CN112577521A - Combined navigation error calibration method and electronic equipment - Google Patents

Combined navigation error calibration method and electronic equipment Download PDF

Info

Publication number
CN112577521A
CN112577521A CN202011350339.6A CN202011350339A CN112577521A CN 112577521 A CN112577521 A CN 112577521A CN 202011350339 A CN202011350339 A CN 202011350339A CN 112577521 A CN112577521 A CN 112577521A
Authority
CN
China
Prior art keywords
data
gnss
error
carrier
inputting
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202011350339.6A
Other languages
Chinese (zh)
Other versions
CN112577521B (en
Inventor
赵方
吴凡
罗海勇
高喜乐
包林封
龚依林
肖逸敏
陆爽秋
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing University of Posts and Telecommunications
Original Assignee
Beijing University of Posts and Telecommunications
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 Beijing University of Posts and Telecommunications filed Critical Beijing University of Posts and Telecommunications
Priority to CN202011350339.6A priority Critical patent/CN112577521B/en
Publication of CN112577521A publication Critical patent/CN112577521A/en
Application granted granted Critical
Publication of CN112577521B publication Critical patent/CN112577521B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C25/00—Manufacturing, calibrating, cleaning, or repairing instruments or devices referred to in the other groups of this subclass
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/10—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration
    • G01C21/12—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning
    • G01C21/16—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning by integrating acceleration or speed, i.e. inertial navigation
    • G01C21/165—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning by integrating acceleration or speed, i.e. inertial navigation combined with non-inertial navigation instruments
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
    • G01C21/28—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network with correlation of data from several navigational instruments
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
    • G01C21/34—Route searching; Route guidance
    • G01C21/3446—Details of route searching algorithms, e.g. Dijkstra, A*, arc-flags, using precalculated routes
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Manufacturing & Machinery (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Navigation (AREA)

Abstract

One or more embodiments of the present specification provide a combined navigation error calibration method and an electronic device, where the method includes: importing accelerometer data and gyroscope data into a mechanical arrangement part to obtain position data, speed data and attitude data; updating the speed, the position and the attitude of a carrier in different modes according to whether the real-time dynamic carrier phase difference technology RTK obtains the GNSS information of the global positioning system; obtaining the covariance of NHC measurement noise through the trained residual attention network model; and inputting the NHC measurement noise covariance into a Kalman filter module to obtain a position estimation value of the carrier. The invention corrects the combined navigation error, and effectively improves the accuracy, robustness and practicability of the combined navigation.

