delta(P dr)  (18)
Z＝[Z p Z a] T  (19)
Figure PCTCN2019103662-appb-000025
基于公式(15)可以确定，本申请中可以基于航位推算角度、航位推算位置误差、陀螺仪漂移、里程误差系数、GPS角度误差以及GPS位置误差，进行卡尔曼滤波融合定位。其中，A dr表示DR角度，delta(P dr)表示DR位置误差，E表示陀螺仪漂移，delta(K)表示DR里程误差系数，delta(A g)表示GPS角度误差，delta(P g)表示GPS位置误差。F1 3×3表示3行3列的矩阵F1，F2 3×3表示3行3列的矩阵F2，F3 3×3、F4 3×3、F5 3×3、F6 3×3以此类推，0 3×3表示3行3列的零矩阵，其中，零矩阵指矩阵中的所有元素均为0，
Figure PCTCN2019103662-appb-000026
表示3行3列的从B坐标系到N坐标系的转换矩阵，1 3×3表示3行3列的单位矩阵，其中，单位矩阵指矩阵中的所有对角线元素均为1，其他元素均为0，0 3×6表示3行6列的零矩阵，0 3×9表示3行9列的零矩阵，0 3×12表示3行12列的零矩阵，0 3×15表示3行15列的零矩阵，P表示当前位置信息，一个实施例的当前位置信息包括经度、纬度和高 度，Z p表示GPS定位信息和DR定位信息的位置差，Z a表示GPS定位信息和DR定位信息的角度差。
据此，在初始融合定位模型中包含卡尔曼滤波融合定位方式时，通过卡尔曼滤波融合定位方式对样本变化更新信息进行融合定位，可以包括下述步骤A1至步骤A3。
步骤A1：基于样本航位推算位置信息，根据如上建立的误差方程(8)、(9)，确定对应的样本航位推算误差。
步骤A2：通过初始融合定位模型中的卡尔曼滤波融合定位方式，基于样本GPS定位信息以及样本航位推算位置信息进行融合定位，获得卡尔曼滤波融合定位方式对应的初始融合定位位置、初始融合定位性能参数及样本航位推算误差修正值。
一个实施例中，在通过卡尔曼滤波融合定位方式对样本变化更新信息进行融合定位的过程中，可以以预测值加权参数Q为固定量，以测量值加权参数R为调节量进行调节。如可以基于GPS定位信息的相对误差DOP调节测量值加权参数R，从而可以在GPS信号不好的地方，使融合结果偏向于DR信息。
步骤A3：基于样本航位推算误差修正值修正初始融合定位位置，获得对应的样本融合定位位置以及对应的融合定位性能参数。
步骤303，基于融合定位性能参数，从各融合定位方式中确定各样本变化更新信息对应的优选融合定位方式。
其中，在基于融合定位性能参数确定优选融合定位方式时，基于融合定位性能参数具体包含的信息内容的不同，可以采用任何可能的方式进行。
以融合定位性能参数仅包含融合定位处理的资源占用量为例，则可以以最小的资源占用量对应的融合定位方式，作为该样本变化更新信息对应的优选融合定位方式。以融合定位性能参数包含两个以上的参数为例，则可以对这两个参数进行综合考虑(如加权处理之后)，从中选择加权后的最佳值(如最大值或最小值等)对应的融合定位方式，作为该样本变化更新信息对应的优选融合定位方式。在其他实施例中，也可以采用其他方式确定样本变化更新信息对应的优选融合定位方式。
步骤304，根据各样本变化更新信息分别对应的优选融合定位方式，确定出变化更新信息与融合定位方式分类结果之间的对应关系。
一个实施例中，可以结合所有的样本变化更新信息，确定出变化更新信息与融合定位方式分类结果之间的对应关系，如图4所示，其具体包括下述步骤401和步骤402。
步骤401，将各样本变化更新信息分别对应的优选融合定位方式，分别确定为该样本变化更新信息对应的样本融合定位方式分类结果。
即基于分类的思想，对应任意一个样本变化更新信息，将其对应的优选融合定位方式作为对该样本变化更新信息的分类。
步骤402，基于各样本变化更新信息及对应的样本融合定位方式分类结果，确定出变化更新信息与融合定位方式之间的对应关系。
在确定该对应关系时，可以采用任何可能的方式进行，例如曲线拟合等等，本实施例不做具体限定。
在另一个实施例中，可以仅基于大部分的样本变化更新信息，确定出变化更新信息与融合定位方式分类结果之间的对应关系。这里的大部分的样本变化更新信息，对应的优选融合定位方式相同，且在预定数目的样本变化更新信息中所占的数量最多。如图5所示，其具体包括下述步骤501至步骤503。
步骤501，确定各优选融合定位方式对应的样本变化更新信息的样本数目。
假设预定数目为N，初始定位模型中包含两个融合定位方式：卡尔曼滤波融合定位方式、粒子滤波融合定位方式，其中这N个样本变化更新信息中有K1个样本变化更新信息对应的优选融合定位方式均为卡尔曼滤波融合定位方式，则卡尔曼滤波融合定位方式对应的样本变化更新信息的样本数目为K1，有K2个样本变化更新信息对应的优选融合定位方式均为粒子滤波融合定位方式，则粒子滤波融合定位方式对应的样本变化更新信息的样本数目为K2。
步骤502，将值最大的样本数目对应的优选融合定位方式，确定为预定数目的样本变化更新信息对应的融合定位方式。
