WO2022007437A1 - 传感器安装偏差角的标定方法、组合定位系统和车辆 - Google Patents

传感器安装偏差角的标定方法、组合定位系统和车辆 Download PDF

Info

Publication number
WO2022007437A1
WO2022007437A1 PCT/CN2021/083032 CN2021083032W WO2022007437A1 WO 2022007437 A1 WO2022007437 A1 WO 2022007437A1 CN 2021083032 W CN2021083032 W CN 2021083032W WO 2022007437 A1 WO2022007437 A1 WO 2022007437A1
Authority
WO
WIPO (PCT)
Prior art keywords
imu
deviation angle
installation deviation
lidar
observation
Prior art date
Application number
PCT/CN2021/083032
Other languages
English (en)
French (fr)
Inventor
王磊杰
温丰
刘镇波
姜锐
Original Assignee
华为技术有限公司
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 华为技术有限公司 filed Critical 华为技术有限公司
Priority to EP21838295.0A priority Critical patent/EP4170282A4/en
Publication of WO2022007437A1 publication Critical patent/WO2022007437A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C25/00Manufacturing, calibrating, cleaning, or repairing instruments or devices referred to in the other groups of this subclass
    • G01C25/005Manufacturing, calibrating, cleaning, or repairing instruments or devices referred to in the other groups of this subclass initial alignment, calibration or starting-up of inertial devices
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01SRADIO 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
    • G01S7/00Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/48Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S17/00
    • G01S7/497Means for monitoring or calibrating
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/10Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration
    • G01C21/12Navigation; 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
    • G01C21/16Navigation; 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
    • G01C21/165Navigation; 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
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/10Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration
    • G01C21/12Navigation; 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
    • G01C21/16Navigation; 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
    • G01C21/165Navigation; 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
    • G01C21/1652Navigation; 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 with ranging devices, e.g. LIDAR or RADAR
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01SRADIO 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
    • G01S17/00Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
    • G01S17/86Combinations of lidar systems with systems other than lidar, radar or sonar, e.g. with direction finders
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01SRADIO 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
    • G01S7/00Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/48Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S17/00
    • G01S7/497Means for monitoring or calibrating
    • G01S7/4972Alignment of sensor

Definitions

  • the present application relates to the technical field of automatic driving, and in particular, to a method for calibrating a sensor installation deviation angle, a combined positioning system, and a vehicle.
  • the positioning system of the autonomous vehicle is used to determine the position of the vehicle and provide the most basic information for the path planning and navigation of the vehicle.
  • the performance of the positioning system will directly determine the stability and reliability of the autonomous vehicle.
  • the positioning system is generally a combined positioning system including a variety of sensors, such as inertial measurement unit (IMU) used by inertial navigation system (INS), global navigation satellite system (GNSS) ) receiver, LiDAR (light detection and ranging, LiDAR), odometer, vision sensor, etc.
  • IMU inertial measurement unit
  • INS inertial navigation system
  • GNSS global navigation satellite system
  • LiDAR light detection and ranging
  • LiDAR odometer
  • vision sensor etc.
  • Each sensor in the combined positioning system is usually installed in different positions of the vehicle, so it is necessary to calibrate the installation position relationship between each sensor before use, such as calibrating the installation deviation angle between each sensor, so as to accurately measure the position of each sensor.
  • the solution result or the positioning result is filtered and calculated to improve the positioning accuracy.
  • the calibration of the installation deviation angle between each sensor mainly adopts the offline calibration method, which relies on special calibration tools and uses some optical means or indirect calibration methods.
  • the operation is cumbersome and the calibration time is long.
  • the vehicle's installation position relationship between sensors changes due to long-term vibration, bumps, etc., it needs to be re-calibrated, otherwise the positioning accuracy of the combined positioning system will be affected.
  • the embodiments of the present application provide a method for calibrating the installation deviation angle of sensors, a combined positioning system and a vehicle, which can realize online calibration of the installation deviation angle of sensors.
  • the calibration result can be updated online. , so as to always ensure the positioning accuracy of the combined positioning system.
  • an embodiment of the present application provides a method for calibrating a sensor installation deviation angle.
  • the method is applied to a combined positioning system.
  • the combined positioning system includes a plurality of sensors, and the installation positions of the plurality of sensors are different; the method includes: according to each The attitude information output by the sensors is used to construct the error model of the installation deviation angle between each sensor; the system state equation is constructed, and the state variables of the system state equation include the installation deviation angle; the system observation equation is constructed, and the observation amount of the system observation equation includes the error model Determine the observation matrix corresponding to the installation deviation angle; filter and solve the system state equation and the system observation equation, and use the estimated result of the installation deviation angle obtained when the filtering solution results converge as the calibration result.
  • the method provided by the embodiment of the present application integrates the installation deviation angle between the various sensors into the state variable of the combined positioning system, so that the installation deviation angle between the various sensors can be calibrated during the driving process of the vehicle (that is, online calibration). ), the calibration process does not need to consider the physical structure of the sensor, and the calibration accuracy is high.
  • the method of the embodiment of the present application does not require the use of a calibration tool, and only requires the vehicle to generate a maneuver to complete the calibration, and the operation is simple and fast.
  • the method of the embodiment of the present application can update the calibration result online, so as to always ensure the positioning accuracy of the combined positioning system.
  • the plurality of sensors include an inertial measurement unit IMU of an inertial navigation system INS and a global navigation satellite system GNSS receiver, the GNSS receiver includes a GNSS antenna, and the installation positions of the IMU and the GNSS antenna are different;
  • the attitude information output by the sensors is used to construct the error model of the installation deviation angle between each sensor, including: constructing the observation amount of the error model, the observation amount is the attitude deviation between the IMU and the GNSS antenna; constructing the attitude matrix corresponding to the INS to the GNSS The transfer relationship equation of the corresponding attitude matrix; the observation amount and the transfer relationship equation of the simultaneous error model construct an error model, and the error model is the error model of the installation deviation angle between the IMU and the GNSS antenna; wherein, the attitude information includes One or more of roll angle, pitch angle, and heading angle, and the transfer relationship equation includes the attitude error of the INS and the installation deviation angle between the IMU and the GNSS antenna.
  • the observations of the error model include:
  • ⁇ m , ⁇ m are the roll angle, pitch angle and heading angle output by the INS, respectively;
  • ⁇ g , ⁇ g are the roll angle, pitch angle and heading angle output by GNSS, respectively.
  • the transition relationship equation includes:
  • the error model includes:
  • Z 1 , Z 2 , and Z 3 are the observed quantities of the error model; are the E-axis attitude error N-axis attitude error and U-axis attitude error of INS based on the east-north-sky coordinate system; are the X-axis installation deviation angle, the Y-axis installation deviation angle, and the Z-axis installation deviation angle between the IMU and the GNSS antenna based on the upper right coordinate system; other parameters are the attitude matrix corresponding to GNSS parameters in .
  • system state equation includes:
  • X 1 and X 2 are state variables, X 1 is the inertial navigation error, and X 2 is the installation deviation angle between the IMU and the GNSS antenna; is the derivative of X 1 ; F 1 is the system matrix of the inertial navigation error model of the combined positioning system, and G 1 and W 1 are the inertial navigation error driving noise; is the derivative of X 2 ; F 2 is the system matrix of the error model of the installation deviation angle between the IMU and the GNSS antenna, and G 2 and W 2 are the error driving noise of the installation deviation angle between the IMU and the GNSS antenna.
  • system state equation includes:
  • Z is the system observation quantity
  • H is the system observation matrix
  • X is the system state equation
  • V is the system observation noise matrix
  • H 1 is the observation matrix corresponding to the attitude error of the INS
  • H 2 is the inertial navigation error model except the attitude error.
  • the observation matrix corresponding to other state variables other than H 3 is the observation matrix corresponding to the installation deviation angle between the IMU and the GNSS antenna.
  • the system state equation and the system observation equation are filtered and solved, and the optimal estimation result of the installation deviation angle obtained when the filter solution result converges is used as the calibration result, including: the system state equation Perform Kalman filtering calculation with the system observation equation to obtain the optimal estimation result of the installation deviation angle between the IMU and the GNSS antenna; according to the optimal estimation result, the GNSS positioning result is compensated, and the inertial navigation solution result for the INS is obtained. Correction is performed; when the filtering solution results converge, the optimal estimation result of the installation deviation angle between the IMU and the GNSS antenna is saved as the calibration result in the configuration file.
  • the multiple sensors include IMU and LiDAR, and the installation positions of the IMU and LiDAR are different; according to the attitude information output by each sensor, an error model of the installation deviation angle between each sensor is constructed, including : Construct the observation amount of the error model, the observation amount is the attitude deviation between the IMU and LiDAR; construct the transfer relationship equation from the attitude matrix corresponding to the INS to the attitude matrix corresponding to the LiDAR; build the observation amount and the transfer relationship equation of the simultaneous error model Error model, the error model is the error model of the installation deviation angle between the IMU and the LiDAR; the attitude information includes one or more of the roll angle, pitch angle and heading angle, and the transfer relationship equation includes the attitude error of the INS and the IMU and the Mounting deviation angle between LiDARs.
  • the observations of the error model include:
  • ⁇ m , ⁇ m are the roll angle, pitch angle and heading angle output by the INS, respectively;
  • ⁇ l , ⁇ l are the roll, pitch, and heading angles output by LiDAR, respectively.
  • the transition relationship equation includes:
  • ⁇ 1 is the installation deviation angle between the IMU and the LiDAR antenna
  • ⁇ 1 ⁇ represents the cross product of ⁇ 1.
  • the error model includes:
  • Z 1 , Z 2 , and Z 3 are the observed quantities of the error model; are the E-axis attitude error N-axis attitude error and U-axis attitude error of INS based on the east-north-sky coordinate system; are the X-axis installation deviation angle, the Y-axis installation deviation angle, and the Z-axis installation deviation angle based on the upper right coordinate system between the IMU and the LiDAR antenna; other parameters are the attitude matrix corresponding to the LiDAR parameters in .
  • system state equation includes:
  • X 1 and X 3 are state variables, X 1 is the inertial navigation error, and X 3 is the installation deviation angle between the IMU and the LiDAR; is the derivative of X 1 ; F 1 is the system matrix of the inertial navigation error model of the combined positioning system, and G 1 and W 1 are the inertial navigation error driving noise; is the derivative of X 3 ; F 3 is the system matrix of the error model of the installation deviation angle between the IMU and the LiDAR, and G 3 and W 3 are the error driving noise of the installation deviation angle between the IMU and the LiDAR.
  • system state equation includes:
  • Z is the system observation quantity
  • H is the system observation matrix
  • X is the system state equation
  • V is the system observation noise matrix
  • H 1 is the observation matrix corresponding to the attitude error of the INS
  • H 2 is the inertial navigation error model except the attitude error.
  • the observation matrix corresponding to other state variables other than H 3 is the observation matrix corresponding to the installation deviation angle between the IMU and the LiDAR.
  • the system state equation and the system observation equation are filtered and solved, and the optimal estimation result of the installation deviation angle obtained when the filter solution result converges is used as the calibration result, including: the system state equation Perform Kalman filter calculation with the system observation equation to obtain the optimal estimation result of the installation deviation angle between the IMU and LiDAR; according to the optimal estimation result, the positioning result of the LiDAR is compensated, and the inertial navigation solution result of the INS is calculated. Correction: when the filter solution results converge, the optimal estimation result of the installation deviation angle between the IMU and the LiDAR is saved in the configuration file as the calibration result.
  • an embodiment of the present application provides a combined positioning system
  • the system includes a processor, a memory and a plurality of sensors, and the installation positions of the plurality of sensors are different;
  • the memory includes program instructions, and when the program instructions are executed by the processor,
  • the combined positioning system is used to perform the following steps: constructing an error model of the installation deviation angle between each sensor according to the attitude information output by each sensor; constructing a system state equation, and the state variables of the system state equation include the installation deviation angle; constructing a system observation Equation, the observation amount of the system observation equation includes the observation matrix corresponding to the installation deviation angle determined according to the error model; the system state equation and the system observation equation are filtered and solved, and the estimated result of the installation deviation angle obtained when the filtering solution results converge as the calibration result.
  • the combined positioning system provided by the embodiment of the present application integrates the installation deviation angle between the various sensors into the state variables of the combined positioning system, so that the installation deviation angle between the various sensors can be calibrated during the driving process of the vehicle (ie Online calibration), the calibration process does not need to consider the physical structure of the sensor, and the calibration accuracy is high.
  • the combined positioning system of the embodiment of the present application does not require the use of a calibration tool, and only requires the vehicle to maneuver to complete the calibration, and the operation is simple and fast.
  • the method of the embodiment of the present application can update the calibration result online, so as to always ensure the positioning accuracy of the combined positioning system.
  • the plurality of sensors include an inertial measurement unit IMU of an inertial navigation system INS and a global navigation satellite system GNSS receiver, the GNSS receiver includes a GNSS antenna, and the installation positions of the IMU and the GNSS antenna are different; program instructions When run by the processor, the combined positioning system is used to perform the following steps, so as to construct the error model of the installation deviation angle between each sensor according to the attitude information output by each sensor: construct the observation quantity of the error model, and the observation quantity is the IMU
  • the attitude deviation between the GNSS antenna and the INS; the transfer relationship equation from the attitude matrix corresponding to INS to the attitude matrix corresponding to GNSS is constructed; the observation amount and transfer relationship equation of the simultaneous error model are used to construct an error model, and the error model is the difference between the IMU and the GNSS antenna.
  • the combined positioning system is used to perform the following steps, so as to realize the filtering and solving of the system state equation and the system observation equation, and when the filtering and solving results converge
  • the obtained optimal estimation result of the installation deviation angle is used as the calibration result: Kalman filtering is performed on the system state equation and the system observation equation to obtain the optimal estimation result of the installation deviation angle between the IMU and the GNSS antenna; according to the optimal estimation The result compensates the positioning result of GNSS and corrects the inertial navigation solution result of INS; when the filter solution result converges, the optimal estimation result of the installation deviation angle between the IMU and the GNSS antenna is saved as the calibration result to in the configuration file.
  • the plurality of sensors include an IMU and a lidar LiDAR, and the installation positions of the IMU and the LiDAR are different; when the program instructions are executed by the processor, the combined positioning system is used to perform the following steps to realize the The attitude information output by the sensor, construct the error model of the installation deviation angle between each sensor: construct the observation amount of the error model, the observation amount is the attitude deviation between the IMU and LiDAR; construct the attitude matrix corresponding to the INS to the attitude corresponding to the LiDAR
  • the transfer relationship equation of the matrix; the observation amount and transfer relationship equation of the simultaneous error model construct an error model, and the error model is the error model of the installation deviation angle between the IMU and the LiDAR; among them, the attitude information includes roll angle, pitch angle and heading angle
  • One or more of the transfer relation equations include the attitude error of the INS and the mounting deviation angle between the IMU and the LiDAR.