Description

Combined navigation error calibration method and electronic equipment
Technical Field
One or more embodiments of the present disclosure relate to the field of integrated navigation technologies, and in particular, to an integrated navigation error calibration method and an electronic device.
Background
In recent years, with the rapid rise of unmanned technology, the great popularization of location services, the development of mobile mapping technology and the increase of high-precision navigation requirements, a high-reliability autonomous positioning navigation system is urgently needed. In the aspect of industrial requirements, the autonomous navigation system can be used for moving measurement results, provides a precise pose scheme, and brings revolutionary changes to the surveying and mapping industry; in the aspect of public demands, intelligent carriers represented by smart phones, unmanned planes, automatic driving automobiles and mobile robots highly depend on precise position information when autonomous movement is carried out, and an autonomous navigation system is the basis and the core of environment perception and decision control.
A Global Navigation Satellite System (GNSS) is a common outdoor positioning method, but because GNSS positioning has a very high requirement on the environment, in a complex environment (such as a city, an overpass, a tunnel, a shade road, and the like), GNSS satellite signals are shielded by a shielding object, the signal strength is weakened, observation noise is greatly increased, and effective data is lost, so that accurate positioning cannot be provided by GNSS basically. The Inertial Navigation System (INS) can generate navigation information (position, carrier attitude, real-time movement velocity, etc.) by integration without external force and by means of raw data of inertial devices (sensors such as gyros, accelerometers, etc.). Because it does not rely on external force, so can apply in all complicated environment. However, the positioning error increases with time, and the required accuracy cannot be achieved when long-time positioning is required. The prediction results of the GNSS and the INS are fused through Kalman filtering, interference and loss of GNSS signals can be prevented, and errors of the INS can be corrected.
However, there are still certain problems with combined navigation systems (GNSS/INS). Firstly, the convergence speed and the positioning accuracy of the system greatly depend on a zero offset calibration result of an Inertial Measurement Unit (IMU) device; the zero offset of the IMU can be calibrated through a turntable test, but under the condition that the condition is not allowed, the calibration process is very complicated and depends on hardware equipment; secondly, the process noise covariance matrix (including random walk coefficients, zero-offset standard deviation, etc.) of the INS system is difficult to determine and may change with the environment, and the existing integrated navigation system reduces the parameter to a fixed value, which may affect the result of integrated navigation and even cause filter divergence. Meanwhile, stall detection is the most basic and critical to limit the increase of errors by using stall information of each stationary point, but erroneous detection may cause performance degradation. In addition to this, the determination of the measurement noise also has a great influence on the combined navigation performance.
Therefore, a calibration method for combined navigation GNSS/INS errors is needed to improve the accuracy, robustness and practicality of combined navigation.
Disclosure of Invention
In view of the above, an object of one or more embodiments of the present disclosure is to provide a method and an electronic device for calibrating a combined navigation error, so as to improve accuracy, robustness and practicability of combined navigation.
In view of the above, one or more embodiments of the present specification provide a combined navigation error calibration method, including:
importing accelerometer data and gyroscope data of a carrier into a mechanical layout part to obtain position data, speed data and posture data of the carrier;
when the global satellite navigation GNSS information of the carrier is obtained through a real-time dynamic carrier phase differential technology RTK, the position data and the speed data are imported into a GNSS Kalman filtering module to obtain a state estimation matrix and a state error matrix through calculation, and the state estimation matrix, the state error matrix and the speed data are input into the GNSS Kalman filtering module to update the speed data, the position data and the attitude data of the carrier;
when the GNSS information of the carrier is not obtained through RTK, the speed data is input into the Kalman filtering module so as to update the speed data, the position data and the attitude data of the carrier based on an initial state estimation matrix and a state error matrix;
storing the accelerometer data, the gyroscope data, the speed data and the attitude data in a sliding window at intervals of a preset time period, inputting the data in the sliding window into a trained residual attention network model as test data to obtain vehicle-mounted constraint NHC measurement noise covariance, wherein the sliding window limits the data input into the residual attention network model each time;
and inputting the NHC measurement noise covariance into the NHC Kalman filtering module to obtain a position estimation value of the carrier.
Further, when the GNSS information of the carrier is not obtained by RTK, the inputting the velocity data into a kalman filtering module to update the velocity data, the position data, and the attitude data of the carrier based on an initial state estimation matrix and a state error matrix may further include:
storing the speed data and the attitude data in a sliding window, and inputting the speed data and the attitude data serving as training data into a time convolution neural network TCN model to obtain a process noise covariance and an NHC (positive temperature coefficient) measurement noise covariance;
substituting the process noise covariance into Kalman filtering, substituting the NHC measurement noise covariance into a Kalman filtering module to obtain a position estimation value of the carrier, feeding back the difference between the position estimation value of the carrier and the true position of the carrier as a loss function to the mechanical arrangement part, and finishing the training of the TCN model;
and inputting the data in the sliding window into a TCN model after training as test data to obtain the process noise covariance and the NHC measurement noise covariance, inputting the process noise covariance into a Kalman filtering module, inputting the NHC measurement noise covariance into the NHC Kalman filtering module, and outputting the position estimation value of the carrier.
Further, the method further comprises:
an error Learning module MT-e & R extracts effective characteristics from NMEA-0183 protocol information and integrates the effective characteristics into characteristic data in a time series form, wherein the MT-e & R comprises an e-Learning task and an R-Learning task;
inputting the characteristic data into an e-learning task to pre-train network parameters so as to obtain positioning errors of the original position measurement of the GNSS in the x direction and the y direction under a geodetic coordinate system e, and recording the positioning errors as positioning errors
Figure BDA0002801071070000031
Inputting the characteristic data into the R-Learning task to obtain a GNSS measurement noise covariance R;
the e-Learning task utilizes
Figure BDA0002801071070000041
Calculating a loss function according to the GNSS position true value y, and adjusting network parameters through the loss function;
the GNSS measurement error e and the measurement noise covariance matrix R obtained by the R-Learning task through network Learning are jointly put into a Kalman filtering system to obtain the correction quantity of the GNSS position information
Figure BDA0002801071070000042
Correcting the GNSS position information
Figure BDA0002801071070000043
The GNSS measurement noise covariance matrix R*And inputting the position data, the speed data and the attitude data into a Kalman filtering module to obtain a navigation position estimation value of the carrier.