如上所述，如果K1大于K2，则将卡尔曼滤波融合定位方式确定为这N个样本变化更新信息对应的样本融合定位方式分类结果。
一个实施例中，也可以是将值最大、且与预定数目的比例超过预定比例的样本数目对应的优选融合定位方式，确定为预定数目的样本变化更新信息对应的融合定位方式分类结果。如上所述，如果K1大于K2，在同时满足K1/N大于或者等于预定比例时，再将卡尔曼滤波融合定位方式确定为这N个样本变化更新信息对应的样本融合定位方式分类结果。通过预定比例的设定，可以在某一 个的融合定位方式具有较大的优势时，才将该融合定位方式作为样本融合定位方式分类结果，以提升性能。
步骤503，基于预定数目的样本变化更新信息，或者最大样本数目的样本变化更新信息，以及融合定位方式分类结果，确定出变化更新信息与融合定位方式分类结果之间的对应关系。
可选地，在确定该对应关系时，可以采用任何可能的方式进行，例如曲线拟合等等，本实施例不做具体限定。
在上述训练完成后，在满足训练结束条件时，则可以基于确定出的变化更新信息与融合定位方式分类结果之间的对应关系，确定为融合定位模型。否则，更新初始融合定位模型，如更新初始融合定位模型中各输入参数特征相对应的权重，并进入下一轮训练过程。其中，训练结束条件可以结合实际需要进行设定，如可以基于样本GPS定位信息、各融合定位方式对应的样本融合定位位置以及融合定位性能参数分析是否收敛作为训练结束条件，本实施例中不对训练结束条件做具体限定。
基于上述训练过程获得的融合定位模型，确定了该融合定位模型的各个输入参数特征相对应的权重，即各个输入参数特征对应的权重可有所不同。从而在实际基于该融合定位模型进行定位时，基于变化更新信息对应的各输入参数特征对应的值以及各输入参数特征对应的权重，确定出该变化更新信息属于各融合定位方式的分类的概率，从而据此获得匹配的融合定位方式进行实际的融合定位。
如上训练得到的融合定位模型，可用于实际的融合定位的过程中。在该训练过程是由第三方(例如服务器)进行的情况下，该融合定位模型可以发送至各定位终端(如车载导航、移动终端等)，由各定位终端采用该融合定位模型完成融合定位的过程。也可以是将该融合定位模型存储到该第三方，在定位过程中，由该第三方基于终端上传的GPS定位信息和姿态变化信息(或基于姿态变化信息确定的惯性定位信息)完成融合定位过程，并将得到的最终定位位置返回给终端使用。在该训练过程由定位终端(如车载导航设备)进行的情况下，该定位终端可直接存储该融合定位模型，并应用在后续的定位过程中。
步骤305，获取当前变化更新信息，其中，当前变化更新信息包括：GPS定位信息以及航位推算位置信息，航位推算位置信息根据终端姿态变化信息确定的。
可选地，上述GPS定位信息指当前时刻通过GPS定位模块检测获得的GPS定位信息，例如GPS角度、以及GPS位置。上述当前姿态变化信息可以是通过惯性测量单元检测获得的相关参数确定的。
惯性测量单元是测量物体三轴姿态角(或角速率)以及加速度的装置，一个实施例中的惯性测量单元可包括加速度计和陀螺仪，加速度计是测量加速度的仪表，陀螺仪又叫角速度传感器，是测量物体偏转、倾斜时的转动角速度或者说用来传感方向的仪表，此时，对应的当前姿态变化信息可包括测量获得的加速度和角速度等参数。
步骤306，确定关于当前变化更新信息匹配的融合定位方式。
可选地，本实施例中，通过将当前变化更新信息输入融合定位模型，从而输出得到与该当前变化更新信息对应的融合定位方式。
步骤307，根据当前变化更新信息确定误差信息。
可选地，基于GPS定位信息以及航位推算位置信息确定误差信息。其中，误差信息可以包括：GPS定位误差，航位推算误差以及GPS定位信息与航位推算位置信息之间的定位信息间误差中的至少一种。
步骤308，根据融合定位方式，对当前变化更新信息中的GPS定位信息和航位推算信息结合误差信息进行融合定位，得到当前定位位置。
基于匹配的融合定位方式进行融合定位时，可以基于该匹配的融合定位方式的已有的融合定位过程进行，或者是基于对该融合定位方式加以改进后的过程进行，本实施例不做进行限定。在匹配的融合定位方式为第一融合定位方式(如卡尔曼滤波融合定位方式)时，结合上述确定的误差信息，对当前变化更新信息中对应的信息进行融合定位，确定当前定位位置，如一个实施例中可以结合上述确定的误差信息对融合定位得到的初始定位位置进行更新，从而获得最终的当前定位位置。
一个实施例中，以匹配的融合定位方式为卡尔曼滤波融合定位方式为例，可以采用卡尔曼滤波定位方式进行融合定位，获得最终的定位位置的方式，可以是目前已有以及以后可能出现的任何的卡尔曼滤波定位方式，此处不再详加赘述。如上所述，在进行融合定位处理时，可以以预测值加权参数Q为固定量，以测量值加权参数R为调节量，且该测量值加权参数R，可以基于GPS定位信息的相对误差进行调节。
应该理解的是，尽管上述实施例的说明中，给出了具体的几种融合定位算 法，但在实际技术实现过程中，还可以在融合定位模型中扩充或者增加其他的融合定位方式，并不能加以限定。