  • the combined positioning system is used to perform the following steps, so as to realize the filtering and solving of the system state equation and the system observation equation, and when the filtering and solving results converge
  • the obtained optimal estimation result of the installation deviation angle is used as the calibration result: Kalman filtering is performed on the system state equation and system observation equation to obtain the optimal estimation result of the installation deviation angle between the IMU and LiDAR; according to the optimal estimation result Compensate the positioning results of the LiDAR and correct the inertial navigation solution results of the INS; when the filter solution results converge, save the optimal estimation result of the installation deviation angle between the IMU and the LiDAR as the calibration result to the configuration file middle.
  • an embodiment of the present application provides a vehicle, the vehicle includes a combined positioning system, the combined positioning system includes a processor, a memory and a plurality of sensors, different sensors are installed in different positions of the vehicle; the memory includes program instructions, and the program The instructions, when executed by the processor, cause the combined positioning system to perform the method of the aforementioned first aspect and any implementations thereof.
  • the embodiments of the present application further provide a computer-readable storage medium, where instructions are stored in the computer-readable storage medium, when the computer-readable storage medium runs on a computer, the computer executes the methods of the above aspects.
  • embodiments of the present application further provide a computer program product containing instructions, which, when executed on a computer, cause the computer to execute the methods of the above aspects.
  • an embodiment of the present application further provides a chip system, where the chip system includes a processor for supporting the above-mentioned apparatus or system to implement the functions involved in the above-mentioned aspects, for example, generating or processing the functions involved in the above-mentioned method. information.
  • Fig. 1 is the data logic block diagram of the combined positioning system adopting the combined positioning algorithm
  • Fig. 2 is the position schematic diagram of each sensor in the combined positioning system on the vehicle
  • FIG. 3 is a schematic diagram of an installation deviation angle shown in an embodiment of the present application.
  • FIG. 4 is a hardware structure diagram of a combined positioning system provided by an embodiment of the present application.
  • FIG. 5 is a data logic block diagram of an expanded combined positioning system provided by an embodiment of the present application.
  • FIG. 6 is a flowchart of the method for calibrating the sensor installation deviation angle provided by the embodiment of the present application applied to the calibration of the installation deviation angle between the IMU and the GNSS antenna;
  • Fig. 7 is the flow chart that the calibration method of the sensor installation deviation angle that the embodiment of the present application provides is applied to the calibration of IMU and LiDAR installation deviation angle;
  • FIG. 8 is a schematic diagram of a software module of a combined positioning system provided by an embodiment of the present application.
  • FIG. 9 is a data logic block diagram when the software module of the combined positioning system provided by the embodiment of the present application is applied to calibrate the installation deviation angle between the IMU and the GNSS antenna;
  • FIG. 10 is a data logic block diagram when the software module of the combined positioning system provided by the embodiment of the present application is applied to calibrate the installation deviation angle between the IMU and the LiDAR.
  • Autonomous vehicles also known as unmanned vehicles, computer-driven vehicles or wheeled mobile robots.
  • Autonomous vehicles can sense their surrounding environment with technologies such as radar, global navigation satellite system (GNSS), and machine vision, and determine their own position, plan navigation routes, update map information, and avoid obstacles based on the sensing data. , and ultimately achieve autonomous vehicle driving without any or little human active operation.
  • technologies such as radar, global navigation satellite system (GNSS), and machine vision, and determine their own position, plan navigation routes, update map information, and avoid obstacles based on the sensing data. , and ultimately achieve autonomous vehicle driving without any or little human active operation.
  • GNSS global navigation satellite system
  • autonomous vehicles need to address three core questions of driving: Where? (vehicle location); where to go? (determine the destination); how to get there? (route plan).
  • the positioning system of autonomous vehicles is mainly used to solve the problem of "where?". Therefore, as one of the key modules of an autonomous vehicle, the performance of the positioning system will directly determine the stability and reliability of the autonomous vehicle.
  • positioning systems usually include sensors for vehicle positioning, such as inertial measurement units, IMUs, GNSS receivers, LiDAR, odometers, vision sensors, etc., in order to solve the "where?" problem.
  • the IMU can include, for example, a three-axis gyroscope and an accelerometer, which can be used to measure the angular velocity and acceleration of an autonomous vehicle; a GNSS receiver can be used to measure GNSS satellite signals, and LiDAR can be used to measure the distance between the vehicle and other objects. Since different sensors are used to measure different data, each sensor application has both advantages and limitations in the positioning system. It is difficult to use any sensor alone to meet the requirements of high positioning accuracy, strong anti-interference ability, and fast data update.
  • the current positioning system usually combines the data of the above-mentioned multiple sensors, and calculates the position and attitude of the vehicle according to the combined positioning algorithm, so as to achieve the effect of learning from each other's strengths and complementing each other's advantages, thereby improving the stability and reliability of the positioning system.
  • the data logic block diagram of the combined positioning system using the combined positioning algorithm is shown in Figure 1, wherein the combined positioning system outputs the position, velocity and attitude information based on the solution results of the INS 100, for example, according to the angular velocity and acceleration measured by the IMU 110 (hereinafter referred to as inertial navigation pose), the positioning result of the GNSS 120, the positioning result of the LiDAR 130, etc. are used as the input of the combined filter 140, and the filtering result is used to perform error correction on the solution result of the INS 100 to determine the final positioning result, Such as pose and its covariance.
  • the system variables of the combined filter 140 are generally inertial navigation-related errors (referred to as inertial navigation errors), such as: position error, velocity error, attitude error, accelerometer zero offset and gyroscope zero deviation, etc.
  • the combined filter 140 in this embodiment of the present application may be, for example, a Kalman filter (KF), which is a recursive filter (autoregressive filter) that can Estimate the state of a dynamic system from measurements that contain noise.
  • KF Kalman filter
  • the Kalman filter will consider the joint distribution of each time according to the value of each measurement at different times, and then generate an estimate of the unknown variable, so it is more accurate than the estimation method based on only a single measurement.
  • FIG. 2 is a schematic diagram of the positions of each sensor in the combined positioning system on the vehicle.
  • the sensors of the combined positioning system include an IMU 110, a LiDAR 130 and a GNSS antenna, and the GNSS antenna may include a GNSS master antenna 121 and a GNSS slave antenna 122, referred to as GNSS dual antennas.
  • the various sensors in the combined positioning system are usually installed in different positions of the vehicle, for example, the IMU 110 is installed in the middle of the vehicle, the GNSS dual antenna is installed on the top of the vehicle, and the GNSS main antenna 121 and the GNSS slave antenna 122 are distributed front and rear Set up to receive GNSS signals from satellites, the LiDAR 130 is mounted to the head of the vehicle to measure objects in front of the vehicle.
  • the IMU 110 is installed in the middle of the vehicle
  • the GNSS dual antenna is installed on the top of the vehicle
  • the GNSS main antenna 121 and the GNSS slave antenna 122 are distributed front and rear Set up to receive GNSS signals from satellites
  • the LiDAR 130 is mounted to the head of the vehicle to measure objects in front of the vehicle.
  • each sensor will describe its solution result or positioning result based on its own coordinate system. Since the installation positions of each sensor are different, the coordinate system of each sensor is also different.
  • the combined positioning system includes: an environment coordinate system (E, N, U), a LiDAR coordinate system (X l , Y l , Z l ), an IMU coordinate system (X b , Y b , Z b ), GNSS coordinate system (X g , Y g , Z g ) and rear axle ground projection coordinate system (X m , Y m , Z m ).
  • the environmental coordinate system (E, N, U) may be, for example, a station center coordinate system, also called an east-north-celestial coordinate system ENU.
  • the station center coordinate system takes the position of the target (such as the GNSS antenna) as the coordinate origin, including the E-axis pointing to the east, the N-axis pointing to the north direction, and the U-axis pointing to the sky.
  • LiDAR coordinate system (X l , Y l , Z l ), IMU coordinate system (X b , Y b , Z b ), GNSS coordinate system (X g , Y g , Z g ) and rear axle
  • the ground projection coordinate system (X m , Y m , Z m ) can all use the upper right coordinate system.
  • the upper right coordinate system takes the forward direction of the vehicle (such as a vehicle) as the Y-axis direction, the right side of the vehicle as the X-axis direction, and the upward direction perpendicular to the X-axis and the Y-axis as the Z-axis
  • the origin of each coordinate system can be the center of its corresponding sensor.
  • the installation position relationship between the sensors may include arm parameters and installation deviation angles.
  • the lever arm parameter mainly describes the installation position distance between each sensor, which can generally be measured directly by measuring equipment;
  • the installation deviation angle mainly describes the installation angle deviation between the various sensors, taking into account the installation position relationship of the sensors on the vehicle and the difference between the sensors. Due to its own structural characteristics, the installation deviation angle between each sensor is generally difficult to directly and accurately measure by measuring equipment.
  • the installation deviation angle can include the installation deviation angle between the IMU and the GNSS antenna, and the installation deviation angle between the IMU and the LiDAR.
  • FIG. 3 is a schematic diagram of the installation deviation angle shown in the embodiment of the present application.
  • the heading installation deviation angle between the IMU 110 and the GNSS antenna refers to the connection between the GNSS dual antennas (from the GNSS main antenna 121 to the GNSS antenna).
  • the heading installation deviation angle between the IMU 110 and the LiDAR 130 refers to the forward axis (yl) of the LiDAR coordinate system and the forward axis of the IMU coordinate system ( yb) of the included angle ⁇ lb.
  • the calibration of the installation deviation angle between the IMU and the GNSS antenna mainly adopts the offline calibration method, and most of them are realized by optical means.
  • the heading installation deviation angle as an example, the headings of the IMU and the GNSS antenna can be calibrated respectively by optical means, and then the deviation value of the two headings is calculated as the heading installation deviation angle between the IMU and the GNSS antenna.
  • the phase center of the GNSS antenna is located inside the antenna, it is difficult to accurately measure the axes of the dual antennas, so the calibration accuracy of this method is low; moreover, this method needs to rely on dedicated calibration tools, such as laser scanners, theodolites, etc., to operate It is cumbersome and the calibration time is long; in addition, when the relationship between the installation position of the IMU and the GNSS changes due to long-term vibration, bumps, etc., it needs to be re-calibrated, otherwise the positioning accuracy of the combined positioning system will be affected.
  • the calibration method of the installation deviation angle between the IMU and the LiDAR is also mainly an offline calibration method, and most of them use an indirect calibration method.
  • the installation deviation angle of the heading as an example, you can use the measurement common point or auxiliary measurement coordinate system to establish the relationship between the IMU and the LiDAR.
  • This method also needs to rely on a dedicated calibration tool, which is cumbersome to operate and takes a long time to calibrate.
  • the relationship between the installation position of the IMU and the LiDAR changes, it needs to be calibrated again, otherwise the positioning accuracy of the combined positioning system will be affected.
  • an embodiment of the present application provides a method for calibrating the installation deviation angle of a sensor.
  • the combined positioning system in this embodiment of the present application may be a positioning system including two or more positioning sensors, and the positioning system uses the calculation results or positioning results of the above two or more sensors to comprehensively determine the final positioning results.
  • the embodiments of the present application do not specifically limit which positioning sensors the combined positioning system includes.
  • the combined positioning system may include IMU and GNSS; in other embodiments, the combined positioning system may include IMU and LiDAR ; In other embodiments, the combined positioning system may include IMU, LiDAR and GNSS; in other embodiments, the combined positioning system may include other sensors in addition to the above positioning sensors.
  • FIG. 4 is a hardware structure diagram of a combined positioning system provided by an embodiment of the present application.
  • the combined positioning system may include: IMU 110, GNSS 120, LiDAR 130, odometer 140, vision sensor 150 and other sensors, processor 160, memory 170 and communication module 180, etc.
  • sensors such as IMU 110, GNSS 120, LiDAR 130 are used to measure and receive various measurement data
  • the processor 160 is used to calculate the measurement data according to the corresponding combined positioning algorithm, such as inertial navigation calculation, Kalman Filter calculation, etc., to obtain positioning results.
  • the processor 160 may include, for example, one or more processing units, such as a central processing unit (CPU), a microcontroller (MCU), an image signal processor (ISP), and a neural network processor. (neural-network processing unit, NPU), etc. Wherein, different processing units may be independent devices, or may be integrated in one or more processors, such as integrated in a system on a chip (system on a chip, SoC).
  • a memory may also be provided in the processor 160 for storing instructions and data. In some embodiments, the memory in the processor is a cache memory. This memory can hold instructions or data that the processor has just used or recycled.
  • the memory 170 may include one or more storage units, for example, may include volatile memory (volatile memory), such as: dynamic random access memory (dynamic random access memory, DRAM), static random access memory (static random access memory, SRAM), etc.; may also include non-volatile memory (non-volatile memory, NVM), such as: read-only memory (read-only memory, ROM), flash memory (flash memory), and the like.
  • volatile memory volatile memory
  • DRAM dynamic random access memory
  • static random access memory static random access memory
  • SRAM static random access memory
  • NVM non-volatile memory
  • the memory 170 may be used to store program codes and instructions executable by the processor, such as: program codes and instructions for processing sensor data, programs and instructions for inertial navigation solutions, programs for implementing Kalman filtering algorithms, and instructions etc.
  • the memory 170 may also store an in-vehicle operating system of the autonomous driving computing platform, such as a Harmony OS, a UQX operating system, and the like.
  • the communication module 180 is used to enable the vehicle to communicate with the outside world in various network modes of 2G, 3G, 4G and 5G, for example, to realize vehicle-to-everything (V2X) of the vehicle, etc.
  • the communication module may include, for example, a baseband chip , power amplifiers and antennas, etc.
  • some or all of the components of the communication module 180 may be integrated with one or more processors, for example, a baseband chip may be integrated with a CPU, NPU, ISP, etc. in an SoC.
  • the combined positioning system may also include a communication interface, such as a controller area network (CAN) interface, allowing the processor to communicate with other chips and devices of the vehicle, such as an electronic control unit (ECU).
  • a communication interface such as a controller area network (CAN) interface
  • ECU electronice control unit
  • the hardware structure of the combined positioning system illustrated in the embodiments of the present application does not constitute a specific limitation on the combined positioning system.
  • the hardware structure of the combined positioning system may include more or less components than shown, or some components are combined, or some components are split, or different components are arranged.
  • the illustrated components may be implemented in hardware, software, or a combination of software and hardware.
  • the combined positioning system in the embodiments of the present application can be installed on various types of vehicles applying automatic driving technology or navigation and positioning, including but not limited to various vehicles: for example, vehicles (cars), ships, trains, subways, airplanes, etc. , and various robots, such as: service robots, transport robots, autonomous guided vehicles (AGVs), unmanned ground vehicles (UGVs), etc., as well as various construction machinery, such as tunnel boring machines Wait.