Further, the e-Learning task comprises the steps of building a long-short term memory network (LSTM) to learn the GNSS measurement error; the R-Learning task includes an input layer, an LSTM network layer, a Dense network layer, and a connection layer.
Further, the effective characteristics comprise the number of satellites used for positioning, a precision factor, a difference state and a correction time period; the time series is of the form { χ%t|t＝t1,…,tTTherein xtFor a valid signature, t is time.
Further, the loss function is denoted as Lz，LzThe calculation is made by the following formula:
Figure BDA0002801071070000044
wherein alpha satisfies the condition
Figure BDA0002801071070000045
N denotes the size of the batch block, T denotes the length of the time step, ei,tIndicating the position error of the ith sample at time t.
Further, the method further comprises:
training a one-dimensional Convolutional Neural Network (CNN) model by taking IMU (inertial measurement unit) measurement data as input to obtain a predicted zero-velocity probability, wherein the IMU measurement data comprises acceleration and angular velocity;
calculating a first error between a zero-speed label and the zero-speed probability through an improved cross entropy loss function;
judging whether the epoch number reaches the training period or not, and if so, ending the training; otherwise, adjusting the network parameters of the convolutional neural network model by using the first error to continue training until the training is finished;
and when the zero-speed probability is greater than or equal to a preset threshold value, inputting the zero-speed information serving as measurement information into a Kalman filtering module to obtain a first navigation error, and correcting a navigation result by using the first navigation error.
Further, the improved cross-entropy loss function includes:
calculate the gradient g ═ p for all sampleszupt-lgtL, where pzuptDenotes the probability of zero velocity,/gtRepresents a zero speed tag;
dividing all gradients into k intervals;
counting the number m of samples in the ith interval and the length len of the interval i, wherein i belongs to [1, k ];
calculating gradient density GD of ith intervaliWherein, GDi＝m/len；
Calculating the gradient density coordination coefficient beta of the ith intervaliWherein beta isi＝N/GDiN is the number of samples;
the error L of the sample is calculated, wherein,
Figure BDA0002801071070000051
further, inputting the course data, the IMU measurement data and the speed data into a time domain convolution neural network (TCN) model for training to obtain a Global Navigation Satellite System (GNSS) position increment delta PpseG；
Converting the delta PpseGAdding to GNSS position PGTo obtain a GNSS pseudo-position PpseG；
GNSS pseudo-position P calculated by loss functionpseGAnd GNSS position PGA second error of (2);
judging whether the number of epochs reaches the training number, if so, finishing the training, and determining the GNSS pseudo position PpseGInputting the measurement information into a Kalman filtering module to obtain a second navigation error, and correcting a navigation result by using the second navigation error;
otherwise, adjusting the parameters of the TCN model by using the second error to continue training until the training is finished.
Based on the same inventive concept, one or more embodiments of the present specification further provide an electronic device, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and the processor implements any of the above methods when executing the program.
As can be seen from the above, the combined navigation error calibration method provided in one or more embodiments of the present disclosure avoids errors introduced by artificial setting and fixed noise, reduces positioning errors caused by the influence of the environment on sensor parameters, and effectively improves the accuracy, robustness, and practicability of combined navigation.
Drawings
In order to more clearly illustrate one or more embodiments or prior art solutions of the present specification, the drawings that are needed in the description of the embodiments or prior art will be briefly described below, and it is obvious that the drawings in the following description are only one or more embodiments of the present specification, and that other drawings may be obtained by those skilled in the art without inventive effort from these drawings.
FIG. 1 is a flow chart of a combined navigation error calibration method according to one embodiment of the present disclosure;
FIG. 2 is a flow chart of a combined navigation error calibration method according to one embodiment of the present disclosure;
FIG. 3 is a flow chart of a combined navigation error calibration method according to one embodiment of the present disclosure;
FIG. 4 is a flow chart of a combined navigation error calibration method according to one embodiment of the present disclosure;
FIG. 5 is a flow chart of a combined navigation error calibration method according to one embodiment of the present disclosure;
fig. 6 is a schematic structural diagram of an electronic device according to one or more embodiments of the present disclosure.
Detailed Description
For the purpose of promoting a better understanding of the objects, aspects and advantages of the present disclosure, reference is made to the following detailed description taken in conjunction with the accompanying drawings.
It is to be noted that unless otherwise defined, technical or scientific terms used in one or more embodiments of the present specification should have the ordinary meaning as understood by those of ordinary skill in the art to which this disclosure belongs. The use of "first," "second," and similar terms in one or more embodiments of the specification is not intended to indicate any order, quantity, or importance, but rather is used to distinguish one element from another. The word "comprising" or "comprises", and the like, means that the element or item listed before the word covers the element or item listed after the word and its equivalents, but does not exclude other elements or items.
As described in the background section, in the error correction method of the existing integrated navigation system GNSS/INS, the prediction results of the GNSS and the INS are fused by kalman filtering, so that interference and loss of GNSS signals can be prevented, and the error of the INS can be corrected; however, the convergence rate and the positioning accuracy of the integrated navigation system depend too much on the zero offset calibration result of the IMU device, the conventional integrated navigation system sets the process noise covariance matrix as a fixed value to influence the navigation result, and meanwhile, the determination of zero-speed detection and measurement noise greatly influences the integrated navigation, and finally influences the accuracy, robustness and practicability of the integrated navigation.
In view of the above, one or more embodiments of the present disclosure provide a combined navigation error calibration method, referring to fig. 1, including:
step S101: and importing accelerometer data and gyroscope data of the carrier into a mechanical layout part to obtain position data, speed data and posture data of the carrier.
Step S102: when the global satellite navigation GNSS information of the carrier is obtained through a real-time dynamic carrier phase differential technology (RTK), the position data and the speed data are imported into a GNSS Kalman filtering module to obtain a state estimation matrix and a state error matrix through calculation, and the state estimation matrix, the state error matrix and the speed data are input into the GNSS Kalman filtering module to update the speed data, the position data and the attitude data of the carrier.
Step S103: when the GNSS information of the carrier is not obtained by RTK, the velocity data is input to the kalman filtering module to update the velocity data, the position data, and the attitude data of the carrier based on an initial state estimation matrix and a state error matrix.
Step S104: storing the accelerometer data, the gyroscope data, the speed data and the attitude data in a sliding window at intervals of a preset time period, inputting the data in the sliding window into a trained residual attention network model as test data to obtain vehicle-mounted constraint (NHC) measurement noise covariance, wherein the sliding window limits the data input into the residual attention network model each time.