可选地，在上述根据匹配的融合定位方式进行融合定位，确定当前定位位置之后，还可以包括步骤：在自当前时刻之后的预定时间段内的各定位时刻，均采用该第一融合定位方式进行融合定位，确定各定位时刻对应的定位位置。从而可以在获得匹配的融合定位方式之后，可将该匹配的融合定位方式用于后续预定时间段内的各定位时刻的融合定位，无需时时动态进行融合定位方式的匹配，提高了处理效率，且提高了平滑性。
如在一个实施例中，在当前时刻获得匹配的融合定位方式(一个实施例中可以限定为是基于融合定位模型匹配出的融合定位方式)之后，自当前时刻开始的预定时间段内的各定位时刻，均采用该匹配融合定位方式进行融合定位，而在超过该预定时间段之后，则基于融合定位模型重新匹配出融合定位方式，以验证匹配出的融合定位方式是否还继续具有较优的融合定位性能，若连续一定次数N1或连续指定时长时间段T1内均不一致，则可以连续一定次数N2或连续指定时长时间段T2均基于融合定位模型重新匹配出融合定位方式，直至连续一定次数N3或连续指定时长时间段T3匹配出的融合定位方式均保持一致。其中，N1、N2、N3的值可以相同，也可以不相同，T1、T2、T3的值可以相同，也可以不相同。
一个实施例中，若连续一定次数N1或连续指定时长时间段T1内均不一致，也可以直接重新执行训练过程，从而训练出新的融合定位模型以进行应用。在另一个实施例中，也可以在当前时刻获得匹配的融合定位方式(一个实施例中可以限定为是基于融合定位模型匹配出的融合定位方式)之后，自当前时刻开始的预定时间段之后，也可以是直接重新执行训练过程，从而训练出新的融合定位模型新的融合定位模型以进行应用。
综上所述，本申请实施例提供的定位方法，通过获取GPS定位信息以及航位推算位置信息，基于这些信息确定匹配的融合定位方式以及误差信息，然后根据匹配的容和定位方式和误差信息进行融合定位得到当前定位位置，从而可以动态分类出匹配的融合定位方式来进行融合定位，可以有效提高融合定位的精度，能够为不同场景的性能提供较好的鲁棒性。
基于与上述定位方法相同的思想，一个实施例中提供一种定位装置，该装 置可应用于任何可以获得GPS信息和惯性测量单元测量获得的惯性传感参数，并快进行融合定位的设备，例如车载导航设备等。
参考图6所示，该定位装置包括获取模块610、确定模块620和定位模块630；
获取模块610，用于获取当前变化更新信息，所述当前变化更新信息中包括GPS定位信息和航位推算位置信息，所述航位推算位置信息是根据终端姿态变化信息确定的；
确定模块620，用于确定与所述当前变化更新信息匹配的融合定位方式，所述融合定位方式用于通过所述当前变化更新信息中的信息进行融合定位；
所述确定模块620，还用于根据所述当前变化更新信息确定误差信息，所述误差信息用于表示所述当前变化更信息的误差情况；
定位模块630，用于根据所述融合定位方式，对当前变化更新信息中的所述GPS定位信息和所述航位推算位置信息结合所述误差信息进行融合定位，得到当前定位位置。
在一个可选的实施例中，所述确定模块620，还用于对里程计记录的里程计数据进行里程计误差处理，基于里程计误差处理后的里程计数据进行速度更新之后，根据所述当前姿态变化信息进行航位推算获得所述航位推算位置信息，所述姿态变化信息中包括速度信息；
所述确定模块620，还用于根据陀螺旋转速度以及所述陀螺仪连续两次采集的角速度增量进行姿态变化信息的更新；根据更新后的所述当前姿态变化信息进行航位推算获得所述航位推算位置信息；
所述确定模块620，还用于基于加速度计采集获得的加速度进行姿态对准之后，根据所述当前姿态变化信息进行航位推算获得所述航位推算位置信息；
所述确定模块620，还用于基于GPS方位角、以及所述加速度计采集获得的加速度进行姿态对准之后，根据所述当前姿态变化信息进行航位推算获得所述航位推算位置信息。
在一个可选的实施例中，所述确定模块620，还用于基于当前变化更新信息确定模型输入特征；将所述模型输入特征输入训练获得的融合定位模型，确定 与所述当前变化更新信息匹配的所述融合定位方式。
在一个可选的实施例中，所述确定模块620，还用于在自当前时刻之后的预定时间段内的各定位时刻，采用所述融合定位方式进行所述融合定位，确定各定位时刻对应的定位位置。
在一个可选的实施例中，所述定位模块630，还用于在当前时刻之前最近一次匹配的融合定位方式之后的预定时间段内，应用所述融合定位方式进行所述融合定位。
在一个可选的实施例中，所述融合定位方式包括卡尔曼滤波融合定位方式和粒子滤波融合定位方式中的任意一种。
在一个可选的实施例中，所述获取模块610，还用于获取预定数目的样本变化更新信息；对任意一个样本变化更新信息，通过所述初始融合定位模型中的各融合定位方式，分别对该样本变化更新信息进行融合定位，并获得各融合定位方式对该样本变化更新信息进行融合定位的融合定位性能参数；
所述确定模块620，还用于基于融合定位性能参数，从各融合定位方式中确定各样本变化更新信息对应的优选融合定位方式；根据各所述样本变化更新信息分别对应的优选融合定位方式，确定出变化更新信息与各融合定位方式之间的对应关系。