  • various vehicles for example, vehicles (cars), ships, trains, subways, airplanes, etc.
  • various robots such as: service robots, transport robots, autonomous guided vehicles (AGVs), unmanned ground vehicles (UGVs), etc.
  • AGVs autonomous guided vehicles
  • UUVs unmanned ground vehicles
  • construction machinery such as tunnel boring machines Wait.
  • the following is an example to illustrate the technical solution of the method for calibrating the sensor installation deviation angle provided by the embodiment of the present application by taking the combined positioning system including sensors such as IMU, LiDAR, and GNSS as an example.
  • the embodiment of the present application expands the state variables of the current combined positioning system.
  • the data logic block diagram of the expanded combined positioning system is shown in FIG. 5 .
  • the guided pose, the positioning result of GNSS 120, the positioning result of LiDAR 130, etc. are used as the input of the combined positioning filter, and the original inertial navigation related errors, such as: position error, velocity error, attitude error, accelerometer zero deviation and
  • the installation deviation angle between the IMU and the GNSS antenna, as well as the installation deviation angle between the IMU and the LiDAR are also expanded into the system variables of the combined positioning system, so that the IMU and the IMU can be adjusted by the combined filter.
  • the installation deviation angle between the GNSS antennas and the installation deviation angle between the IMU and the LiDAR are calibrated online; when the state variables of the combined positioning filter are converged, the installation deviation between the IMU and the GNSS antenna in the online calibration result is used for online calibration. Angle, and the installation deviation angle between IMU and LiDAR to compensate the positioning results of GNSS and LiDAR to improve the positioning accuracy.
  • the method for calibrating the installation deviation angle of the sensor provided by the present application may specifically consist of two parts.
  • the first part includes the calibration of the installation deviation angle of the IMU and the GNSS dual antenna, and the second part includes the installation deviation between the IMU and the LiDAR. Angle calibration square.
  • FIG. 6 is a flowchart showing that the method for calibrating the installation deviation angle of the sensor provided by the embodiment of the present application is applied to the calibration of the installation deviation angle between the IMU and the GNSS antenna. As shown in Figure 6, the method may include the following steps:
  • Step S101 acquiring the measurement result of the IMU.
  • the measurement result of the IMU may include, for example, angular velocity and acceleration information measured by the IMU.
  • the vehicle can be maneuvered, such as acceleration, deceleration, constant speed driving, left turn and right turn, etc.
  • the combined positioning system can The angular velocity and acceleration information measured by the IMU is acquired in real time and continuously.
  • step S102 the measurement result of the IMU is calculated to obtain the attitude information of the INS.
  • the attitude information of INS can be expressed by attitude matrix.
  • information such as the position and speed of the INS can also be obtained.
  • the above position, velocity and attitude information can also be collectively referred to as inertial navigation pose.
  • the combined positioning system in the process of vehicle maneuvering, can solve the measurement result data of the IMU in real time and continuously, so as to continuously obtain the current position, speed and attitude information of the vehicle.
  • the initial position and initial speed of the vehicle in the initial stage of the solution, can be determined according to the GNSS positioning result as the initial data of the IMU; then, the combined positioning system can solve the solution according to the initial data and the measurement result data of the current IMU. Calculate the latest position, speed and attitude information of the vehicle. In the non-initial stage of the calculation, the combined positioning system can obtain the latest position, speed and attitude information of the vehicle according to the position, speed and attitude information of the vehicle obtained by the previous calculation and the measurement result data of the current IMU.
  • step S103 an error model of the installation deviation angle between the INS and the GNSS antenna is constructed according to the attitude information of the INS and the attitude information of the GNSS.
  • the error model is described by the state variables of the combined positioning filter.
  • the embodiment of the present application may construct an attitude matrix corresponding to the INS And, the attitude matrix corresponding to GNSS
  • attitude matrix For example, it can be a 3 ⁇ 3 matrix, which indicates that the pose output by the INS is mapped from the rear axle ground projection coordinate system (Xm, Ym, Zm) to the environmental coordinate system (E, N, U) through translation and rotation.
  • the attitude matrix corresponding to the above-mentioned IMU is denoted as
  • attitude matrix For example, it can be a 3 ⁇ 3 matrix, which indicates that the pose of the GNSS output is mapped from the GNSS coordinate system (Xg, Yg, Zg) to the environment coordinate system (E, N, U) through translation and rotation.
  • attitude matrix corresponding to the IMU Attitude matrix corresponding to GNSS It can be expressed as the following formula (1):
  • the attitude matrix parameters in etc. are used to represent the translation and rotation parameters required to map the pose output of the INS from the rear wheel axle ground projection coordinate system (Xm, Ym, Zm) to the environment coordinate system (E, N, U);
  • the attitude matrix parameters in Etc. is used to represent the translation and rotation parameters required to map the pose of the GNSS output from the GNSS coordinate system (Xg, Yg, Zg) to the environment coordinate system (E, N, U).
  • the roll angle output by the INS is expressed as ⁇ m
  • the pitch angle is expressed as ⁇ m
  • the heading angle is expressed as
  • the roll angle of the GNSS positioning output is expressed as ⁇ g
  • the pitch angle is expressed as ⁇ g
  • the heading angle is expressed as Then, the observed quantity of this error model can be expressed as the following formula (2)
  • ⁇ 0 represents the deviation of the roll angle output by the IMU and GNSS
  • ⁇ 0 represents the deviation of the pitch angle output by the IMU and GNSS
  • the above deviation is referred to as the attitude deviation between the IMU and the GNSS antenna in this embodiment of the present application.
  • the heading angle has an obvious influence on the accuracy of the positioning result. Therefore, in the embodiment of the present application, when outputting the calibration results subsequently, only the output IMU and GNSS may be considered. Heading mount deviation angle between dual antennas.
  • this embodiment of the present application can map the attitude matrix corresponding to the IMU Perform the following transformations to construct the transfer relationship equation from the attitude matrix corresponding to INS to the attitude matrix corresponding to GNSS to establish its attitude matrix corresponding to GNSS the relationship between.
  • the attitude matrix can be decomposed into two transformation matrices, namely and and in,
  • it can be a 3 ⁇ 3 matrix, which represents the mapping of the pose from the environment coordinate system (E, N, U) to the erroneous environment coordinate system (Xn', Yn', Zn') through translation and rotation, etc.
  • it can be a 3 ⁇ 3 matrix, which represents the mapping of the pose from the rear wheel axle ground projection coordinate system (Xm, Ym, Zm) to the environment coordinate system (E, N, U) through translation and rotation.
  • transformation matrices can be further decomposed into two transformation matrices, namely and and in,
  • it can be a 3 ⁇ 3 matrix, which represents mapping the pose from the GNSS coordinate system (Xg, Yg, Zg) to the environment coordinate system (E, N, U) through translation and rotation;
  • it can be a 3 ⁇ 3 matrix, which represents the mapping of the pose from the rear axle ground projection coordinate system (Xm, Ym, Zm) to the GNSS coordinate system (Xg, Yg, Zg) through translation and rotation.
  • the embodiment of the present application can introduce the attitude error of the INS and the installation deviation angle between the IMU and the GNSS antenna into the state variables of the combined positioning filter based on the formula (3).
  • attitude error of the INS described in the east-north-sky coordinate system Can be:
  • the installation deviation angle ⁇ between the IMU and the GNSS antenna may be:
  • ⁇ 0 ⁇ represents the cross product of ⁇ 0 .
  • the above formula (4) can be used as a transfer relation equation from the attitude matrix corresponding to INS to the attitude matrix corresponding to GNSS.
  • This formula (6) can be called an error model of the installation deviation angle between the IMU and the GNSS dual antenna.
  • Z 1 , Z 2 , Z 3 are the observed quantities of the error model; is the attitude error of the aforementioned INS, such as the platform error angle; is the installation deviation angle between the aforementioned IMU and the GNSS antenna.
  • step S104 a system state equation is constructed, and the state variables of the system state equation include the installation deviation angle between the INS and the GNSS antenna.
  • the original state equation in the combined positioning system is an equation constructed based on the inertial navigation error model, and the state variables in the state equation are the inertial navigation error, such as: position error, velocity error, attitude error, accelerometer zero deviation and gyro
  • the error model of the installation deviation angle between the IMU and the GNSS antenna can be introduced into the original state equation, so that the state variables of the new state equation include both the inertial navigation error-related variables , also includes the error of the installation deviation angle between the IMU and the GNSS antenna.
  • the embodiment of the present application defines a new system state equation of the combined positioning system including:
  • X 1 and X 2 are state variables
  • X 1 is the inertial navigation error
  • X 2 is the installation deviation angle between the IMU and the GNSS antenna.
  • G 1 , W 1 are the inertial navigation error driving noise, where, G 1 may be the noise driving matrix of the inertial navigation error, and W 1 may be the state noise matrix of the inertial navigation error.
  • G 1 and W 1 are Gaussian white noise.
  • ⁇ v E , ⁇ v N , and ⁇ v U are the velocity errors of the INS in the east, north, and sky directions, respectively
  • ⁇ , ⁇ L, and ⁇ H are the longitude, latitude, and altitude errors of the INS, respectively
  • ⁇ x , ⁇ y , ⁇ z are the components of the random constant drift of the gyroscope in the X, Y, and Z axes
  • X 1 parameter items included in the above X 1 are only used as an example, and do not constitute a limitation on X 1. In some other implementation manners, X 1 may include more parameter items or fewer parameter items. The application examples are not described in detail here.
  • G 2 and W 2 are both Gaussian white noise.
  • X 2 is the installation deviation angle in the X-axis direction
  • Y-axis direction is the installation deviation angle in the Y-axis direction
  • Z-axis direction is the installation deviation angle in the Z-axis direction.
  • the parameter items included in the above X 2 are only used as an example, and do not constitute a limitation on X 2. In some other implementation manners, X 2 may include more parameter items or fewer parameter items. The application examples are not described in detail here.
  • formulas (7)-(9) constitute the system state equation of the combined positioning system of the embodiment of the present application. It can be seen that the state variables of the system state equation not only include the inertial navigation error, but also include the IMU and GNSS. The installation deviation angle between the antennas, thus, when the combined positioning filter is used to estimate the state variables of the system state equation online, the installation deviation angle between the IMU and the GNSS antenna can be calibrated online at the same time.
  • step S105 a system observation equation is constructed, and the observation amount of the system observation equation includes an observation matrix corresponding to the installation deviation angle between the INS and the GNSS antenna determined according to the error model.
  • the observed quantities of the observation equation of the system are the observation matrix corresponding to the inertial navigation error and the observation matrix corresponding to the error of the installation deviation angle between the IMU and the GNSS antenna.
  • system observation equation may be:
  • H is the system observation matrix
  • X is the new system state equation established in step S105 in the embodiment of the present application
  • V is the system observation noise matrix.
  • H 1 is the attitude error corresponding to the observation matrix
  • H 2 in addition to the error model for the attitude error of the other state variables corresponding to the observation matrix
  • H 3 mounted observation matrix corresponding to the angle deviation between the IMU and GNSS antenna.
  • H 2 can be obtained according to the inertial navigation error model
  • H 1 and H 3 are as follows:
  • step S106 the combined positioning filter is used to filter and solve the system state equation and the system observation equation, so as to obtain the optimal estimation result of the installation deviation angle between the IMU and the GNSS antenna.
  • the combined positioning filter may be a Kalman filter, that is, Kalman filtering is performed on the system state equation and the system observation equation to obtain the optimal estimation result of the heading installation deviation angle between the IMU and the GNSS.
  • Step S107 Compensate the GNSS positioning result according to the optimal estimation result of the installation deviation angle between the IMU and the GNSS antenna.
  • the embodiment of the present application can also correct the inertial navigation solution results, such as the position, velocity and attitude, according to the optimal estimation result of the state variable obtained by the Kalman filter, so as to correct the accumulated error of the inertial navigation solution results and improve the combination The positioning accuracy of the positioning system.
  • Step S108 when the filter solution result of the combined positioning filter pair converges, save the optimal estimation result of the installation deviation angle between the IMU and the GNSS antenna as the calibration result in the configuration file.
  • the installation deviation angle between the IMU and the GNSS antenna will not change, and the result of the Kalman filter pair will always be in a state of convergence. ; If the vehicle vibrates during driving or vehicle maintenance causes the installation position relationship between the IMU and GNSS, then the Kalman filter pair will enter the convergence state again after a short period of non-convergence. , at this time, the optimal estimation result of the installation deviation angle between the IMU and the GNSS antenna will change, and the combined positioning system can save the latest calibration result to the configuration file after the Kalman filter reconverges.
  • the method of the embodiment of the present application integrates the installation deviation angle between the IMU and the GNSS antenna into the state variable of the combined positioning system, and realizes the adjustment between the IMU and the GNSS antenna of the combined positioning system during the driving process of the vehicle.
  • the installation deviation angle is calibrated (that is, online calibration).
  • the calibration process does not need to consider the physical structure of the IMU and GNSS, and the calibration accuracy is high.
  • the method of the embodiment of the present application does not require the use of a calibration tool, and only requires the vehicle to generate a maneuver to complete the calibration, and the operation is simple and fast.
  • the method in this embodiment of the present application can update the calibration result online, thereby always ensuring the accuracy of the combined positioning system.
  • FIG. 7 is a flowchart of the application of the method for calibrating the installation deviation angle of the sensor provided by the embodiment of the present application to the calibration of the installation deviation angle of the IMU and the LiDAR. As shown in Figure 7, the method may include the following steps:
  • Step S201 acquiring the measurement result of the IMU.
  • the measurement result of the IMU may include, for example, angular velocity and acceleration information measured by the IMU.
  • step S202 the measurement result of the IMU is calculated to obtain the attitude information of the INS.
  • step S201 and step S202 For the specific implementation manner of step S201 and step S202, reference may be made to step S101 and step S102, which will not be repeated here.
  • Step S203 constructing an error model of the installation deviation angle between the INS and the LiDAR according to the attitude information of the INS and the attitude information of the LiDAR.
  • the parameters of the error model can be expressed by combining the state variables of the positioning filter.
  • the embodiment of the present application may construct an attitude matrix corresponding to the INS And the attitude matrix corresponding to LiDAR
  • attitude matrix For example, it can be a 3 ⁇ 3 matrix, which indicates that the pose output by LiDAR is mapped from the LiDAR coordinate system (Xl, Yl, Zl) to the environment coordinate system (E, N, U) through translation and rotation.