In this step, the predetermined time may be set to 5s, the data frequency of the accelerometer and the gyroscope is 100Hz, that is, 100 pieces of data are acquired in 1s, the method moves backward 0.1s each time in the form of a sliding window, stores 5s before the current time, that is, 500 pieces of data, and trains the data in the sliding window when prediction is needed. On one hand, the efficiency of training and prediction is guaranteed, and on the other hand, the fitting performance of the model to the current data change trend is guaranteed.
The trained residual attention network model can be trained in the following way:
and inputting the NHC measurement noise covariance into an NHC Kalman filtering module, outputting a position estimation value, feeding back the difference between the position estimation value and the true position to a mechanical arrangement part as a loss function, and finishing training the residual attention network model.
Step S105: and inputting the NHC measurement noise covariance into the NHC Kalman filtering module to obtain a position estimation value of the carrier.
In the embodiment, the residual error attention network model is used for measuring the noise covariance of the vehicle-mounted constraint in the integrated navigation algorithm, so that the adaptive noise parameter adjustment can be carried out in real time according to different road section conditions, and the errors caused by using fixed noise and artificial setting are reduced.
As an alternative embodiment, referring to fig. 2, after step S103, the method may further include:
step S201: and storing the speed data and the attitude data in a sliding window, and inputting the speed data and the attitude data serving as training data into a time convolution neural network (TCN) model to obtain a process noise covariance and an NHC (positive temperature coefficient) measurement noise covariance.
Step S202: and substituting the process noise covariance into Kalman filtering, substituting the NHC measurement noise covariance into a Kalman filtering module to obtain a position estimation value of the carrier, feeding back the difference between the position estimation value of the carrier and the true position of the carrier as a loss function to the mechanical arrangement part, finishing the training of the TCN model, wherein the state error is updated as one part of the Kalman filtering.
Step S203: and inputting the data in the sliding window into a TCN model after training as test data to obtain the process noise covariance and the NHC measurement noise covariance, inputting the process noise covariance into a Kalman filtering module, inputting the NHC measurement noise covariance into the NHC Kalman filtering module, and outputting the position estimation value of the carrier.
The method provided by the embodiment does not need a user to learn the relevant knowledge of the sensor to set the process noise parameters and the measurement noise parameters in the integrated navigation process, avoids errors possibly caused by artificial setting, and reduces the positioning errors caused by environmental problems to the sensor parameters.
As an alternative embodiment, with reference to fig. 3, the combined navigation error calibration method further includes:
step S301: and the error Learning module MT-e & R extracts effective characteristics from the NMEA-0183 protocol information and integrates the effective characteristics into characteristic data in a time series form, wherein the MT-e & R comprises an e-Learning task and an R-Learning task.
In the step, the e-Learning task comprises the steps of building a long short term memory network (LSTM) to learn the GNSS measurement error; the R-Learning task includes an input layer, an LSTM network layer, a Dense network layer, and a connection layer.
Further, the effective characteristics include the number of satellites used for positioning, a precision factor, a differential state and a correction time period; the time series is of the form { χ%t|t＝t1,…,tTTherein xtFor a valid signature, t is time.
Step S302: inputting the characteristic data into an e-learning task to pre-train network parameters so as to obtain positioning errors of the original position measurement of the GNSS in the x direction and the y direction under a geodetic coordinate system e, and recording the positioning errors as positioning errors
Figure BDA0002801071070000091
The unit is meter.
Step S303: and inputting the characteristic data into the R-Learning task, and outputting the GNSS measurement noise covariance R.
Step S304: e-Learning task utilization
Figure BDA0002801071070000092
And calculating a loss function according to the GNSS position true value y, and adjusting the network parameters through the loss function.
In this step, the loss function is denoted as Lz，LzThe calculation is made by the following formula,
Figure BDA0002801071070000093
wherein alpha satisfies the condition
Figure BDA0002801071070000094
N denotes the size of the batch block, T denotes the length of the time step, ei,tIndicating the position error of the ith sample at time t.
Step S305: the GNSS measurement error e and the measurement noise covariance matrix R obtained by the R-Learning task through network Learning are jointly put into a Kalman filtering system to obtain the correction quantity of the GNSS position information
Figure BDA0002801071070000096
Step S306: correcting the GNSS position information
Figure BDA0002801071070000097
GNSS measurement noise covariance matrix R*And inputting the position data, the speed data and the attitude data into a Kalman filtering module, and outputting a navigation position estimation value of the carrier.
In this embodiment, NMEA-0183 protocol information is introduced into a GNSS measurement error correction process, an error learning module MT-e & R is constructed, and a correlation between a position error and a position confidence is extracted from the NMEA-0183 protocol information, so that the GNSS measurement error and GNSS measurement noise are evaluated.
As an alternative embodiment, referring to fig. 4, the combined navigation error calibration method further includes:
step S401: training a one-dimensional Convolutional Neural Network (CNN) model with IMU measurement data as input to obtain a predicted zero-velocity probability, wherein the IMU measurement data comprises acceleration and angular velocity.
Step S402: and calculating a first error of the zero-speed label and the zero-speed probability through an improved cross entropy loss function.
In this step, the pseudo code of the improved cross entropy loss function is shown in table 1:
TABLE 1
Figure BDA0002801071070000101
Step S403: judging whether the number of training periods (epochs) is reached or not, and if so, finishing the training; otherwise, the network parameters of the convolutional neural network model are adjusted by the first error to continue training until the training is finished.
In this step, the training period (epoch) represents a process of training all the training samples once.
Step S404: and when the zero-speed probability is greater than or equal to a preset threshold value, inputting the zero-speed information serving as measurement information into a Kalman filtering module to obtain a first navigation error, and correcting a navigation result by using the first navigation error.
In this step, if the integrated navigation is in a vehicle-mounted environment of a complex urban road, the vehicle is stationary, which is mostly caused by traffic lights and traffic congestion, and the real speed of the vehicle should be 0 at this time, and the vehicle heading is not changed because the vibration acceleration of the engine is not 0. Therefore, the Z-axis of the true angular velocity is also 0, and the actual measurement of the gyroscope should be zero offset at rest due to the inertial error of the gyroscope.
Specifically, the speed of the mechanical programming solution when the vehicle is stationary
Figure BDA0002801071070000102
Measuring information for speed error; z-axis measurement of gyroscope
Figure BDA0002801071070000103
Zero offset epsilon from gyroscope z axisbzDifference of (2)
Figure BDA0002801071070000104
Is another kind of measurement information; wherein,
Figure BDA0002801071070000105
Figure BDA0002801071070000111