在一个可选的实施例中，所述确定模块620，还用于将各所述样本变化更新信息分别对应的优选融合定位方式，分别确定为该样本变化更新信息对应的样本融合定位方式分类结果；基于各所述样本变化更新信息及对应的样本融合定位方式分类结果，确定出变化更新信息与各融合定位方式之间的对应关系。
在一个可选的实施例中，所述确定模块620，还用于确定各优选融合定位方式对应的样本变化更新信息的样本数目；将值最大的样本数目对应的优选融合定位方式，确定为所述预定数目的样本变化更新信息对应的融合定位方式；基于所述预定数目的样本变化更新信息或者最大样本数目的样本变化更新信息，以及融合定位方式分类结果，确定出变化更新信息与融合定位方式分类结果之间的对应关系。
在一个可选的实施例中，所述确定模块620，还用于在满足训练结束条件时，基于确定出的变化更新信息与融合定位方式分类结果之间的对应关系，确定所述融合定位模型，否则，更新所述初始融合定位模型，进入下一轮训练过程。
综上所述，本申请实施例提供的定位装置，通过获取GPS定位信息以及航位推算位置信息，基于这些信息确定匹配的融合定位方式以及误差信息，然后根据匹配的容和定位方式和误差信息进行融合定位得到当前定位位置，从而可以动态分类出匹配的融合定位方式来进行融合定位，可以有效提高融合定位的精度，能够为不同场景的性能提供较好的鲁棒性。
一个实施例中提供了一种定位设备700，该定位设备700可以是任何能够结合惯性测量单元750测得的惯性传感参数进行定位处理的设备，例如终端或者服务器，这里的终端可以是任何可以进行惯性测量的终端，例如车载导航设备，服务器可以是任何可以获得惯性测量单元750得到的惯性传感参数并据此进行融合定位的服务器。一个实施例中的定位设备700的内部结构图可以如图7所示。该计算机设备包括通过系统总线连接的处理器710、存储器720，还可以包括通过系统总线连接的网络接口730，在该计算机设备为终端设备时，还可以包括显示屏和输入装置740。在定位设备700为车载导航设备等可以自行测量惯性传感参数的设备时，其还可以包括通过系统总线连接的惯性测量单元750、GPS定位设备770、加速度计和里程计760等设备。
其中，该计算机设备的处理器710用于提供计算和控制能力。该计算机设备的存储器720包括非易失性存储介质721、内存储器722。该非易失性存储介质721存储有操作系统7211和计算机程序7212。该内存储器722为非易失性存储介质721中的操作系统7211和计算机程序7212以及计算机程序7221的运行提供环境。该计算机设备的网络接口730用于与外部的终端通过网络连接通信。该计算机程序被处理器710执行时以实现一种定位方法。该计算机设备的显示屏可以是液晶显示屏或者电子墨水显示屏，该计算机设备的输入装置740可以是显示屏上覆盖的触摸层，也可以是计算机设备外壳上设置的按键、轨迹球或触控板，还可以是外接的键盘、触控板或鼠标等。
本领域技术人员可以理解，图7中示出的结构，仅仅是与本申请方案相关的部分结构的框图，并不构成对本申请方案所应用于其上的惯性导航设备的限定，具体的惯性导航设备可以包括比图中所示更多或更少的部件，或者组合某 些部件，或者具有不同的部件布置。
据此，在一个实施例中，还提供了一种惯性导航设备，包括存储器720和处理器710，存储器720中存储有计算机程序，该处理器710执行计算机程序时实现如上所述的任意实施例中的方法的步骤。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过计算机程序来指令相关的硬件来完成，所述的计算机程序可存储于一非易失性计算机可读取存储介质中，该计算机程序在执行时，可包括如上述各方法的实施例的流程。其中，本申请所提供的各实施例中所使用的对存储器、存储、数据库或其它介质的任何引用，均可包括非易失性和/或易失性存储器。非易失性存储器可包括只读存储器(ROM)、可编程ROM(PROM)、电可编程ROM(EPROM)、电可擦除可编程ROM(EEPROM)或闪存。易失性存储器可包括随机存取存储器(RAM)或者外部高速缓冲存储器。作为说明而非局限，RAM以多种形式可得，诸如静态RAM(SRAM)、动态RAM(DRAM)、同步DRAM(SDRAM)、双数据率SDRAM(DDRSDRAM)、增强型SDRAM(ESDRAM)、同步链路(Synchlink)DRAM(SLDRAM)、存储器总线(Rambus)直接RAM(RDRAM)、直接存储器总线动态RAM(DRDRAM)、以及存储器总线动态RAM(RDRAM)等。
以上实施例的各技术特征可以进行任意的组合，为使描述简洁，未对上述实施例中的各个技术特征所有可能的组合都进行描述，然而，只要这些技术特征的组合不存在矛盾，都应当认为是本说明书记载的范围。
以上所述实施例仅表达了本申请的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对发明专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在不脱离本申请构思的前提下，还可以做出若干变形和改进，这些都属于本申请的保护范围。因此，本申请专利的保护范围应以所附权利要求为准。