  • attitude matrix corresponding to the IMU Attitude matrix corresponding to LiDAR It can be expressed as the following formula (10):
  • the attitude matrix parameters in etc. are used to represent the translation and rotation parameters required to map the pose output of the INS from the rear wheel axle ground projection coordinate system (Xm, Ym, Zm) to the environment coordinate system (E, N, U);
  • the attitude matrix parameters in Etc. is used to represent the translation and rotation parameters required to map the LiDAR output pose from the LiDAR coordinate system (Xl, Yl, Zl) to the environment coordinate system (E, N, U).
  • the roll angle calculated by the IMU is expressed as ⁇ m
  • the pitch angle is expressed as ⁇ m
  • the heading angle is expressed as
  • the roll angle of LiDAR positioning output is expressed as ⁇ l
  • the pitch angle is expressed as ⁇ l
  • the heading angle is expressed as Then, the attitude observation of the combined positioning system can be expressed as the following formula (11)
  • ⁇ 1 represents the deviation of the roll angle output by the IMU and LiDAR
  • ⁇ 1 represents the deviation of the pitch angle output by the IMU and LiDAR
  • the above deviation is referred to as the attitude deviation between the IMU and the LiDAR in this embodiment of the present application.
  • the heading angle has an obvious influence on the accuracy of the positioning result. Therefore, when outputting the calibration results in the following embodiments of the present application, only the output of the IMU and the LiDAR may be considered. Heading installation deviation angle between.
  • this embodiment of the present application can map the attitude matrix corresponding to the INS Perform the following transformations to construct the transfer relationship equation from the attitude matrix corresponding to INS to the attitude matrix corresponding to LiDAR to establish its attitude matrix corresponding to GNSS the relationship between.
  • the attitude matrix can be decomposed into two transformation matrices, namely and and in,
  • it can be a 3 ⁇ 3 matrix, which represents the mapping of the pose from the environment coordinate system (E, N, U) to the erroneous environment coordinate system (Xn', Yn', Zn') through translation and rotation, etc.
  • it can be a 3 ⁇ 3 matrix, which represents the mapping of the pose from the rear wheel axle ground projection coordinate system (Xm, Ym, Zm) to the environment coordinate system (E, N, U) through translation and rotation.
  • transformation matrices can be further decomposed into two transformation matrices, namely and and in,
  • it can be a 3 ⁇ 3 matrix, which represents mapping the pose from the LiDAR coordinate system (X l , Y l , Z l ) to the environmental coordinate system (E, N, U) through translation and rotation;
  • it can be a 3 ⁇ 3 matrix, which represents the mapping of the pose from the rear axle ground projection coordinate system (Xm, Ym, Zm) to the LiDAR coordinate system (X l , Y l , Z l ) through translation and rotation, etc. .
  • the embodiment of the present application can introduce the attitude error of the INS and the installation deviation angle between the IMU and the LiDAR into the state variable of the combined positioning filter based on the formula (12).
  • attitude error of the INS described in the east-north-sky coordinate system Can be:
  • the installation deviation angle ⁇ 1 between the IMU and the LiDAR may be:
  • ⁇ 1 ⁇ represents the cross product of ⁇ 1.
  • the above formula (13) can be used as the transfer relation equation from the attitude matrix corresponding to INS to the attitude matrix corresponding to LiDAR.
  • This formula (15) can be referred to as an error model of the installation deviation angle between the IMU and the LiDAR.
  • Z 1 , Z 2 , Z 3 are the observed quantities of the error model; is the attitude error of the aforementioned INS, such as the platform error angle; is the installation deviation angle between the aforementioned IMU and LiDAR.
  • step S204 a system state equation is constructed, and the state variables of the system state equation include the installation deviation angle between the INS and the LiDAR.
  • the original state equation in the combined positioning system is an equation constructed based on the inertial navigation error model, and the state variables in the state equation are the inertial navigation error, such as: position error, velocity error, attitude error, accelerometer zero deviation and gyro
  • the error model of the installation deviation angle between the IMU and the LiDAR can be introduced into the original state equation, so that the state variables of the new state equation include both the inertial navigation error-related variables, Also included are variables related to the error in the mounting deviation angle between the IMU and the LiDAR.
  • the embodiment of the present application defines a new system state equation of the combined positioning system including:
  • X 1 is the inertial navigation error in the state variable
  • X 3 is the installation deviation angle between the IMU and the LiDAR.
  • G 1 , W 1 are the inertial navigation error driving noise, where, G 1 may be the noise driving matrix of the inertial navigation error, and W 1 may be the state noise matrix of the inertial navigation error.
  • G 1 and W 1 are Gaussian white noise.
  • ⁇ v E , ⁇ v N , and ⁇ v U are the velocity errors of the INS in the east, north, and sky directions, respectively
  • ⁇ , ⁇ L, and ⁇ H are the longitude, latitude, and altitude errors of the INS, respectively
  • ⁇ x , ⁇ y , ⁇ z are the components of the random constant drift of the gyroscope in the X, Y, and Z axes
  • X 1 parameter items included in the above X 1 are only used as an example, and do not constitute a limitation on X 1. In some other implementation manners, X 1 may include more parameter items or fewer parameter items. The application examples are not described in detail here.
  • G 3 , W 3 is the error driving noise of the installation deviation angle between IMU and LiDAR, where G 3 can be the noise driving matrix of the error of the installation deviation angle between IMU and LiDAR, W 3 can be the difference between IMU and LiDAR
  • the state noise matrix of the error of the installation deviation angle, optionally, G 3 and W 3 are both Gaussian white noise.
  • X 3 is the installation deviation angle in the X-axis direction between the IMU and the LiDAR
  • Y-axis direction between the IMU and the LiDAR is the installation deviation angle in the Y-axis direction between the IMU and the LiDAR.
  • Z-axis direction between the IMU and the LiDAR is the installation deviation angle in the Z-axis direction between the IMU and the LiDAR.
  • formulas (16)-(18) constitute the system state equation of the combined positioning system of the embodiment of the present application. It can be seen that the state variables of the system state equation not only include the inertial navigation error, but also include the IMU and LiDAR Therefore, when the combined positioning filter is used to estimate the state variables of the system state equation online, the installation deviation angle between the IMU and the LiDAR can be calibrated online at the same time.
  • step S205 a system observation equation is constructed, and the observation amount of the system observation equation includes an observation matrix corresponding to the installation deviation angle between the INS and the LiDAR determined according to the error model.
  • the observed quantities of the observation equation of the system are the observation matrix corresponding to the inertial navigation error and the observation matrix corresponding to the error of the installation deviation angle between the IMU and the LiDAR.
  • system observation equation may be:
  • H is the system observation matrix
  • X is the new system state equation established in step S205 in the embodiment of the present application
  • V is the system observation noise matrix.
  • H 1 is the attitude error corresponding to the observation matrix
  • H 2 in addition to the error model for the attitude error of the other state variables corresponding to the observation matrix
  • H 3 is mounted between the deviation angle corresponding to the IMU and the LiDAR observation matrix.
  • H 2 can be obtained according to the inertial navigation error model, and H 1 and H 3 are as follows:
  • step S206 the combined positioning filter is used to filter and solve the system state equation and the system observation equation, so as to obtain the optimal estimation result of the installation deviation angle between the IMU and the LiDAR.
  • the combined positioning filter can be a Kalman filter, that is, Kalman filtering is performed on the system state equation and the system observation equation to obtain the optimal estimation result of the installation deviation angle between the IMU and the LiDAR.
  • Step S207 the LiDAR positioning result is compensated according to the optimal estimation result of the installation deviation angle between the IMU and the LiDAR.
  • the embodiment of the present application can also correct the inertial navigation solution results, such as the position, velocity and attitude, according to the optimal estimation result of the state variable obtained by the Kalman filter, so as to correct the accumulated error of the inertial navigation solution results and improve the combination The positioning accuracy of the positioning system.
  • Step S208 when the filter solution result of the combined positioning filter pair converges, save the optimal estimation result of the installation deviation angle between the IMU and the LiDAR as the calibration result in the configuration file.
  • the installation position relationship between the IMU and the LiDAR does not change, then the installation deviation angle between the IMU and the LiDAR will not change, and the filtering solution result of the Kalman filter pair will always be in a convergent state; If the vehicle vibrates during driving or vehicle maintenance causes the installation position relationship between the IMU and the LiDAR, the Kalman filter pair will enter the convergence state again after a short period of non-convergence. At this time, the optimal estimation result of the installation deviation angle between the IMU and the LiDAR will change, and the combined positioning system can save the latest calibration result to the configuration file after the Kalman filter reconverges.
  • the method of the embodiment of the present application integrates the installation deviation angle between the IMU and the LiDAR into the state variable of the combined positioning system, and realizes the installation deviation between the IMU and the LiDAR of the combined positioning system during the driving process of the vehicle.
  • the angle is calibrated (that is, online calibration), the calibration process does not need to consider the physical structure of the IMU and LiDAR, and the calibration accuracy is high.
  • the method of the embodiment of the present application does not require the use of a calibration tool, and only requires the vehicle to generate a maneuver to complete the calibration, and the operation is simple and fast.
  • the method of the embodiment of the present application can update the calibration result online, thereby always ensuring the positioning accuracy of the combined positioning system.
  • the method in the embodiments of the present application can be applied to a combined positioning system including an IMU and a GNSS to calibrate the installation deviation angle between the IMU and the GNSS dual antennas;
  • the combined positioning system may include corresponding hardware structures and/or software modules for performing each function.
  • the present application can be implemented in hardware or a combination of hardware and computer software with the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein. Whether a function is performed by hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Skilled artisans may implement the described functionality using different methods for each particular application, but such implementations should not be considered beyond the scope of this application.
  • the combined positioning system can implement corresponding functions through the hardware structure shown in FIG. 4 .
  • each sensor can be installed in different positions of the vehicle, and the memory and processor can be, for example, a trip computer of the vehicle, such as an electronic control unit (ECU), or an autonomous driving computing platform, such as a mobile data center (mobile data center) , MDC) memory and processor.
  • ECU electronice control unit
  • MDC mobile data center
  • the memory includes program instructions for implementing corresponding functions, and when the program instructions are run by the processor, the combined positioning system is used to perform the following steps: constructing an error model of the installation deviation angle between the sensors according to the attitude information output by the sensors ; Construct the system state equation, the state variables of the system state equation include the installation deviation angle; construct the system observation equation, the observation amount of the system observation equation includes the observation matrix corresponding to the installation deviation angle determined according to the error model; the system state equation and the system observation equation Perform the filtering calculation, and take the estimation result of the installation deviation angle obtained when the filtering calculation result converges as the calibration result.
  • the combined positioning system is used to calibrate the installation deviation angle between the IMU and the GNSS antenna
  • the plurality of sensors may include at least an IMU and a GNSS receiver
  • the GNSS receiver includes GNSS antennas, such as a master antenna and a slave antenna, Referred to as GNSS dual antenna, each antenna of IMU and GNSS is installed in different positions of the vehicle; when the program instructions are run by the processor, the combined positioning system is used to perform the following steps, so as to realize the construction of each sensor according to the attitude information output by each sensor
  • the error model of the installation deviation angle between the two According to the attitude information output by the INS and GNSS, construct the observation amount of the error model, and the observation amount is the attitude deviation between the IMU and the GNSS antenna; construct the attitude matrix from the INS corresponding to the GNSS corresponding The transfer relationship equation of the attitude matrix; the observation amount and transfer relationship equation of the simultaneous error model build an error model, and the error model is the error model of the installation deviation angle between the
  • the combined positioning system is used to perform the following steps, so as to realize the filtering and solving of the system state equation and the system observation equation, and the installation deviation obtained when the filtering and solving results converge.
  • the optimal estimation result of the angle is used as the calibration result: Kalman filtering is performed on the system state equation and the system observation equation to obtain the optimal estimation result of the installation deviation angle between the IMU and the GNSS antenna; The positioning result is compensated, and the inertial navigation solution result of the INS is corrected; when the filtering solution result converges, the optimal estimation result of the installation deviation angle between the IMU and the GNSS antenna is saved as the calibration result in the configuration file.
  • the combined positioning system is used to calibrate the installation deviation angle between the IMU and the LiDAR, so the plurality of sensors at least include the IMU and the LiDAR, and the IMU and the LiDAR are installed in different positions of the vehicle; when the program instructions are executed by the processor,
  • the combined positioning system is used to perform the following steps to construct an error model of the installation deviation angle between each sensor according to the attitude information output by each sensor: According to the attitude information output by INS and LiDAR, construct the observation amount of the error model, observe The quantity is the attitude deviation between the IMU and LiDAR; the transfer relationship equation from the attitude matrix corresponding to the INS to the attitude matrix corresponding to the LiDAR is constructed; the observation quantity and the transfer relationship equation of the simultaneous error model are used to construct the error model, and the error model is the IMU and LiDAR.
  • the error model of the installation deviation angle between the two wherein, the attitude information includes one or more of roll angle, pitch angle and heading angle, and the transfer relationship equation includes the attitude error of the INS and the
  • the combined positioning system is used to perform the following steps, so as to realize the filtering and solving of the system state equation and the system observation equation, and the installation deviation obtained when the filtering and solving results converge.
  • the optimal estimation result of the angle is used as the calibration result: Kalman filtering is performed on the system state equation and system observation equation to obtain the optimal estimation result of the installation deviation angle between the IMU and the LiDAR; the positioning of the LiDAR is based on the optimal estimation result.
  • the results are compensated, and the inertial navigation solution results of the INS are corrected; when the filter solution results converge, the optimal estimation result of the installation deviation angle between the IMU and the LiDAR is saved as the calibration result in the configuration file.
  • the combined positioning system may implement corresponding functions through the software modules shown in FIG. 8 .
  • the combined positioning system may include an inertial navigation solution module 310 , an error model building module 320 , a system state equation building module 330 , a system state equation building module 340 and a filter solution module 350 .
  • the functions of the above modules are described in detail below:
  • the inertial navigation calculation module 310 is configured to perform inertial navigation calculation on the measurement result of the IMU to obtain the attitude information of the INS.
  • the error model building module 320 is used to build the error model of the installation deviation angle between the INS and the GNSS antenna according to the attitude information of the INS and the attitude information of the GNSS;
  • a system state equation building module 330 configured to build a system state equation, the state variables of the system state equation include the installation deviation angle between the INS and the GNSS antenna;
  • the system observation equation building module 340 is configured to construct a system observation equation, and the observation amount of the system observation equation includes an observation matrix corresponding to the installation deviation angle between the INS and the GNSS antenna determined according to the error model.
  • the filtering and solving module 350 is used for filtering and solving the system state equation and the system observation equation, and takes the estimation result of the installation deviation angle between the INS and the GNSS antenna obtained when the filtering and solving result converges as the calibration result.