in the embodiment, the current zero-speed probability is predicted through the one-dimensional convolutional neural network, when the zero-speed probability is greater than a set value, the vehicle is judged to be in a static state, further, the zero-speed probability when the vehicle moves approaches 0, and the zero-speed probability when the vehicle is static approaches 1, so that the static and non-static motion modes can be easily separated even if a fixed threshold is used, and the dependency of the method on the threshold is reduced.
As an alternative embodiment, referring to fig. 5, the combined navigation error calibration method further includes:
step S501: associating heading data, said IMU measurement data, andinputting the speed data into a time domain convolution neural network (TCN) model for training to obtain a GNSS position increment delta PpseG。
Step S502: will be delta PpseGAdding to GNSS position PGTo obtain a GNSS pseudo-position PpseG。
Step S503: GNSS pseudo-position P calculated by loss functionpseGAnd GNSS position PGThe second error of (2).
Step S504: judging whether the number of epochs reaches the training number, if so, finishing the training, and determining the GNSS pseudo position PpseGAnd inputting the measurement information into a Kalman filtering module to obtain a second navigation error, and correcting the navigation result by using the second navigation error.
Step S505: otherwise, adjusting the parameters of the time domain convolution neural network model by using the second error to continue training until the training is finished.
In the embodiment, the GNSS pseudo position is predicted by the TCN model, the integrated navigation system during the GNSS signal interruption is assisted, the mapping between the INS data and the GNSS position increment is directly established, the output of the TCN model only relates to the GNSS position increment information, and the error caused by the two parts is avoided.
Therefore, the combined navigation error calibration method provided by one or more embodiments of the specification can effectively improve the accuracy, robustness and practicability of combined navigation.
It should be noted that the method of one or more embodiments of the present disclosure may be performed by a single device, such as a computer or server. The method of the embodiment can also be applied to a distributed scene and completed by the mutual cooperation of a plurality of devices. In such a distributed scenario, one of the devices may perform only one or more steps of the method of one or more embodiments of the present disclosure, and the devices may interact with each other to complete the method.
It should be noted that the above description describes certain embodiments of the present disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
Based on the same inventive concept, corresponding to any of the above-mentioned embodiments, one or more embodiments of the present specification further provide an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the computer program, the combined navigation error calibration method according to any of the above-mentioned embodiments is implemented.
Fig. 6 is a schematic diagram illustrating a more specific hardware structure of an electronic device according to this embodiment, where the electronic device may include: a processor 1010, a memory 1020, an input/output interface 1030, a communication interface 1040, and a bus 1050. Wherein the processor 1010, memory 1020, input/output interface 1030, and communication interface 1040 are communicatively coupled to each other within the device via bus 1050.
The processor 1010 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an Application Specific Integrated Circuit (ASIC), or one or more Integrated circuits, and is configured to execute related programs to implement the technical solutions provided in the embodiments of the present disclosure.
The Memory 1020 may be implemented in the form of a ROM (Read Only Memory), a RAM (Random Access Memory), a static storage device, a dynamic storage device, or the like. The memory 1020 may store an operating system and other application programs, and when the technical solution provided by the embodiments of the present specification is implemented by software or firmware, the relevant program codes are stored in the memory 1020 and called to be executed by the processor 1010.
The input/output interface 1030 is used for connecting an input/output module to input and output information. The i/o module may be configured as a component in a device (not shown) or may be external to the device to provide a corresponding function. The input devices may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output devices may include a display, a speaker, a vibrator, an indicator light, etc.
The communication interface 1040 is used for connecting a communication module (not shown in the drawings) to implement communication interaction between the present apparatus and other apparatuses. The communication module can realize communication in a wired mode (such as USB, network cable and the like) and also can realize communication in a wireless mode (such as mobile network, WIFI, Bluetooth and the like).
Bus 1050 includes a path that transfers information between various components of the device, such as processor 1010, memory 1020, input/output interface 1030, and communication interface 1040.
It should be noted that although the above-mentioned device only shows the processor 1010, the memory 1020, the input/output interface 1030, the communication interface 1040 and the bus 1050, in a specific implementation, the device may also include other components necessary for normal operation. In addition, those skilled in the art will appreciate that the above-described apparatus may also include only those components necessary to implement the embodiments of the present description, and not necessarily all of the components shown in the figures.
The electronic device of the above embodiment is used to implement the corresponding combined navigation error calibration method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiment, which are not described herein again.
Those of ordinary skill in the art will understand that: the discussion of any embodiment above is meant to be exemplary only, and is not intended to intimate that the scope of the disclosure, including the claims, is limited to these examples; within the spirit of the present disclosure, features from the above embodiments or from different embodiments may also be combined, steps may be implemented in any order, and there are many other variations of different aspects of one or more embodiments of the present description as described above, which are not provided in detail for the sake of brevity.
In addition, well-known power/ground connections to Integrated Circuit (IC) chips and other components may or may not be shown in the provided figures, for simplicity of illustration and discussion, and so as not to obscure one or more embodiments of the disclosure. Furthermore, devices may be shown in block diagram form in order to avoid obscuring the understanding of one or more embodiments of the present description, and this also takes into account the fact that specifics with respect to implementation of such block diagram devices are highly dependent upon the platform within which the one or more embodiments of the present description are to be implemented (i.e., specifics should be well within purview of one skilled in the art). Where specific details (e.g., circuits) are set forth in order to describe example embodiments of the disclosure, it should be apparent to one skilled in the art that one or more embodiments of the disclosure can be practiced without, or with variation of, these specific details. Accordingly, the description is to be regarded as illustrative instead of restrictive.
While the present disclosure has been described in conjunction with specific embodiments thereof, many alternatives, modifications, and variations of these embodiments will be apparent to those of ordinary skill in the art in light of the foregoing description. For example, other memory architectures (e.g., dynamic ram (dram)) may use the discussed embodiments.
It is intended that the one or more embodiments of the present specification embrace all such alternatives, modifications and variations as fall within the broad scope of the appended claims. Therefore, any omissions, modifications, substitutions, improvements, and the like that may be made without departing from the spirit and principles of one or more embodiments of the present disclosure are intended to be included within the scope of the present disclosure.