Claims (22)

  1. 一种定位方法，应用于终端中，所述方法包括：
    获取当前变化更新信息，所述当前变化更新信息中包括GPS定位信息和航位推算位置信息，所述航位推算位置信息是根据终端姿态变化信息确定的；
    确定与所述当前变化更新信息匹配的融合定位方式，所述融合定位方式用于通过所述当前变化更新信息中的信息进行融合定位；
    根据所述当前变化更新信息确定误差信息，所述误差信息用于表示所述当前变化更新信息的误差情况；
    根据所述融合定位方式，对当前变化更新信息中的所述GPS定位信息和所述航位推算位置信息结合所述误差信息进行融合定位，得到当前定位位置。
  2. 根据权利要求1所述的方法，其中所述航位推算位置信息的确定方式包括如下方式中的至少一种：
    对里程计记录的里程计数据进行里程计误差处理，基于里程计误差处理后的里程计数据进行速度更新之后，根据所述当前姿态变化信息进行航位推算获得所述航位推算位置信息，所述姿态变化信息中包括速度信息；
    根据陀螺旋转速度以及所述陀螺仪连续两次采集的角速度增量进行姿态变化信息的更新；根据更新后的所述当前姿态变化信息进行航位推算获得所述航位推算位置信息；
    基于加速度计采集获得的加速度进行姿态对准之后，根据所述当前姿态变化信息进行航位推算获得所述航位推算位置信息；
    基于GPS方位角、以及所述加速度计采集获得的加速度进行姿态对准之后，根据所述当前姿态变化信息进行航位推算获得所述航位推算位置信息。
  3. 根据权利要求1所述的方法，其中所述确定与所述当前变化更新信息匹配的融合定位方式，包括：
    基于当前变化更新信息确定模型输入特征；
    将所述模型输入特征输入训练获得的融合定位模型，确定与所述当前变化更新信息匹配的所述融合定位方式。
  4. 根据权利要求1所述的方法，其中所述得到当前定位位置之后，还包括：
    在自当前时刻之后的预定时间段内的各定位时刻，采用所述融合定位方式进行所述融合定位，确定各定位时刻对应的定位位置。
  5. 根据权利要求1所述的方法，其中所述确定与所述当前变化更新信息匹配的融合定位方式之后，还包括：
    在当前时刻之前最近一次匹配的融合定位方式之后的预定时间段内，应用所述融合定位方式进行所述融合定位。
  6. 根据权利要求1至5任一所述的方法，其中
    所述融合定位方式包括卡尔曼滤波融合定位方式和粒子滤波融合定位方式中的任意一种。
  7. 根据权利要求3所述的方法，其中所述将所述模型输入特征输入训练获得的融合定位模型之前，包括：
    获取预定数目的样本变化更新信息；
    对任意一个样本变化更新信息，通过所述初始融合定位模型中的各融合定位方式，分别对该样本变化更新信息进行融合定位，并获得各融合定位方式对该样本变化更新信息进行融合定位的融合定位性能参数；
    基于融合定位性能参数，从各融合定位方式中确定各样本变化更新信息对应的优选融合定位方式；
    根据各所述样本变化更新信息分别对应的优选融合定位方式，确定出变化更新信息与各融合定位方式之间的对应关系。
  8. 根据权利要求7所述的方法，其中所述根据各所述样本变化更新信息分别对应的优选融合定位方式，确定出变化更新信息与各融合定位方式之间的对应关系，包括：
    将各所述样本变化更新信息分别对应的优选融合定位方式，分别确定为该样本变化更新信息对应的样本融合定位方式分类结果；
    基于各所述样本变化更新信息及对应的样本融合定位方式分类结果，确定出变化更新信息与各融合定位方式之间的对应关系。
  9. 根据权利要求7所述的方法，其中所述根据各所述样本变化更新信息分别对应的优选融合定位方式，确定出变化更新信息与各融合定位方式之间的对应关系，包括：
    确定各优选融合定位方式对应的样本变化更新信息的样本数目；
    将值最大的样本数目对应的优选融合定位方式，确定为所述预定数目的样本变化更新信息对应的融合定位方式；
    基于所述预定数目的样本变化更新信息或者最大样本数目的样本变化更新信息，以及融合定位方式分类结果，确定出变化更新信息与融合定位方式分类结果之间的对应关系。
  10. 根据权利要求7所述的方法，其中确定出变化更新信息与融合定位方式分类结果之间的对应关系之后，还包括：
    在满足训练结束条件时，基于确定出的变化更新信息与融合定位方式分类结果之间的对应关系，确定所述融合定位模型，否则，更新所述初始融合定位模型，进入下一轮训练过程。
  11. 一种定位装置，应用于终端中，所述装置包括：
    获取模块，用于获取当前变化更新信息，所述当前变化更新信息中包括GPS 定位信息和航位推算位置信息，所述航位推算位置信息是根据终端姿态变化信息确定的；
    确定模块，用于确定与所述当前变化更新信息匹配的融合定位方式，所述融合定位方式用于通过所述当前变化更新信息中的信息进行融合定位；
    所述确定模块，还用于根据所述当前变化更新信息确定误差信息，所述误差信息用于表示所述当前变化更信息的误差情况；