  • the error model construction module 320 can be specifically configured to: construct the observation amount of the error model according to the attitude information output by the INS and the GNSS; construct the transfer relation equation from the attitude matrix corresponding to the INS to the attitude matrix corresponding to the GNSS ; Simultaneous observation and transfer relationship equations of the error model to construct an error model; wherein, the attitude information includes one or more of roll angle, pitch angle and heading angle, and the transfer relationship equation includes the attitude error of the INS and the relationship between the IMU and the GNSS antenna. installation deviation angle.
  • the filtering and solving module 350 may be specifically configured to: perform Kalman filtering and solving on the system state equation and the system observation equation to obtain the optimal estimation result of the installation deviation angle between the IMU and the GNSS antenna;
  • the optimal estimation result compensates the positioning result of GNSS and corrects the inertial navigation solution result of INS; when the filtering solution result converges, the optimal estimation result of the installation deviation angle between IMU and GNSS antenna is used as calibration
  • the results are saved to the configuration file.
  • the inertial navigation calculation module 310 is configured to perform inertial navigation calculation on the measurement result of the IMU to obtain the attitude information of the INS.
  • the error model building module 320 is used to build an error model of the installation deviation angle between the INS and the LiDAR according to the attitude information of the INS and the attitude information of the LiDAR;
  • a system state equation building module 330 configured to build a system state equation, the state variables of the system state equation include the installation deviation angle between the INS and the LiDAR;
  • the system observation equation building module 340 is configured to construct a system observation equation, and the observation amount of the system observation equation includes an observation matrix corresponding to the installation deviation angle between the INS and the LiDAR determined according to the error model.
  • the filtering and solving module 350 is used for filtering and solving the system state equation and the system observation equation, and takes the estimation result of the installation deviation angle between the INS and the LiDAR obtained when the filtering and solving result converges as the calibration result.
  • the error model building module 320 can specifically be used to: construct the observation amount of the error model according to the attitude information output by the INS and the LiDAR; construct the transfer relation equation from the attitude matrix corresponding to the INS to the attitude matrix corresponding to the LiDAR ;Construct the error model with the observation quantity and transfer relationship equation of the simultaneous error model; wherein, the attitude information includes one or more of roll angle, pitch angle and heading angle, and the transfer relationship equation includes the attitude error of the INS and the relationship between the IMU and LiDAR installation deviation angle.
  • the filtering and solving module 350 may be specifically configured to: perform Kalman filtering on the system state equation and the system observation equation to obtain the optimal estimation result of the installation deviation angle between the IMU and the LiDAR;
  • the optimal estimation result compensates the positioning result of the LiDAR and corrects the inertial navigation solution result of the INS; when the filtering solution result converges, the optimal estimation result of the installation deviation angle between the IMU and the LiDAR is saved as the calibration result into the configuration file.
  • Embodiments of the present application further provide a vehicle, which may include the combined positioning system provided by the foregoing embodiments, and the user executes the method for calibrating the sensor installation deviation angle provided by the foregoing embodiments.
  • Embodiments of the present application further provide a computer-readable storage medium, where instructions are stored in the computer-readable storage medium, when the computer-readable storage medium runs on a computer, the computer executes the methods of the above aspects.
  • Embodiments of the present application also provide a computer program product containing instructions, which, when run on a computer, cause the computer to execute the methods of the above aspects.

Landscapes

  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Automation & Control Theory (AREA)
  • Electromagnetism (AREA)
  • Manufacturing & Machinery (AREA)
  • Navigation (AREA)
  • Position Fixing By Use Of Radio Waves (AREA)

Abstract

一种传感器安装偏差角的标定方法、组合定位系统和车辆。其中,组合定位系统包括多个传感器,多个传感器在车辆的安装位置不同;该标定方法包括:根据各个传感器输出的姿态信息,构建各个传感器之间的安装偏差角的误差模型;构建系统状态方程,系统状态方程的状态变量包括安装偏差角;构建系统观测方程,系统观测方程的观测量包括根据误差模型确定的安装偏差角对应的观测矩阵;对系统状态方程和系统观测方程进行滤波解算,将滤波解算结果收敛时得到的安装偏差角的估计结果作为标定结果,该系统能够对传感器之间的安装偏差角进行在线标定,当传感器之间的安装位置关系发生变化时,能够及时更新标定结果,精度高、操作简单。

Description

传感器安装偏差角的标定方法、组合定位系统和车辆
本申请要求于2020年07月04日提交到国家知识产权局、申请号为202010636730.6、发明名称为“传感器安装偏差角的标定方法、组合定位系统和车辆”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
本申请涉及自动驾驶技术领域,尤其涉及一种传感器安装偏差角的标定方法、组合定位系统和车辆。
背景技术
自动驾驶车辆的定位系统用于确定车辆的位置,为车辆的路径规划和导航提供最基础的信息,定位系统的性能将直接决定自动驾驶车辆的稳定性和可靠性。目前定位系统一般为包含多种传感器的组合定位系统,例如:惯性导航系统(inertial navigation system,INS)使用的惯性测量单元(inertial measurement unit,IMU)、全球卫星导航系统(global navigation satellite system,GNSS)接收机、激光雷达(light detection and ranging,LiDAR)、里程计、视觉传感器等。
组合定位系统中的各个传感器通常安装到车辆的不同位置,因此在使用之前需要对各个传感器之间的安装位置关系标定,例如标定各个传感器之间的安装偏差角,以便于准确地对各个传感器的解算结果或定位结果进行滤波计算,提高定位精度。
目前,各个传感器之间的安装偏差角的标定主要采用离线标定方法,依赖专用的标定工具而使用一些光学手段或者间接标定的方式实现,操作较为繁琐,标定时间较长。当车辆由于长时间振动、颠簸等原因导致各个传感器之间的安装位置关系发生变化时,还需要重新标定,否则会影响组合定位系统的定位精度。
发明内容
本申请实施例提供了一种传感器安装偏差角的标定方法、组合定位系统和车辆,能够实现对传感器安装偏差角的在线标定,当各个传感器之间的安装关系发生变化时,标定结果能够在线更新,从而始终保证组合定位系统的定位精度。
第一方面,本申请实施例提供了一种传感器安装偏差角的标定方法,该方法应用于组合定位系统,组合定位系统包括多个传感器,多个传感器的安装位置不同;该方法包括:根据各个传感器输出的姿态信息,构建各个传感器之间的安装偏差角的误差模型;构建系统状态方程,系统状态方程的状态变量包括安装偏差角;构建系统观测方程,系统观测方程的观测量包括根据误差模型确定的安装偏差角对应的观测矩阵;对系统状态方程和系统观测方程进行滤波解算,将滤波解算结果收敛时得到的安装偏差角的估计结果作为标定结果。
本申请实施例提供的方法,将各个传感器之间的安装偏差角整合到组合定位系统的状态变量中,因此能够实现在车辆行驶过程中对各个传感器之间的安装偏差角进行标定(即在线标定),标定过程不需要考虑传感器的物理结构,标定精度高。并且,本申请实施例的方法不需要借助用的标定工具,仅需要车辆产生机动即可完成标定,操作简单、快速。另外,当各 个传感器之间的安装位置关系发生变化时,本申请实施例的方法能够在线更新标定结果,从而始终保证组合定位系统的定位精度。
在一种可选择的实现方式中,多个传感器包括惯性导航系统INS的惯性测量单元IMU和全球导航卫星系统GNSS接收机,GNSS接收机包括GNSS天线,IMU和GNSS天线的安装位置不同;根据各个传感器输出的姿态信息,构建各个传感器之间的安装偏差角的误差模型,包括:构造误差模型的观测量,观测量为IMU和GNSS天线之间的姿态偏差;构建从INS对应的姿态矩阵到GNSS对应的姿态矩阵的转移关系方程;联立误差模型的观测量和转移关系方程构建误差模型,所述误差模型为所述IMU和GNSS天线之间的安装偏差角的误差模型;其中,姿态信息包括滚转角、俯仰角和航向角中的一个或者多个,转移关系方程包含INS的姿态误差和IMU与GNSS天线之间的安装偏差角。
在一种可选择的实现方式中,误差模型的观测量包括:
Figure PCTCN2021083032-appb-000001
其中,γ m、θ m
Figure PCTCN2021083032-appb-000002
分别为INS输出的滚转角、俯仰角和航向角;γ g、θ g
Figure PCTCN2021083032-appb-000003
分别为GNSS输出的滚转角、俯仰角和航向角。
在一种可选择的实现方式中,转移关系方程包括:
Figure PCTCN2021083032-appb-000004
其中,
Figure PCTCN2021083032-appb-000005
为INS对应的姿态矩阵,
Figure PCTCN2021083032-appb-000006
为GNSS对应的姿态矩阵,
Figure PCTCN2021083032-appb-000007
为INS的姿态误差,λ 0为IMU与GNSS天线之间的安装偏差角,λ 0×表示λ 0的叉乘。
在一种可选择的实现方式中,误差模型包括:
Figure PCTCN2021083032-appb-000008