Claims (10)

1. A combined navigation error calibration method, comprising:
importing accelerometer data and gyroscope data of a carrier into a mechanical layout part to obtain position data, speed data and posture data of the carrier;
when the global satellite navigation GNSS information of the carrier is obtained through a real-time dynamic carrier phase differential technology RTK, the position data and the speed data are imported into a GNSS Kalman filtering module to obtain a state estimation matrix and a state error matrix through calculation, and the state estimation matrix, the state error matrix and the speed data are input into the GNSS Kalman filtering module to update the speed data, the position data and the attitude data of the carrier;
when the GNSS information of the carrier is not obtained through RTK, the speed data is input into the Kalman filtering module so as to update the speed data, the position data and the attitude data of the carrier based on an initial state estimation matrix and a state error matrix;
storing the accelerometer data, the gyroscope data, the speed data and the attitude data in a sliding window at intervals of a preset time period, inputting the data in the sliding window into a trained residual attention network model as test data to obtain vehicle-mounted constraint NHC measurement noise covariance, wherein the sliding window limits the data input into the residual attention network model each time;
and inputting the NHC measurement noise covariance into the NHC Kalman filtering module to obtain a position estimation value of the carrier.
2. The method of claim 1, wherein the velocity data is input to a kalman filter module to update the velocity data, the position data, and the attitude data of the vehicle based on an initial state estimation matrix and a state error matrix when GNSS information of the vehicle is not obtained by RTK, and thereafter further comprising:
storing the speed data and the attitude data in a sliding window, and inputting the speed data and the attitude data serving as training data into a time convolution neural network TCN model to obtain a process noise covariance and an NHC (positive temperature coefficient) measurement noise covariance;
inputting the process noise covariance into a Kalman filtering module, substituting the NHC measurement noise covariance into the NHC Kalman filtering module to obtain a position estimation value of the carrier, feeding back the difference between the position estimation value of the carrier and the true position of the carrier as a loss function to the mechanical arrangement part, and finishing the training of the TCN model;
and inputting the data in the sliding window into a TCN model after training as test data to obtain the process noise covariance and the NHC measurement noise covariance, inputting the process noise covariance into a Kalman filtering module, inputting the NHC measurement noise covariance into the NHC Kalman filtering module, and outputting the position estimation value of the carrier.
3. The method of claim 1, further comprising:
an error Learning module MT-e & R extracts effective characteristics from NMEA-0183 protocol information and integrates the effective characteristics into characteristic data in a time series form, wherein the MT-e & R comprises an e-Learning task and an R-Learning task;
inputting the characteristic data into an e-learning task to pre-train network parameters so as to obtain positioning errors of the original position measurement of the GNSS in the x direction and the y direction under a geodetic coordinate system e, and recording the positioning errors as positioning errors
Figure FDA0002801071060000021
Inputting the characteristic data into the R-Learning task to obtain a GNSS measurement noise covariance R;
the e-Learning task utilizes
Figure FDA0002801071060000022
Calculating a loss function according to the GNSS position true value y, and adjusting network parameters through the loss function;
the GNSS measurement error e and the measurement noise covariance matrix R obtained by the R-Learning task through network Learning are jointly put into a Kalman filtering system to obtain the correction quantity of the GNSS position information
Figure FDA0002801071060000024
Will be described inCorrection of GNSS position information
Figure FDA0002801071060000023
The GNSS measurement noise covariance matrix R*And inputting the position data, the speed data and the attitude data into a Kalman filtering module to obtain a navigation position estimation value of the carrier.
4. The method of claim 3, wherein the e-Learning task comprises Learning GNSS measurement errors by building a long-short term memory network (LSTM); the R-Learning task includes an input layer, an LSTM network layer, a Dense network layer, and a connection layer.
5. The method of claim 3, wherein the valid features include the number of satellites used for positioning, a precision factor, a difference state, a correction period; the time series is of the form { χ%t|t＝t1,…,tTTherein xtFor a valid signature, t is time.
6. The method of claim 3, wherein said loss function is denoted Lz，LzThe calculation is made by the following formula:
Figure FDA0002801071060000031
wherein alpha satisfies the condition
Figure FDA0002801071060000032
N denotes the size of the batch block, T denotes the length of the time step, ei,tIndicating the position error of the ith sample at time t.
7. The method of claim 1, further comprising:
training a one-dimensional Convolutional Neural Network (CNN) model by taking IMU (inertial measurement unit) measurement data as input to obtain a predicted zero-velocity probability, wherein the IMU measurement data comprises acceleration and angular velocity;
calculating a first error between a zero-speed label and the zero-speed probability through an improved cross entropy loss function;
judging whether the epoch number reaches the training period or not, and if so, ending the training; otherwise, adjusting the network parameters of the convolutional neural network model by using the first error to continue training until the training is finished;
and when the zero-speed probability is greater than or equal to a preset threshold value, inputting the zero-speed information serving as measurement information into a Kalman filtering module to obtain a first navigation error, and correcting a navigation result by using the first navigation error.
8. The method of claim 5, wherein the improved cross-entropy loss function comprises:
calculate the gradient g ═ p for all sampleszupt-lgtL, where pzuptDenotes the probability of zero velocity,/gtRepresents a zero speed tag;
dividing all gradients into k intervals;
counting the number m of samples in the ith interval and the length len of the interval i, wherein i belongs to [1, k ];
calculating gradient density GD of ith intervaliWherein, GDi＝m/len；
Calculating the gradient density coordination coefficient beta of the ith intervaliWherein beta isi＝N/GDiN is the number of samples;
the error L of the sample is calculated, wherein,
Figure FDA0002801071060000033
9. the method of claim 1, further comprising,
inputting the course data, the IMU measurement data and the speed data into a time domain convolution neural network (TCN) model for training to obtain a Global Navigation Satellite System (GNSS) position increment delta PpseG；
Converting the delta PpseGAdding to GNSS position PGTo obtain a GNSS pseudo-position PpseG；
GNSS pseudo-position P calculated by loss functionpseGAnd GNSS position PGA second error of (2);
judging whether the number of epochs reaches the training number, if so, finishing the training, and determining the GNSS pseudo position PpseGInputting the measurement information into a Kalman filtering module to obtain a second navigation error, and correcting a navigation result by using the second navigation error;
otherwise, adjusting the parameters of the TCN model by using the second error to continue training until the training is finished.
10. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the method according to any of claims 1 to 9 when executing the program.
CN202011350339.6A 2020-11-26 2020-11-26 Combined navigation error calibration method and electronic equipment Active CN112577521B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011350339.6A CN112577521B (en) 2020-11-26 2020-11-26 Combined navigation error calibration method and electronic equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011350339.6A CN112577521B (en) 2020-11-26 2020-11-26 Combined navigation error calibration method and electronic equipment