    定位模块，用于根据所述融合定位方式，对当前变化更新信息中的所述GPS定位信息和所述航位推算位置信息结合所述误差信息进行融合定位，得到当前定位位置。
  12. 根据权利要求11所述的装置，其中所述确定模块，还用于对里程计记录的里程计数据进行里程计误差处理，基于里程计误差处理后的里程计数据进行速度更新之后，根据所述当前姿态变化信息进行航位推算获得所述航位推算位置信息，所述姿态变化信息中包括速度信息；
    所述确定模块，还用于根据陀螺旋转速度以及所述陀螺仪连续两次采集的角速度增量进行姿态变化信息的更新；根据更新后的所述当前姿态变化信息进行航位推算获得所述航位推算位置信息；
    所述确定模块，还用于基于加速度计采集获得的加速度进行姿态对准之后，根据所述当前姿态变化信息进行航位推算获得所述航位推算位置信息；
    所述确定模块，还用于基于GPS方位角、以及所述加速度计采集获得的加速度进行姿态对准之后，根据所述当前姿态变化信息进行航位推算获得所述航位推算位置信息。
  13. 根据权利要求11所述的装置，其中所述确定模块，还用于基于当前变化更新信息确定模型输入特征；将所述模型输入特征输入训练获得的融合定位模型，确定与所述当前变化更新信息匹配的所述融合定位方式。
  14. 根据权利要求11所述的装置，其中所述确定模块，还用于在自当前时刻之后的预定时间段内的各定位时刻，采用所述融合定位方式进行所述融合定位，确定各定位时刻对应的定位位置。
  15. 根据权利要求11所述的装置，其中所述定位模块，还用于在当前时刻之前最近一次匹配的融合定位方式之后的预定时间段内，应用所述融合定位方式进行所述融合定位。
  16. 根据权利要求11至15任一所述的装置，其中
    所述融合定位方式包括卡尔曼滤波融合定位方式和粒子滤波融合定位方式中的任意一种。
  17. 根据权利要求13所述的装置，其中所述获取模块，还用于获取预定数目的样本变化更新信息；对任意一个样本变化更新信息，通过所述初始融合定位模型中的各融合定位方式，分别对该样本变化更新信息进行融合定位，并获得各融合定位方式对该样本变化更新信息进行融合定位的融合定位性能参数；
    所述确定模块，还用于基于融合定位性能参数，从各融合定位方式中确定各样本变化更新信息对应的优选融合定位方式；根据各所述样本变化更新信息分别对应的优选融合定位方式，确定出变化更新信息与各融合定位方式之间的对应关系。
  18. 根据权利要求17所述的装置，其中所述确定模块，还用于将各所述样本变化更新信息分别对应的优选融合定位方式，分别确定为该样本变化更新信息对应的样本融合定位方式分类结果；基于各所述样本变化更新信息及对应的样本融合定位方式分类结果，确定出变化更新信息与各融合定位方式之间的对应关系。
  19. 根据权利要求17所述的装置，其中所述确定模块，还用于确定各优选融合定位方式对应的样本变化更新信息的样本数目；将值最大的样本数目对应的优选融合定位方式，确定为所述预定数目的样本变化更新信息对应的融合定位方式；基于所述预定数目的样本变化更新信息或者最大样本数目的样本变化更新信息，以及融合定位方式分类结果，确定出变化更新信息与融合定位方式分类结果之间的对应关系。
  20. 根据权利要求17所述的装置，其中所述确定模块，还用于在满足训练结束条件时，基于确定出的变化更新信息与融合定位方式分类结果之间的对应关系，确定所述融合定位模型，否则，更新所述初始融合定位模型，进入下一轮训练过程。
  21. 一种定位设备，包括存储器和处理器，所述存储器存储有计算机程序，其中所述处理器执行所述计算机程序时实现权利要求1至10中任一项所述方法的步骤。
  22. 一种计算机可读存储介质，其上存储有计算机程序，其中所述计算机程序被处理器执行时实现权利要求1至10中任一项所述的方法的步骤。
PCT/CN2019/103662 2018-09-04 2019-08-30 定位方法、装置、设备和计算机可读存储介质 WO2020048394A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP19857105.1A EP3848730A4 (en) 2018-09-04 2019-08-30 POSITIONING PROCESS, APPARATUS AND DEVICE, AND COMPUTER READABLE STORAGE MEDIA
US17/075,880 US11796320B2 (en) 2018-09-04 2020-10-21 Positioning method, apparatus and device, and computer-readable storage medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811025588.0A CN110873888B (zh) 2018-09-04 2018-09-04 定位方法、定位装置、定位设备和计算机存储介质
CN201811025588.0 2018-09-04