其中,Z 1、Z 2、Z 3为误差模型的观测量;
Figure PCTCN2021083032-appb-000009
分别为INS基于东-北-天坐标系的E轴姿态误差N轴姿态误差和U轴姿态误差;
Figure PCTCN2021083032-appb-000010
分别为IMU与GNSS天线之间基于右前上坐标系的X轴安装偏差角、Y轴安装偏差角和Z轴安装偏差角;其他参数为GNSS对应的姿态矩阵
Figure PCTCN2021083032-appb-000011
中的参数。
在一种可选择的实现方式中,系统状态方程包括:
X=[X 1,X 2]
Figure PCTCN2021083032-appb-000012
Figure PCTCN2021083032-appb-000013
其中,X 1和X 2为状态变量,X 1为惯导误差,X 2为IMU与GNSS天线之间的安装偏差角;
Figure PCTCN2021083032-appb-000014
为X 1的导数;F 1为组合定位系统的惯导误差模型的系统矩阵,G 1、W 1为惯导误差驱动噪声;
Figure PCTCN2021083032-appb-000015
为X 2的导数;F 2为IMU与GNSS天线之间的安装偏差角的误差模型的系统矩阵,G 2、W 2为IMU与GNSS天线之间的安装偏差角的误差驱动噪声。
在一种可选择的实现方式中,系统状态方程包括:
Z=HX+V=[H 1 H 2 H 3]
其中,Z为系统观测量,H为系统观测矩阵,X为系统状态方程,V为系统观测噪声矩阵,H 1为INS的姿态误差对应的观测矩阵,H 2为惯导误差模型中除了姿态误差以外的其他状态变量对应的观测矩阵,H 3为IMU与GNSS天线之间的安装偏差角对应的观测矩阵。
在一种可选择的实现方式中,对系统状态方程和系统观测方程进行滤波解算,将滤波解算结果收敛时得到的安装偏差角的最优估计结果作为标定结果,包括:对系统状态方程和系统观测方程进行卡尔曼滤波解算,得到IMU与GNSS天线之间的安装偏差角的最优估计结果;根据最优估计结果对GNSS的定位结果进行补偿,以及对INS的惯导解算结果进行修正;当滤波解算结果收敛时,将IMU与GNSS天线之间的安装偏差角的最优估计结果作为标定结果保存至配置文件中。
在一种可选择的实现方式中,多个传感器包括IMU和激光雷达LiDAR,IMU和LiDAR的安装位置不同;根据各个传感器输出的姿态信息,构建各个传感器之间的安装偏差角的误差模型,包括:构造误差模型的观测量,观测量为IMU和LiDAR之间的姿态偏差;构建从INS对应的姿态矩阵到LiDAR对应的姿态矩阵的转移关系方程;联立误差模型的观测量和转移关系方程构建误差模型,误差模型为IMU和LiDAR之间的安装偏差角的误差模型;其中,姿态信息包括滚转角、俯仰角和航向角中的一个或者多个,转移关系方程包含INS的姿态误差和IMU与LiDAR之间的安装偏差角。
在一种可选择的实现方式中,误差模型的观测量包括:
Figure PCTCN2021083032-appb-000016
其中,γ m、θ m
Figure PCTCN2021083032-appb-000017
分别为INS输出的滚转角、俯仰角和航向角;γ l、θ l
Figure PCTCN2021083032-appb-000018
分别为LiDAR输出的滚转角、俯仰角和航向角。
在一种可选择的实现方式中,转移关系方程包括:
Figure PCTCN2021083032-appb-000019
其中,
Figure PCTCN2021083032-appb-000020
为INS对应的姿态矩阵,
Figure PCTCN2021083032-appb-000021
为LiDAR对应的姿态矩阵,
Figure PCTCN2021083032-appb-000022
为INS的姿态误差,λ 1为IMU与LiDAR天线之间的安装偏差角,λ 1×表示λ 1的叉乘。
在一种可选择的实现方式中,误差模型包括:
Figure PCTCN2021083032-appb-000023
其中,Z 1、Z 2、Z 3为误差模型的观测量;
Figure PCTCN2021083032-appb-000024
分别为INS基于东-北-天坐标系的E轴姿态误差N轴姿态误差和U轴姿态误差;
Figure PCTCN2021083032-appb-000025
分别为IMU与LiDAR天线之间基于右前上坐标系的X轴安装偏差角、Y轴安装偏差角和Z轴安装偏差角;其他参数为LiDAR对应的姿态矩阵
Figure PCTCN2021083032-appb-000026
中的参数。
在一种可选择的实现方式中,系统状态方程包括:
X=[X 1,X 3]
Figure PCTCN2021083032-appb-000027
Figure PCTCN2021083032-appb-000028
其中,X 1和X 3为状态变量,X 1为惯导误差,X 3为IMU与LiDAR之间的安装偏差角;
Figure PCTCN2021083032-appb-000029
为X 1的导数;F 1为组合定位系统的惯导误差模型的系统矩阵,G 1、W 1为惯导误差驱动噪声;
Figure PCTCN2021083032-appb-000030
为X 3的导数;F 3为IMU与LiDAR之间的安装偏差角的误差模型的系统矩阵,G 3、W 3为IMU与LiDAR之间的安装偏差角的误差驱动噪声。
在一种可选择的实现方式中,系统状态方程包括:
Z=HX+V=[H 1 H 2 H 3]
其中,Z为系统观测量,H为系统观测矩阵,X为系统状态方程,V为系统观测噪声矩阵,H 1为INS的姿态误差对应的观测矩阵,H 2为惯导误差模型中除了姿态误差以外的其他状态变量对应的观测矩阵,H 3为IMU与LiDAR之间的安装偏差角对应的观测矩阵。
在一种可选择的实现方式中,对系统状态方程和系统观测方程进行滤波解算,将滤波解算结果收敛时得到的安装偏差角的最优估计结果作为标定结果,包括:对系统状态方程和系统观测方程进行卡尔曼滤波解算,得到IMU与LiDAR之间的安装偏差角的最优估计结果;根据最优估计结果对LiDAR的定位结果进行补偿,以及对INS的惯导解算结果进行修正;当滤波解算结果收敛时,将IMU与LiDAR之间的安装偏差角的最优估计结果作为标定结果保存至配置文件中。
第二方面,本申请实施例提供了一种组合定位系统,该系统包括处理器、存储器和多个传感器,多个传感器的安装位置不同;存储器包括有程序指令,程序指令被处理器运行时,使得组合定位系统用于执行如下步骤:根据各个传感器输出的姿态信息,构建各个传感器之间的安装偏差角的误差模型;构建系统状态方程,系统状态方程的状态变量包括安装偏差角;构建系统观测方程,系统观测方程的观测量包括根据误差模型确定的安装偏差角对应的观测矩阵;对系统状态方程和系统观测方程进行滤波解算,将滤波解算结果收敛时得到的安装偏差角的估计结果作为标定结果。
本申请实施例提供的组合定位系统,将各个传感器之间的安装偏差角整合到组合定位系统的状态变量中,因此能够实现在车辆行驶过程中对各个传感器之间的安装偏差角进行标定(即在线标定),标定过程不需要考虑传感器的物理结构,标定精度高。并且,本申请实施例的组合定位系统不需要借助用的标定工具,仅需要车辆产生机动即可完成标定,操作简单、快速。另外,当各个传感器之间的安装位置关系发生变化时,本申请实施例的方法能够在线更新标定结果,从而始终保证组合定位系统的定位精度。
在一种可选择的实现方式中,多个传感器包括惯性导航系统INS的惯性测量单元IMU和全球导航卫星系统GNSS接收机,GNSS接收机包括GNSS天线,IMU和GNSS天线的安装位置不同;程序指令被处理器运行时,使得组合定位系统用于执行如下步骤,以实现根据各个传感器输出的姿态信息,构建各个传感器之间的安装偏差角的误差模型:构造误差模型的观测量,观测量为IMU和GNSS天线之间的姿态偏差;构建从INS对应的姿态矩阵到GNSS对应的姿态矩阵的转移关系方程;联立误差模型的观测量和转移关系方程构建误差模型,误差模型为IMU和GNSS天线之间的安装偏差角的误差模型;其中,姿态信息包括滚转角、俯仰角和航向角中的一个或者多个,转移关系方程包含INS的姿态误差和IMU与GNSS天线之间的安装偏差角。
在一种可选择的实现方式中,程序指令被处理器运行时,使得组合定位系统用于执行如下步骤,以实现对系统状态方程和系统观测方程进行滤波解算,将滤波解算结果收敛时得到 的安装偏差角的最优估计结果作为标定结果:对系统状态方程和系统观测方程进行卡尔曼滤波解算,得到IMU与GNSS天线之间的安装偏差角的最优估计结果;根据最优估计结果对GNSS的定位结果进行补偿,以及对INS的惯导解算结果进行修正;当滤波解算结果收敛时,将IMU与GNSS天线之间的安装偏差角的最优估计结果作为标定结果保存至配置文件中。
在一种可选择的实现方式中,多个传感器包括IMU和激光雷达LiDAR,IMU和LiDAR的安装位置不同;程序指令被处理器运行时,使得组合定位系统用于执行如下步骤,以实现根据各个传感器输出的姿态信息,构建各个传感器之间的安装偏差角的误差模型:构造误差模型的观测量,观测量为IMU和LiDAR之间的姿态偏差;构建从INS对应的姿态矩阵到LiDAR对应的姿态矩阵的转移关系方程;联立误差模型的观测量和转移关系方程构建误差模型,误差模型为IMU和LiDAR之间的安装偏差角的误差模型;其中,姿态信息包括滚转角、俯仰角和航向角中的一个或者多个,转移关系方程包含INS的姿态误差和IMU与LiDAR之间的安装偏差角。
在一种可选择的实现方式中,程序指令被处理器运行时,使得组合定位系统用于执行如下步骤,以实现对系统状态方程和系统观测方程进行滤波解算,将滤波解算结果收敛时得到的安装偏差角的最优估计结果作为标定结果:对系统状态方程和系统观测方程进行卡尔曼滤波解算,得到IMU与LiDAR之间的安装偏差角的最优估计结果;根据最优估计结果对LiDAR的定位结果进行补偿,以及对INS的惯导解算结果进行修正;当滤波解算结果收敛时,将IMU与LiDAR之间的安装偏差角的最优估计结果作为标定结果保存至配置文件中。
进一步地,本申请实施例提供的组合定位系统,还可以执行前述第一方面的其他实现方式。
第三方面,本申请实施例提供了一种车辆,该车辆包括组合定位系统,组合定位系统包括处理器、存储器和多个传感器,不同传感器安装在车辆的不同位置;存储器包括有程序指令,程序指令被处理器运行时,使得组合定位系统用于执行前述第一方面及其任意实现方式中的方法。
第四方面,本申请实施例还提供一种计算机可读存储介质,计算机可读存储介质中存储有指令,当其在计算机上运行时,使得计算机执行上述各方面的方法。
第五方面,本申请实施例还提供了一种包含指令的计算机程序产品,当其在计算机上运行时,使得计算机执行上述各方面的方法。
第六方面,本申请实施例还提供了一种芯片系统,该芯片系统包括处理器,用于支持上述装置或系统实现上述方面中所涉及的功能,例如,生成或处理上述方法中所涉及的信息。
附图说明
图1是采用组合定位算法的组合定位系统的数据逻辑框图;
图2是组合定位系统中的各个传感器在车辆上的位置示意图;
图3是本申请实施例示出的安装偏差角的示意图;
图4是本申请实施例提供的组合定位系统的硬件结构图;
图5是本申请实施例提供的扩充后的组合定位系统的数据逻辑框图;
图6是本申请实施例提供的传感器安装偏差角的标定方法应用于IMU与GNSS天线之间的安装偏差角的标定的流程图;
图7是本申请实施例提供的传感器安装偏差角的标定方法应用于IMU与LiDAR安装偏 差角的标定的流程图;
图8是本申请实施例提供的组合定位系统的软件模块的示意图;
图9是本申请实施例提供的组合定位系统的软件模块应用于标定IMU和GNSS天线之间的安装偏差角时的数据逻辑框图;
图10是本申请实施例提供的组合定位系统的软件模块应用于标定IMU和LiDAR之间的安装偏差角时的数据逻辑框图。
具体实施方式
自动驾驶车辆(autonomous vehicles或self-driving automobile),也称无人驾驶车辆,电脑驾驶车辆或者轮式移动机器人。自动驾驶车辆能够以雷达、全球导航卫星系统(global navigation satellite system,GNSS)及机器视觉等技术感测其周围环境,并根据感测数据确定自身位置、规划导航路线、更新地图信息、躲避障碍等,最终实现在没有任何人类主动操作或者少有人类主动操作的情况下自动地驾驶车辆。
一般来说,自动驾驶车辆需要解决驾驶的三个核心问题:在哪里?(车辆定位);去哪里?(确定目的地);怎么去?(路径规划)。其中,自动驾驶车辆的定位系统主要用于解决“在哪里?”的问题。因此,作为自动驾驶车辆的关键模块之一,定位系统的性能将直接决定自动驾驶车辆的稳定性和可靠性。
目前,定位系统为了解决“在哪里?”的问题通常会包含有用于车辆定位的传感器,例如:惯性测量单元IMU、GNSS接收机、激光雷达LiDAR、里程计、视觉传感器等。其中,IMU例如可以包括三轴陀螺仪和加速度计,可用于测量自动驾驶车辆的角速度和加速度;GNSS接收机可用于测量GNSS的卫星信号,LiDAR可用于测量车辆与其他物体的距离。由于不同的传感器用于测量不同的数据,各个传感器应用在定位系统中既有优势也有局限性,单独使用任何一种传感器都难以达到定位精度高、抗干扰能力强、数据更新快的要求,因此目前的定位系统通常会将上述多个传感器的数据相结合,根据组合定位算法解算出车辆的位姿,达到取长补短、优势互补的效果,从而提高定位系统的稳定性和可靠性。
目前,采用组合定位算法的组合定位系统的数据逻辑框图如图1所示,其中,组合定位系统以INS 100的解算结果,例如根据IMU 110测量的角速度和加速度输出的位置、速度和姿态信息(以下称惯导位姿),GNSS 120的定位结果,LiDAR 130的定位结果等作为组合滤波器140的输入,利用滤波结果对INS 100的解算结果进行误差修正,以确定最终的定位结果,例如位姿及其协方差。在图1以及其他常见的定位系统中,组合滤波器140的系统变量一般为惯导相关误差(简称惯导误差),例如:位置误差、速度误差、姿态误差、加速度计零偏差和陀螺仪零偏差等。
本申请实施例中的组合滤波器140例如可以是卡尔曼滤波器(kalman filter,KF),卡尔曼滤波器是一种递归滤波器(自回归滤波器),它能够从一系列的不完全及包含噪声的测量量中,估计动态系统的状态。卡尔曼滤波会根据各测量量在不同时间下的值,考虑各时间下的联合分布,再产生对未知变量的估计,因此会比只以单一测量量为基础的估计方式要准确。
图2是组合定位系统中的各个传感器在车辆上的位置示意图。如图2所示,该组合定位系统的传感器包括IMU 110、LiDAR 130和GNSS天线,GNSS天线可以包括一个GNSS主天线121和一个GNSS从天线122,简称GNSS双天线。组合定位系统中的各个传感器通常会被安装到车辆的不同位置,例如,IMU 110被安装到车辆的中部,GNSS双天线被安装到 车辆的顶部,并且GNSS主天线121和GNSS从天线122前后分布设置,以便于接收卫星的GNSS信号,LiDAR 130被安装到车辆的头部,以便于测量车辆行驶前方的物体。
一般来说,各个传感器会基于自身的坐标系描述其解算结果或定位结果,由于各个传感器的安装位置不同,因此各个传感器的坐标系也不同。示例地,如图2所示,组合定位系统中包括:环境坐标系(E,N,U),LiDAR坐标系(X l,Y l,Z l),IMU坐标系(X b,Y b,Z b),GNSS坐标系(X g,Y g,Z g)和后轮轴地面投影坐标系(X m,Y m,Z m)。可以理解的是,由于各个传感器的坐标系不同,因此将各个传感器的解算结果或定位结果进行滤波计算时,需要依赖各个传感器之间的安装位置关系以归化坐标系。各个传感器之间的安装位置关系可以通过标定获得,因此,标定的准确性就成为了组合定位系统定位精度的重要因素。
在一个实施例中,环境坐标系(E,N,U)例如可以是站心坐标系,也称东-北-天坐标系ENU。站心坐标系以目标(例如GNSS天线)所在位置为坐标原点,包括指向东方向的E轴、指向北方向的N轴和指向天空的U轴。
在一个实施例中,LiDAR坐标系(X l,Y l,Z l),IMU坐标系(X b,Y b,Z b),GNSS坐标系(X g,Y g,Z g)和后轮轴地面投影坐标系(X m,Y m,Z m)均可以使用右前上坐标系。在组合定位系统中,右前上坐标系以载具(例如车辆)的前进方向为Y轴方向,以载具的右侧为X轴方向,以垂直于X轴和Y轴的朝上方向为Z轴方向,各个坐标系的原点可以是其对应的传感器的中心。
本申请实施例中,各个传感器之间的安装位置关系可以包括臂参数以及安装偏差角。其中,杆臂参数主要描述各个传感器之间的安装位置距离,一般可以通过测量设备直接测量得到;安装偏差角主要描述各个传感器之间的安装角度偏差,考虑到车辆上传感器安装位置关系以及传感器的自身结构特性,各个传感器之间的安装偏差角一般很难通过测量设备直接精确测量。一般来说,安装偏差角可以包括IMU和GNSS天线之间的安装偏差角,以及IMU与LiDAR之间的安装偏差角。
图3是本申请实施例示出的安装偏差角的示意图,如图3所示,IMU 110和GNSS天线之间的航向安装偏差角是指GNSS双天线之间的连线(从GNSS主天线121到GNSS从天线122)与IMU坐标系前向轴(yb)的夹角λgb;IMU 110与LiDAR 130之间的航向安装偏差角指LiDAR坐标系前向轴(yl)与IMU坐标系前向轴(yb)的夹角λlb。