Publications (2)

Publication Number Publication Date
CN112577521A true CN112577521A (en) 2021-03-30
CN112577521B CN112577521B (en) 2023-11-17

Family

ID=75123729

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011350339.6A Active CN112577521B (en) 2020-11-26 2020-11-26 Combined navigation error calibration method and electronic equipment

Country Status (1)

Country Link
CN (1) CN112577521B (en)

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112985462A (en) * 2021-04-21 2021-06-18 智道网联科技(北京)有限公司 Method and device for correcting inertial measurement data based on convolutional neural network model
CN113124901A (en) * 2021-04-01 2021-07-16 中铁第四勘察设计院集团有限公司 Position correction method and device, electronic device and storage medium
CN113780520A (en) * 2021-08-19 2021-12-10 桂林理工大学 Landslide MEMS acceleration sensor error compensation method
CN113917512A (en) * 2021-12-13 2022-01-11 智道网联科技(北京)有限公司 Positioning method and device for automatic driving vehicle, electronic equipment and storage medium
CN114147717A (en) * 2021-12-09 2022-03-08 乐聚(深圳)机器人技术有限公司 Robot motion trajectory estimation method and device, controller and storage medium
CN114488227A (en) * 2022-01-26 2022-05-13 西南交通大学 Multipath error correction method based on spatial correlation
CN114637036A (en) * 2022-05-18 2022-06-17 深圳华大北斗科技股份有限公司 Non-integrity-constrained adaptive noise measurement method
CN114966762A (en) * 2022-05-18 2022-08-30 中国人民解放军空军工程大学 Fault detection method of GNSS/INS tightly-combined navigation system
CN115932913A (en) * 2022-11-24 2023-04-07 国网思极位置服务有限公司 Satellite positioning pseudo-range correction method and device
CN117706583A (en) * 2023-12-29 2024-03-15 无锡物联网创新中心有限公司 High-precision positioning method and system

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103339472A (en) * 2010-12-03 2013-10-02 高通股份有限公司 Inertial sensor aided heading and positioning for GNSS vehicle navigation
US20170227363A1 (en) * 2015-01-08 2017-08-10 Uti Limited Partnership Method and apparatus for enhanced pedestrian navigation based on wlan and mems sensors
US20180188384A1 (en) * 2017-01-04 2018-07-05 Qualcomm Incorporated Systems and methods for using a sliding window of global positioning epochs in visual-inertial odometry
CN109059904A (en) * 2018-06-01 2018-12-21 浙江亚特电器有限公司 Combinated navigation method for mobile carrier
CN111272174A (en) * 2020-02-27 2020-06-12 中国科学院计算技术研究所 Combined navigation method and system
US20200208982A1 (en) * 2018-12-28 2020-07-02 Uti Limited Partnership Sensor placement and methods for aiding navigation during gnss signal outage

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103339472A (en) * 2010-12-03 2013-10-02 高通股份有限公司 Inertial sensor aided heading and positioning for GNSS vehicle navigation
US20170227363A1 (en) * 2015-01-08 2017-08-10 Uti Limited Partnership Method and apparatus for enhanced pedestrian navigation based on wlan and mems sensors
US20180188384A1 (en) * 2017-01-04 2018-07-05 Qualcomm Incorporated Systems and methods for using a sliding window of global positioning epochs in visual-inertial odometry
CN109059904A (en) * 2018-06-01 2018-12-21 浙江亚特电器有限公司 Combinated navigation method for mobile carrier
WO2019228437A1 (en) * 2018-06-01 2019-12-05 浙江亚特电器有限公司 Integrated navigation method for mobile vehicle
US20200208982A1 (en) * 2018-12-28 2020-07-02 Uti Limited Partnership Sensor placement and methods for aiding navigation during gnss signal outage
CN111272174A (en) * 2020-02-27 2020-06-12 中国科学院计算技术研究所 Combined navigation method and system