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/075,880 Continuation US11796320B2 (en) 2018-09-04 2020-10-21 Positioning method, apparatus and device, and computer-readable storage medium

Publications (1)

Publication Number Publication Date
WO2020048394A1 true WO2020048394A1 (zh) 2020-03-12

Family

ID=69716012

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/103662 WO2020048394A1 (zh) 2018-09-04 2019-08-30 定位方法、装置、设备和计算机可读存储介质

Country Status (4)

Country Link
US (1) US11796320B2 (zh)
EP (1) EP3848730A4 (zh)
CN (1) CN110873888B (zh)
WO (1) WO2020048394A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111836359A (zh) * 2020-07-01 2020-10-27 Oppo广东移动通信有限公司 多径场景的定位方法、装置、电子设备以及存储介质
CN113406682A (zh) * 2021-06-22 2021-09-17 腾讯科技（深圳）有限公司 一种定位方法、装置、电子设备及存储介质
CN113984044A (zh) * 2021-10-08 2022-01-28 杭州鸿泉物联网技术股份有限公司 一种基于车载多感知融合的车辆位姿获取方法及装置

Families Citing this family (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI752366B (zh) * 2019-10-29 2022-01-11 優式機器人股份有限公司 位移處理方法及模組與除草裝置
CN111859177A (zh) * 2020-04-10 2020-10-30 北京嘀嘀无限科技发展有限公司 一种定位方法和系统
CN111879305B (zh) * 2020-06-16 2022-03-18 华中科技大学 一种面向高危生产环境的多模态感知定位模型与系统
CN111780756A (zh) * 2020-07-20 2020-10-16 北京百度网讯科技有限公司 车辆航位推算方法、装置、设备以及存储介质
CN112468959B (zh) * 2020-11-24 2023-05-16 宏景科技股份有限公司 位置确定方法、装置、计算机设备和存储介质
CN112556696B (zh) * 2020-12-03 2022-01-07 腾讯科技（深圳）有限公司 一种对象定位方法、装置、计算机设备以及存储介质
US11859978B2 (en) 2021-01-15 2024-01-02 ALMA Technologies Ltd. System and method for estimating a velocity of a vehicle using inertial sensors
US20220228866A1 (en) * 2021-01-15 2022-07-21 ALMA Technologies Ltd. System and method for providing localization using inertial sensors
CN112596089B (zh) * 2021-03-02 2021-05-18 腾讯科技（深圳）有限公司 融合定位方法、装置、电子设备及存储介质
CN113175936B (zh) * 2021-06-28 2021-08-31 智道网联科技(北京)有限公司 基于神经网络模型的车辆轨迹推算方法及装置
CN113568430A (zh) * 2021-07-31 2021-10-29 北京晟天行科技有限公司 一种无人机机翼执行数据的修正控制方法
CN114167858B (zh) * 2021-11-12 2022-11-11 广州文远知行科技有限公司 车辆的导航控制方法、装置、导航控制器和车辆控制系统
CN114370870B (zh) * 2022-01-05 2024-04-12 中国兵器工业计算机应用技术研究所 适用于位姿测量卡尔曼滤波的滤波器更新信息筛选方法
CN117031521B (zh) * 2023-10-08 2024-01-30 山东大学 一种室内外无缝环境下的弹性融合定位方法及系统
CN117433540B (zh) * 2023-12-20 2024-04-02 北京海兰信数据科技股份有限公司 船舶的航行方法、终端设备及可读存储介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1908587A (zh) * 2006-08-18 2007-02-07 湖南大学 Gps/dr车载组合定位系统及定位方法
CN102778686A (zh) * 2012-08-07 2012-11-14 东南大学 基于移动gps/ins节点的协同车辆定位方法
CN104502942A (zh) * 2015-01-06 2015-04-08 上海华测导航技术有限公司 基于卫星导航和航位推测实现农业机械定位的系统及方法
US20150142311A1 (en) * 2013-11-18 2015-05-21 General Motors Llc Vehicular GPS/DR Navigation with Environmental-Adaptive Kalman Filter Gain
CN104819716A (zh) * 2015-04-21 2015-08-05 北京工业大学 一种基于mems的ins/gps组合的室内外个人导航算法
CN108303721A (zh) * 2018-02-12 2018-07-20 北京经纬恒润科技有限公司 一种车辆定位方法及系统