目前,IMU和GNSS天线之间的安装偏差角的标定主要采用离线标定方法,并且大多采用光学手段实现。以航向安装偏差角为例,可以通过光学手段分别对IMU和GNSS天线的航向进行标定,然后计算二者航向的偏差值即为IMU和GNSS天线之间的航向安装偏差角。然而,由于GNSS天线的相位中心位于天线内部,致使难以准确测量双天线的轴线,因此该方法的标定精度较低;并且,该方法需要依赖专用的标定工具,例如激光扫描仪、经纬仪等,操作较为繁琐,标定时间较长;另外,当车辆由于长时间振动、颠簸等原因导致IMU与GNSS安装位置关系发生变化时,需要重新标定,否则会影响组合定位系统的定位精度。
目前,IMU与LiDAR之间的安装偏差角的标定方法同样主要为离线标定方法,并且大多采用间接标定法,以航向安装偏差角为例,可以借助测量公共点或者辅助测量坐标系,建立IMU与LiDAR之间的联系,从而进一步确定IMU与LiDAR之间的航向安装偏差角。该方法同样需要依赖专用的标定工具,操作较为繁琐,标定时间较长,并且,当IMU与LiDAR安装位置关系发生变化时,需要重新标定,否则会影响组合定位系统的定位精度。
为了解决现有技术中存在的技术问题,本申请实施例提供了一种传感器安装偏差角的标 定方法。
本申请实施例中的组合定位系统,可以是包括两个或者两个以上的定位传感器的定位系统,并且该定位系统使用上述两个或者两个以上的传感器的解算结果或者定位结果综合确定最终的定位结果。对于组合定位系统具体包含哪些定位传感器,本申请实施例不做具体限定,例如:在一些实施例中,组合定位系统可以包括IMU和GNSS;在另一些实施例中,组合定位系统包括IMU和LiDAR;在另一些实施例中,组合定位系统可以包括IMU、LiDAR和GNSS;在另一些实施例中,组合定位系统除包含上述定位传感器之外,还可以包括其他传感器。
图4是本申请实施例提供的组合定位系统的硬件结构图。在一个实施例中,该组合定位系统可以包括:IMU 110、GNSS 120、LiDAR 130、里程计140、视觉传感器150等传感器,处理器160、存储器170和通信模块180等。其中,IMU 110、GNSS 120、LiDAR 130等传感器用于测量和接收各类测量数据,处理器160用于对根据相应的组合定位算法对测量数据进行解算,例如:惯性导航解算、卡尔曼滤波计算等,以得到定位结果。
处理器160例如可以包括一个或者多个处理单元,例如中央处理器(central processing unit,CPU)、微控制器(microcontroller,MCU)、图像信号处理器(image signal processor,ISP)、神经网络处理器(neural-network processing unit,NPU)等。其中,不同的处理单元可以是独立的器件,也可以集成在一个或多个处理器中,例如集成在系统芯片(system on a chip,SoC)中。处理器160中还可以设置存储器,用于存储指令和数据。在一些实施例中,处理器中的存储器为高速缓冲存储器。该存储器可以保存处理器刚用过或循环使用的指令或数据。
存储器170可以包括一个或者多个存储单元,例如可以包括易失性存储器(volatile memory),如:动态随机存取存储器(dynamic random access memory,DRAM)、静态随机存取存储器(static random access memory,SRAM)等;还可以包括非易失性存储器(non-volatile memory,NVM),如:只读存储器(read-only memory,ROM)、闪存(flash memory)等。存储器170可以用于存储处理器可执行的程序代码和指令,例如:用于处理传感器数据的程序代码和指令、用于惯性导航解算的程序和指令、用于实现卡尔曼滤波算法的程序和指令等。存储器170还可以存储有自动驾驶计算平台的车机操作系统,例如:鸿蒙操作系统Harmony OS、UQX操作系统等。
通信模块180用于使车辆实现与外界的2G、3G、4G和5G多种网络模式的通信,例如实现车辆的车联万物(vehicle-to-everything,V2X)等,通信模块例如可以包括基带芯片、功率放大器和天线等。在一些实施例中,通信模块180的部分或者全部器件可以与一个或者多个处理器集成在一起,例如基带芯片可以与CPU、NPU、ISP等集成在SoC中。
另外,组合定位系统还可以包括通信接口,例如控制器局域网(controller area network,CAN)接口,使得处理器可以与车辆的其他芯片和设备,例如电子控制单元(electrical control unit,ECU)进行通信。
可以理解的是,本申请实施例示意的组合定位系统的硬件结构并不构成对组合定位系统的具体限定。在本申请另一些实施例中,组合定位系统的硬件结构可以包括比图示更多或更少的部件,或者组合某些部件,或者拆分某些部件,或者不同的部件布置。图示的部件可以以硬件,软件或软件和硬件的组合实现。
本申请实施例中的组合定位系统可以安装到各类应用自动驾驶技术或导航定位的载具之上,包括但不限于各种交通工具:例如车辆(汽车)、轮船、火车、地铁、飞机等,以及各种机器人,例如:服务机器人、运输机器人、自主导引机器人(automated guided vehicle,AGV)、 无人地面车(unmanned ground vehicle,UGV)等,以及各种工程机械,例如:隧道掘进机等。
下面以组合定位系统包括IMU、LiDAR和GNSS等传感器为例,对本申请实施例提供的传感器安装偏差角的标定方法的技术方案进行示例地阐述说明。
本申请实施例对目前组合定位系统的状态变量进行了扩充,扩充后的组合定位系统的数据逻辑框图如图5所示,其中,该组合定位系统以根据IMU 110测量的角速度和加速度确定的惯导位姿、GNSS 120的定位结果、LiDAR 130的定位结果等作为组合定位滤波器的输入,并且在原有的惯导相关的误差,例如:位置误差、速度误差、姿态误差、加速度计零偏差和陀螺仪零偏差等基础之上,将IMU和GNSS天线之间的安装偏差角,以及IMU与LiDAR之间的安装偏差角也扩充到组合定位系统的系统变量中,从而通过组合滤波器对IMU和GNSS天线之间的安装偏差角,以及IMU与LiDAR之间的安装偏差角进行在线标定;当组合定位滤波器的状态变量完成收敛时,根据在线标定结果中的IMU和GNSS天线之间的安装偏差角,以及IMU与LiDAR之间的安装偏差角对GNSS和LiDAR的定位结果进行补偿,以提高定位精度。
在一些实施例中,本申请提供的传感器安装偏差角的标定方法具体可以由两部分组成,第一部分包括IMU与GNSS双天线安装偏差角的标定,第二部分包括IMU与LiDAR之间的安装偏差角的标定方。下面分别结合更多附图分别对本方法的上述两部分进行具体地解释说明。
图6是本申请实施例提供的传感器安装偏差角的标定方法应用于IMU与GNSS天线之间的安装偏差角的标定的流程图。如图6所示,该方法可以包括以下步骤:
步骤S101,获取IMU的测量结果。
其中,IMU的测量结果例如可以包括IMU测量得到的角速度和加速度信息。
具体实现中,在IMU和GNSS等传感器固定安装在载具上之后,可以使载具产生机动,例如加速、减速、匀速行驶、左转和右转等,在载具机动期间,组合定位系统可以实时并且连续地获取IMU测量得到的角速度和加速度信息。
步骤S102,对IMU的测量结果进行解算,以得到INS的姿态信息。
INS的姿态信息可以通过姿态矩阵来表达。另外,步骤S102对测量结果进行解算之后还可以得到INS的位置和速度等信息。上述位置、速度和姿态信息也可以概括称为惯导位姿。
具体实现中,在载具产生机动的过程中,组合定位系统可以实时并且连续地对IMU的测量结果数据进行解算,从而连续地获取车辆当前的位置、速度和姿态信息等。
在一个实施例中,在解算的初始阶段,可以根据GNSS定位结果确定车辆的初始位置和初始速度,作为IMU的初始数据;然后,组合定位系统可以根据初始数据和当前IMU的测量结果数据解算得到车辆最新的位置、速度和姿态信息。在解算的非初始阶段,组合定位系统可以根据上一次解算得到的车辆的位置、速度和姿态信息和当前IMU的测量结果数据得到车辆最新的位置、速度和姿态信息。
步骤S103,根据INS的姿态信息和GNSS姿态信息构建INS与GNSS天线之间的安装偏差角的误差模型。
其中,该误差模型以通过组合定位滤波器的状态变量进行描述。
在一个实施例中,为了建立IMU与GNSS之间的安装偏差角的误差模型,本申请实施例可以构建INS对应的姿态矩阵
Figure PCTCN2021083032-appb-000031
以及,GNSS对应的姿态矩阵
Figure PCTCN2021083032-appb-000032
姿态矩阵
Figure PCTCN2021083032-appb-000033
例如可以是一个3×3大小的矩阵,其表示INS输出的位姿从后轮轴地面投 影坐标系(Xm,Ym,Zm)经过平移和旋转等映射到环境坐标系(E,N,U)。一般来说,由于IMU的安装位置可能存在误差,因此,将此IMU输出的位姿从后轮轴地面投影坐标系(Xm,Ym,Zm)进行映射之后,可能会映射到有误差的环境坐标系(Xn′,Yn′,Zn′),因此,本申请实施例将上述IMU对应的姿态矩阵记作
Figure PCTCN2021083032-appb-000034
姿态矩阵
Figure PCTCN2021083032-appb-000035
例如可以是一个3×3大小的矩阵,其表示GNSS输出的位姿从GNSS坐标系(Xg,Yg,Zg)经过平移和旋转等映射到环境坐标系(E,N,U)。
示例地,IMU对应的姿态矩阵
Figure PCTCN2021083032-appb-000036
和GNSS对应的姿态矩阵
Figure PCTCN2021083032-appb-000037
可以表示为以下公式(1):
Figure PCTCN2021083032-appb-000038
其中,姿态矩阵
Figure PCTCN2021083032-appb-000039
中的参数
Figure PCTCN2021083032-appb-000040
等用来表示INS输出的位姿从后轮轴地面投影坐标系(Xm,Ym,Zm)映射到环境坐标系(E,N,U)需要的平移和旋转等参数;姿态矩阵
Figure PCTCN2021083032-appb-000041
中的参数
Figure PCTCN2021083032-appb-000042
等用来表示GNSS输出的位姿从GNSS坐标系(Xg,Yg,Zg)映射到环境坐标系(E,N,U)需要的平移和旋转等参数。
另外,为便于描述,本申请实施例这里将INS输出的滚转角表示为γ m、俯仰角表示为θ m、航向角表示为
Figure PCTCN2021083032-appb-000043
以及,将GNSS定位输出的滚转角表示为γ g、俯仰角表示为θ g、航向角表示为
Figure PCTCN2021083032-appb-000044
那么,该误差模型的观测量可以表示为以下公式(2)
Figure PCTCN2021083032-appb-000045
其中,δγ 0表示IMU与GNSS输出的滚转角的偏差、δθ 0表示IMU与GNSS输出的俯仰角的偏差、
Figure PCTCN2021083032-appb-000046
表示IMU与GNSS输出的航向角的偏差,为便于概括描述,本申请实施例将上述偏差称作IMU和GNSS天线之间的姿态偏差。
需要补充说明的是,在车辆自动驾驶的定位和导航过程中,航向角对定位结果的准确性影响比较明显,因此,本申请实施例在后续输出标定结果时,也可以仅考虑输出IMU与GNSS双天线之间的航向安装偏差角。
进一步地,本申请实施例可以对IMU对应的姿态矩阵
Figure PCTCN2021083032-appb-000047
进行以下变换,构造从INS对应的姿态矩阵到GNSS对应的姿态矩阵的转移关系方程,以建立其与GNSS对应的姿态矩阵
Figure PCTCN2021083032-appb-000048
之间的联系。
其中,姿态矩阵
Figure PCTCN2021083032-appb-000049
可以分解成两个变换矩阵,即
Figure PCTCN2021083032-appb-000050
Figure PCTCN2021083032-appb-000051
并且
Figure PCTCN2021083032-appb-000052
其中,
Figure PCTCN2021083032-appb-000053
例如可以是一个3×3大小的矩阵,其表示将位姿从环境坐标系(E,N,U)经过平移和旋转等映射到有误差的环境坐标系(Xn′,Yn′,Zn′);
Figure PCTCN2021083032-appb-000054
例如可以是一个3×3大小的矩阵,其表示将位姿从后轮轴地面投影坐标系(Xm,Ym,Zm)经过平移和旋转等映射到环境坐标系(E,N,U)。
Figure PCTCN2021083032-appb-000055
可以进一步分解成两个变换矩阵,即
Figure PCTCN2021083032-appb-000056
Figure PCTCN2021083032-appb-000057
并且
Figure PCTCN2021083032-appb-000058
其中,
Figure PCTCN2021083032-appb-000059
例如可以是一个3×3大小的矩阵,其表示将位姿从GNSS坐标系(Xg,Yg,Zg)经过平移和旋转等映射到环境坐标系(E,N,U);
Figure PCTCN2021083032-appb-000060
例如可以是一个3×3大小的矩阵,其表示将位姿从后轮轴地面投影坐标系(Xm,Ym,Zm)经过平移和旋转等映射到GNSS坐标系(Xg,Yg,Zg)。
进而得到以下公式(3):
Figure PCTCN2021083032-appb-000061
进一步地,本申请实施例可以基于公式(3)向组合定位滤波器的状态变量中引入INS的姿态误差和IMU与GNSS天线之间的安装偏差角。
示例地,以东-北-天坐标系描述的INS的姿态误差
Figure PCTCN2021083032-appb-000062
可以为:
Figure PCTCN2021083032-appb-000063
其中,
Figure PCTCN2021083032-appb-000064
为INS的E轴姿态误差、
Figure PCTCN2021083032-appb-000065
为INS的N轴姿态误差、
Figure PCTCN2021083032-appb-000066
为INS的U轴姿态误差。则有:
Figure PCTCN2021083032-appb-000067
其中,
Figure PCTCN2021083032-appb-000068
表示
Figure PCTCN2021083032-appb-000069
的叉乘,I为单位矩阵:
Figure PCTCN2021083032-appb-000070
示例地,IMU与GNSS天线之间的安装偏差角λ可以为:
Figure PCTCN2021083032-appb-000071
其中,
Figure PCTCN2021083032-appb-000072
为X轴方向的安装偏差角、
Figure PCTCN2021083032-appb-000073
为Y轴方向的安装偏差角、
Figure PCTCN2021083032-appb-000074
为Z轴方向的安装偏差角。则有:
Figure PCTCN2021083032-appb-000075
其中,λ 0×表示λ 0的叉乘。
由此,根据公式(3)可进一步得到:
Figure PCTCN2021083032-appb-000076
上述公式(4)即可以作为从INS对应的姿态矩阵到GNSS对应的姿态矩阵的转移关系方程。
接下来,将公式(1)、公式(2)与公式(4)联立,可推导出IMU对应的姿态矩阵
Figure PCTCN2021083032-appb-000077
和GNSS对应的姿态矩阵
Figure PCTCN2021083032-appb-000078
中的各个元素间的关系如公式(5)所示:
Figure PCTCN2021083032-appb-000079
将公式(5)的左端按照泰勒级数展开,并且忽略公式中的二阶以上的高阶项,可以得到公式(6):
Figure PCTCN2021083032-appb-000080
该公式(6)可以称作IMU与GNSS双天线之间的安装偏差角的误差模型。其中,Z 1,Z 2,Z 3为该误差模型的观测量;
Figure PCTCN2021083032-appb-000081
为前述提到的INS的姿态误差,例如平台误差角;
Figure PCTCN2021083032-appb-000082
Figure PCTCN2021083032-appb-000083
为前述提到的IMU与GNSS天线之间的安装偏差角。
步骤S104,构建系统状态方程,系统状态方程的状态变量包括INS与GNSS天线之间的 安装偏差角。