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
GAO XILE ET AL.: "RL-AKF: An Adaptive Kalman Filter Navigation Algorithm Based on Reinforcement Learning for Ground Vehicles", 《REMOTE SENSING》, vol. 12, no. 11, pages 1704 *
冯木榉等: "基于简化车辆非完整约束辅助的低成本车载INS/GNSS 组合导航***研究", 《微电子学与计算机》, vol. 37, no. 8, pages 78 - 82 *
李彦杰;杨元喜;何海波;: "附加约束条件对GNSS/INS组合导航结果的影响分析", 武汉大学学报(信息科学版), no. 09, pages 68 - 74 *
郭俊鸽等: "非完整约束下滤波算法在组合导航中的应用", 《计算机工程与应用》, vol. 51, no. 24, pages 215 - 220 *

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113124901A (en) * 2021-04-01 2021-07-16 中铁第四勘察设计院集团有限公司 Position correction method and device, electronic device and storage medium
CN113124901B (en) * 2021-04-01 2022-03-11 中铁第四勘察设计院集团有限公司 Position correction method and device, electronic device and storage medium
CN112985462A (en) * 2021-04-21 2021-06-18 智道网联科技(北京)有限公司 Method and device for correcting inertial measurement data based on convolutional neural network model
CN112985462B (en) * 2021-04-21 2021-07-23 智道网联科技(北京)有限公司 Method and device for correcting inertial measurement data based on convolutional neural network model
CN113780520A (en) * 2021-08-19 2021-12-10 桂林理工大学 Landslide MEMS acceleration sensor error compensation method
CN114147717A (en) * 2021-12-09 2022-03-08 乐聚(深圳)机器人技术有限公司 Robot motion trajectory estimation method and device, controller and storage medium
CN114147717B (en) * 2021-12-09 2024-05-24 乐聚(深圳)机器人技术有限公司 Robot motion track estimation method, device, controller and storage medium
CN113917512A (en) * 2021-12-13 2022-01-11 智道网联科技(北京)有限公司 Positioning method and device for automatic driving vehicle, electronic equipment and storage medium
CN114488227B (en) * 2022-01-26 2023-10-20 西南交通大学 Multipath error correction method based on spatial correlation
CN114488227A (en) * 2022-01-26 2022-05-13 西南交通大学 Multipath error correction method based on spatial correlation
CN114637036B (en) * 2022-05-18 2022-08-19 深圳华大北斗科技股份有限公司 Non-integrity-constrained adaptive noise measurement method
CN114966762A (en) * 2022-05-18 2022-08-30 中国人民解放军空军工程大学 Fault detection method of GNSS/INS tightly-combined navigation system
CN114637036A (en) * 2022-05-18 2022-06-17 深圳华大北斗科技股份有限公司 Non-integrity-constrained adaptive noise measurement method
CN114966762B (en) * 2022-05-18 2024-05-28 中国人民解放军空军工程大学 Fault detection method of GNSS/INS (Global navigation satellite System/inertial navigation System) tightly integrated navigation system
CN115932913A (en) * 2022-11-24 2023-04-07 国网思极位置服务有限公司 Satellite positioning pseudo-range correction method and device
CN115932913B (en) * 2022-11-24 2024-03-08 国网思极位置服务有限公司 Satellite positioning pseudo-range correction method and device
CN117706583A (en) * 2023-12-29 2024-03-15 无锡物联网创新中心有限公司 High-precision positioning method and system

Also Published As

Publication number Publication date
CN112577521B (en) 2023-11-17

Similar Documents

Publication Publication Date Title
CN112577521B (en) Combined navigation error calibration method and electronic equipment
CN111121768B (en) Robot pose estimation method and device, readable storage medium and robot
CN106500693B (en) A kind of AHRS algorithm based on adaptive extended kalman filtering
WO2017064790A1 (en) Positioning device and positioning method
CN111707260B (en) Positioning method based on frequency domain analysis and convolutional neural network
CN113465628B (en) Inertial measurement unit data compensation method and system
CN112082547B (en) Method and device for optimizing integrated navigation system, electronic equipment and storage medium
CN113933818A (en) Method, device, storage medium and program product for calibrating laser radar external parameter
CN108534744A (en) A kind of attitude angle acquisition methods, device and handle
CN109507706B (en) GPS signal loss prediction positioning method
CN111026081B (en) Error calculation method, device, equipment and storage medium
KR101576424B1 (en) Automatic calibration method of magnetometer for indoor positioning
CN112946681B (en) Laser radar positioning method fusing combined navigation information
CN116007620A (en) Combined navigation filtering method, system, electronic equipment and storage medium
CN114383605B (en) Indoor positioning and optimizing method based on MEMS sensor and sparse landmark point
CN111197994B (en) Position data correction method, position data correction device, computer device, and storage medium
CN116224407B (en) GNSS and INS integrated navigation positioning method and system
CN116242373A (en) High-precision navigation positioning method and system for fusing multi-source data
CN113959433B (en) Combined navigation method and device
CN114019954A (en) Course installation angle calibration method and device, computer equipment and storage medium
CN112985385A (en) Positioning and orientation system and positioning and orientation method applying high-precision map
CN113137975B (en) Inertial correction method and device for astronomical inertial integrated navigation and electronic equipment
CN111002991B (en) Method and device for processing vehicle-mounted navigation information and computer storage medium
CN113566849B (en) Method and device for calibrating installation angle of inertial measurement unit and computer equipment
CN110579212B (en) Indoor positioning method and device

Legal Events

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