Family Cites Families (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6240367B1 (en) * 1998-11-27 2001-05-29 Ching-Fang Lin Full fusion positioning method for vehicle
US20020158796A1 (en) * 2001-04-25 2002-10-31 Ian Humphrey Integrated GPS and IGS system and method
CN101907714B (zh) 2010-06-25 2013-04-03 陶洋 基于多传感器数据融合的gps辅助定位方法
CN103810382B (zh) * 2014-01-27 2017-01-25 中国电子科技集团公司第十研究所 机载分布式多传感器两级数据融合策略选择方法
DE102014211164A1 (de) * 2014-06-11 2015-12-17 Continental Teves Ag & Co. Ohg Verfahren und System zur Anpassung eines Navigationssystems
US9836056B2 (en) * 2015-06-05 2017-12-05 Bao Tran Smart vehicle
US10571926B1 (en) * 2016-08-29 2020-02-25 Trifo, Inc. Autonomous platform guidance systems with auxiliary sensors and obstacle avoidance
US10571925B1 (en) * 2016-08-29 2020-02-25 Trifo, Inc. Autonomous platform guidance systems with auxiliary sensors and task planning
US11789449B2 (en) * 2018-02-09 2023-10-17 Nvidia Corporation Controlling autonomous vehicles using safe arrival times
US11099558B2 (en) * 2018-03-27 2021-08-24 Nvidia Corporation Remote operation of vehicles using immersive virtual reality environments
WO2019200385A1 (en) * 2018-04-13 2019-10-17 Salandy Defour Ricardo Scott Proximity-based event networking system and wearable augmented reality clothing

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1908587A (zh) * 2006-08-18 2007-02-07 湖南大学 Gps/dr车载组合定位系统及定位方法
CN102778686A (zh) * 2012-08-07 2012-11-14 东南大学 基于移动gps/ins节点的协同车辆定位方法
US20150142311A1 (en) * 2013-11-18 2015-05-21 General Motors Llc Vehicular GPS/DR Navigation with Environmental-Adaptive Kalman Filter Gain
CN104502942A (zh) * 2015-01-06 2015-04-08 上海华测导航技术有限公司 基于卫星导航和航位推测实现农业机械定位的系统及方法
CN104819716A (zh) * 2015-04-21 2015-08-05 北京工业大学 一种基于mems的ins/gps组合的室内外个人导航算法
CN108303721A (zh) * 2018-02-12 2018-07-20 北京经纬恒润科技有限公司 一种车辆定位方法及系统

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111836359A (zh) * 2020-07-01 2020-10-27 Oppo广东移动通信有限公司 多径场景的定位方法、装置、电子设备以及存储介质
CN113406682A (zh) * 2021-06-22 2021-09-17 腾讯科技（深圳）有限公司 一种定位方法、装置、电子设备及存储介质
CN113406682B (zh) * 2021-06-22 2024-03-12 腾讯科技（深圳）有限公司 一种定位方法、装置、电子设备及存储介质
CN113984044A (zh) * 2021-10-08 2022-01-28 杭州鸿泉物联网技术股份有限公司 一种基于车载多感知融合的车辆位姿获取方法及装置

Also Published As

Publication number Publication date
EP3848730A4 (en) 2021-11-03
CN110873888A (zh) 2020-03-10
US20210033738A1 (en) 2021-02-04
US11796320B2 (en) 2023-10-24
EP3848730A1 (en) 2021-07-14
CN110873888B (zh) 2022-05-06

Similar Documents

Publication Publication Date Title
WO2020048394A1 (zh) 定位方法、装置、设备和计算机可读存储介质
CN109059907B (zh) 轨迹数据处理方法、装置、计算机设备和存储介质
CN111562603B (zh) 基于航位推算的导航定位方法、设备及存储介质
WO2017150106A1 (ja) 車載装置、及び、推定方法
WO2010111118A1 (en) Method and apparatus for improving gps receiver accuracy using an embedded map database
JP2017146279A (ja) 推定装置、推定方法及び推定プログラム
US9389316B2 (en) Elevation assistance for location determination
US10197396B2 (en) Always on compass calibration system and methods
CN103941273A (zh) 机载惯性/卫星组合导航系统的自适应滤波方法与滤波器
CN114545472B (zh) 一种gnss/ins组合系统的导航方法和装置
CN114061570A (zh) 车辆定位方法、装置、计算机设备和存储介质
CN111707294B (zh) 基于最优区间估计的行人导航零速区间检测方法和装置
CN110989619B (zh) 用于定位对象的方法、装置、设备和存储介质
CN114413934A (zh) 一种车辆定位系统校正方法和装置
CN113566850B (zh) 惯性测量单元的安装角度标定方法、装置和计算机设备
US20220057517A1 (en) Method for constructing point cloud map, computer device, and storage medium
US11805390B2 (en) Method, apparatus, and computer program product for determining sensor orientation
EP3425339A1 (en) Position estimating device, position estimating method and program
CN112985391B (zh) 一种基于惯性和双目视觉的多无人机协同导航方法和装置
CN113074757B (zh) 车载惯导安装误差角的标定方法
US20180364047A1 (en) Estimation device, estimation method, and non-transitory computer-readable recording medium
CN106646569A (zh) 一种导航定位方法及设备
CN114001730B (zh) 融合定位方法、装置、计算机设备和存储介质
CN110375740B (zh) 车辆导航方法、装置、设备和存储介质
CN109945864B (zh) 室内行车定位融合方法、装置、存储介质及终端设备

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19857105

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2019857105

Country of ref document: EP

Effective date: 20210406