一般来说,组合定位系统中原有的状态方程是基于惯导误差模型构建的方程,状态方程中的状态变量为惯导误差,例如:位置误差、速度误差、姿态误差、加速度计零偏差和陀螺仪零偏差等,而本申请实施例可以将IMU与GNSS天线之间的安装偏差角的误差模型引入到原有的状态方程中,使新的状态方程的状态变量既包含惯导误差相关的变量,也包含IMU与GNSS天线之间的安装偏差角的误差。
在一个实施例中,本申请实施例定义该组合定位系统新的系统状态方程包括:
X=[X 1,X 2]      (7)
其中,X 1和X 2为状态变量,X 1为惯导误差,X 2为IMU与GNSS天线之间的安装偏差角。
则有,X 1的系统状态方程为:
Figure PCTCN2021083032-appb-000084
其中,
Figure PCTCN2021083032-appb-000085
为惯导误差X 1的导数;F 1为惯导误差模型中可以推导的系统矩阵,例如可以是惯导误差模型的非线性转移矩阵;G 1,W 1为惯导误差驱动噪声,其中,G 1可以是惯导误差的噪声驱动矩阵,W 1可以是惯导误差的状态噪声矩阵,可选的,G 1和W 1均为高斯白噪声。
示例地,
Figure PCTCN2021083032-appb-000086
其中,
Figure PCTCN2021083032-appb-000087
分别为东、北、天方向平台失准角,δv E,δv N,δv U分别为INS在东、北、天方向的速度误差,δλ,δL,δH分别为INS的经度、纬度和高度误差,ε xyz为陀螺仪随机常值漂移在X,Y,Z轴的分量,
Figure PCTCN2021083032-appb-000088
为INS的加速度计零偏在X,Y,Z轴的分量。可以理解的是,上述X 1包含的参数项仅仅作为一个示例,不构成对X 1的限定,在一些其他的实现方式中,X 1可以包含更多的参数项或者更少的参数项,本申请实施例此处不做具体展开说明。
以及,X 2的系统状态方程为:
Figure PCTCN2021083032-appb-000089
其中,
Figure PCTCN2021083032-appb-000090
为IMU与GNSS天线之间的安装偏差角X 2的导数;F 2为IMU与GNSS天线之间的安装偏差角的误差模型中可以推导的系统矩阵,例如可以是该误差模型的非线性转移矩阵;G 2,W 2为IMU与GNSS天线之间的安装偏差角的误差驱动噪声,其中,G 2可以是IMU与GNSS天线之间的安装偏差角的误差的噪声驱动矩阵,W 2可以是IMU与GNSS天线之间的安装偏差角的误差的状态噪声矩阵,可选的,G 2和W 2均为高斯白噪声。
示例地,
Figure PCTCN2021083032-appb-000091
其中,
Figure PCTCN2021083032-appb-000092
为X轴方向的安装偏差角、
Figure PCTCN2021083032-appb-000093
为Y轴方向的安装偏差角、
Figure PCTCN2021083032-appb-000094
为Z轴方向的安装偏差角。可以理解的是,上述X 2包含的参数项仅仅作为一个示例,不构成对X 2的限定,在一些其他的实现方式中,X 2可以包含更多的参数项或者更少的参数项,本申请实施例此处不做具体展开说明。
另设,IMU与GNSS天线之间的安装偏差角的误差为
Figure PCTCN2021083032-appb-000095
Figure PCTCN2021083032-appb-000096
对应的系统状态方程为:
Figure PCTCN2021083032-appb-000097
其中,
Figure PCTCN2021083032-appb-000098
Figure PCTCN2021083032-appb-000099
的导数。
以上,公式(7)-(9)即构成了本申请实施例的组合定位系统的系统状态方程,可以看出,该系统状态方程的状态变量不仅包含了惯导误差,还包含了IMU与GNSS天线之间的安装偏差角,由此,在后续使用组合定位滤波器对系统状态方程的状态变量进行在线估计时,就能够同时对IMU与GNSS天线之间的安装偏差角进行在线标定。
步骤S105,构建系统观测方程,系统观测方程的观测量包括根据误差模型确定的INS与GNSS天线之间的安装偏差角对应的观测矩阵。
具体来说,该系观测方程的观测量为惯导误差对应的观测矩阵和IMU与GNSS天线之间的安装偏差角的误差对应的观测矩阵。
在一个实施例中,该系统观测方程可以为:
Z=HX+V=[H 1 H 2 H 3]
其中,Z为系统观测量,H为系统观测矩阵,X为本申请实施例在步骤S105中建立的新的系统状态方程,V为系统观测噪声矩阵。进一步地,H 1为姿态误差对应的观测矩阵,H 2为误差模型中除了姿态误差以外的其他状态变量对应的观测矩阵,H 3为IMU与GNSS天线之间的安装偏差角对应的观测矩阵。H 2可根据惯导误差模型得到,H 1和H 3如下所示:
Figure PCTCN2021083032-appb-000100
步骤S106,使用组合定位滤波器对系统状态方程和系统观测方程进行滤波解算,以得到IMU与GNSS天线之间的安装偏差角的最优估计结果。
具体实现中,组合定位滤波器可以是卡尔曼滤波器,即对系统状态方程和系统观测方程进行卡尔曼滤波,以得到IMU与GNSS之间的航向安装偏差角的最优估计结果。
步骤S107,根据IMU与GNSS天线之间的安装偏差角的最优估计结果对GNSS定位结果进行补偿。
这样,GNSS与IMU之间消除了由于安装偏差角而产生的实际安装位置关系与其坐标系位置关系的不一致性,从而消除了由安装偏差角造成的组合定位误差,提高组合定位系统的定位精度。
另外,本申请实施例还可以根据卡尔曼滤波得到的状态变量最优估计结果对惯导解算结果,例如位置、速度和姿态等进行校正,以修正惯导解算结果的累积误差,提高组合定位系统的定位精度。
步骤S108,当组合定位滤波器对的滤波解算结果收敛时,将IMU与GNSS天线之间的安装偏差角的最优估计结果作为标定结果保存至配置文件中。
在此之后,如果IMU与GNSS之间的安装位置关系没有发生变化,那么IMU与GNSS天线之间的安装偏差角就不会改变,那么卡尔曼滤波器对的滤波解算结果将一直处于收敛状态;如果车辆在行驶过程中发生振动或者车辆维修等导致IMU与GNSS之间的安装位置关系,那么当卡尔曼滤波器对的滤波解算结果经过短暂的不收敛状态之后也会再次进入到收敛状态,这时,IMU与GNSS天线之间的安装偏差角的最优估计结果会发生变化,组合定位系统可以 在卡尔曼滤波器重新收敛之后,将最新的标定结果保存至配置文件中。
由此可见,本申请实施例的方法将IMU与GNSS天线之间的安装偏差角整合到组合定位系统的状态变量中,实现了在车辆行驶过程中对组合定位系统的IMU与GNSS天线之间的安装偏差角进行标定(即在线标定),标定过程不需要考虑IMU和GNSS的物理结构,标定精度高。并且,本申请实施例的方法不需要借助用的标定工具,仅需要车辆产生机动即可完成标定,操作简单、快速。另外,当IMU与GNSS之间的安装位置关系发生变化时,本申请实施例的方法能够在线更新标定结果,从而始终保证组合定位系统的准确性。
图7是本申请实施例提供的传感器安装偏差角的标定方法应用于IMU与LiDAR安装偏差角的标定的流程图。如图7所示,该方法可以包括以下步骤:
步骤S201,获取IMU的测量结果。
其中,IMU的测量结果例如可以包括IMU测量得到的角速度和加速度信息。
步骤S202,对IMU的测量结果进行解算,以得到INS的姿态信息。
步骤S201和步骤S202的具体实现方式可以参照步骤S101和步骤S102,此处不再赘述。
步骤S203,根据INS的姿态信息和LiDAR姿态信息构建INS与LiDAR之间的安装偏差角的误差模型。
其中,该误差模型的参数可以通过组合定位滤波器的状态变量进行表述。
在一个实施例中,为了建立IMU与LiDAR之间的安装偏差角的误差模型,本申请实施例可以构建INS对应的姿态矩阵
Figure PCTCN2021083032-appb-000101
以及LiDAR对应的姿态矩阵
Figure PCTCN2021083032-appb-000102
姿态矩阵
Figure PCTCN2021083032-appb-000103
的具体实现方式可以参考本申请实施例对步骤S103和公式(1)描述的内容,此处不再赘述。
姿态矩阵
Figure PCTCN2021083032-appb-000104
例如可以是一个3×3大小的矩阵,其表示LiDAR输出的位姿从LiDAR坐标系(Xl,Yl,Zl)经过平移和旋转等映射到环境坐标系(E,N,U)。
示例地,IMU对应的姿态矩阵
Figure PCTCN2021083032-appb-000105
和LiDAR对应的姿态矩阵
Figure PCTCN2021083032-appb-000106
可以表示为以下公式(10):
Figure PCTCN2021083032-appb-000107
其中,姿态矩阵
Figure PCTCN2021083032-appb-000108
中的参数
Figure PCTCN2021083032-appb-000109
等用来表示INS输出的位姿从后轮轴地面投影坐标系(Xm,Ym,Zm)映射到环境坐标系(E,N,U)需要的平移和旋转等参数;姿态矩阵
Figure PCTCN2021083032-appb-000110
中的参数
Figure PCTCN2021083032-appb-000111
等用来表示LiDAR输出的位姿从LiDAR坐标系(Xl,Yl,Zl)映射到环境坐标系(E,N,U)需要的平移和旋转等参数。
另外,为便于描述,本申请实施例这里将IMU解算输出的滚转角表示为γ m、俯仰角表示为θ m、航向角表示为
Figure PCTCN2021083032-appb-000112
以及,将LiDAR定位输出的滚转角表示为γ l、俯仰角表示为θ l、航向角表示为
Figure PCTCN2021083032-appb-000113
那么,该组合定位系统的姿态观测量可以表示为以下公式(11)
Figure PCTCN2021083032-appb-000114
其中,δγ 1表示IMU与LiDAR输出的滚转角的偏差、δθ 1表示IMU与LiDAR输出的俯仰角的偏差、
Figure PCTCN2021083032-appb-000115
表示IMU与LiDAR输出的航向角的偏差,为便于概括描述,本申请实施例将上述偏差称作IMU和LiDAR之间的姿态偏差。
需要补充说明的是,在车辆自动驾驶的定位和导航过程中,航向角对定位结果的准确性影响比较明显,因此,本申请实施例在后续输出标定结果时,可以仅考虑输出IMU与LiDAR 之间的航向安装偏差角。
进一步地,本申请实施例可以对INS对应的姿态矩阵
Figure PCTCN2021083032-appb-000116
进行以下变换,构造从INS对应的姿态矩阵到LiDAR对应的姿态矩阵的转移关系方程,以建立其与GNSS对应的姿态矩阵
Figure PCTCN2021083032-appb-000117
之间的联系。
其中,姿态矩阵
Figure PCTCN2021083032-appb-000118
可以分解成两个变换矩阵,即
Figure PCTCN2021083032-appb-000119
Figure PCTCN2021083032-appb-000120
并且
Figure PCTCN2021083032-appb-000121
其中,
Figure PCTCN2021083032-appb-000122
例如可以是一个3×3大小的矩阵,其表示将位姿从环境坐标系(E,N,U)经过平移和旋转等映射到有误差的环境坐标系(Xn′,Yn′,Zn′);
Figure PCTCN2021083032-appb-000123
例如可以是一个3×3大小的矩阵,其表示将位姿从后轮轴地面投影坐标系(Xm,Ym,Zm)经过平移和旋转等映射到环境坐标系(E,N,U)。
Figure PCTCN2021083032-appb-000124
可以进一步分解成两个变换矩阵,即
Figure PCTCN2021083032-appb-000125
Figure PCTCN2021083032-appb-000126
并且
Figure PCTCN2021083032-appb-000127
其中,
Figure PCTCN2021083032-appb-000128
例如可以是一个3×3大小的矩阵,其表示将位姿从LiDAR坐标系(X l,Y l,Z l)经过平移和旋转等映射到环境坐标系(E,N,U);
Figure PCTCN2021083032-appb-000129
例如可以是一个3×3大小的矩阵,其表示将位姿从后轮轴地面投影坐标系(Xm,Ym,Zm)经过平移和旋转等映射到LiDAR坐标系(X l,Y l,Z l)。
进而得到以下公式(12):
Figure PCTCN2021083032-appb-000130
进一步地,本申请实施例可以基于公式(12)向组合定位滤波器的状态变量中引入INS的姿态误差,以及IMU与LiDAR之间的安装偏差角。
示例地,以东-北-天坐标系描述的INS的姿态误差
Figure PCTCN2021083032-appb-000131
可以为:
Figure PCTCN2021083032-appb-000132
其中,
Figure PCTCN2021083032-appb-000133
为INS的E轴姿态误差、
Figure PCTCN2021083032-appb-000134
为INS的N轴姿态误差、
Figure PCTCN2021083032-appb-000135
为INS的U轴姿态误差。则有:
Figure PCTCN2021083032-appb-000136
其中,
Figure PCTCN2021083032-appb-000137
表示
Figure PCTCN2021083032-appb-000138
的叉乘,I为单位矩阵:
Figure PCTCN2021083032-appb-000139
示例地,IMU与LiDAR之间的安装偏差角λ 1可以为:
Figure PCTCN2021083032-appb-000140
其中,
Figure PCTCN2021083032-appb-000141
为X轴方向的安装偏差角、
Figure PCTCN2021083032-appb-000142
为Y轴方向的安装偏差角、
Figure PCTCN2021083032-appb-000143
为Z轴方向的安装偏差角。则有:
Figure PCTCN2021083032-appb-000144
其中,λ 1×表示λ 1的叉乘。
由此,根据公式(12)可进一步得到:
Figure PCTCN2021083032-appb-000145
上述公式(13)即可以作为从INS对应的姿态矩阵到LiDAR对应的姿态矩阵的转移关系方程。
接下来,将公式(10)、公式(11)与公式(13)联立,可推导出INS对应的姿态矩阵
Figure PCTCN2021083032-appb-000146
和LiDAR对应的姿态矩阵
Figure PCTCN2021083032-appb-000147
中的各个元素间的关系如公式(14)所示:
Figure PCTCN2021083032-appb-000148
将公式(14)的左端按照泰勒级数展开,并且忽略公式中的二阶以上的高阶项,可以得到公式(15):
Figure PCTCN2021083032-appb-000149
该公式(15)可以称作IMU与LiDAR之间的安装偏差角的误差模型。其中,Z 1,Z 2,Z 3为该误差模型的观测量;
Figure PCTCN2021083032-appb-000150
为前述提到的INS的姿态误差,例如平台误差角;
Figure PCTCN2021083032-appb-000151
Figure PCTCN2021083032-appb-000152
为前述提到的IMU与LiDAR之间的安装偏差角。
步骤S204,构建系统状态方程,系统状态方程的状态变量包括INS与LiDAR之间的安装偏差角。
一般来说,组合定位系统中原有的状态方程是基于惯导误差模型构建的方程,状态方程中的状态变量为惯导误差,例如:位置误差、速度误差、姿态误差、加速度计零偏差和陀螺仪零偏差等,而本申请实施例可以将IMU与LiDAR之间的安装偏差角的误差模型引入到原有的状态方程中,使新的状态方程的状态变量既包含惯导误差相关的变量,也包含IMU与LiDAR之间的安装偏差角的误差相关的变量。
在一个实施例中,本申请实施例定义该组合定位系统新的系统状态方程包括:
X=[X 1,X 3]      (16)
其中,X 1为状态变量中的惯导误差,X 3为IMU与LiDAR之间的安装偏差角。
则有,X 1的系统状态方程为:
Figure PCTCN2021083032-appb-000153
其中,
Figure PCTCN2021083032-appb-000154
为惯导误差X 1的导数;F 1为惯导误差模型中可以推导的系统矩阵,例如可以是惯导误差模型的非线性转移矩阵;G 1,W 1为惯导误差驱动噪声,其中,G 1可以是惯导误差的噪声驱动矩阵,W 1可以是惯导误差的状态噪声矩阵,可选的,G 1和W 1均为高斯白噪声。
示例地,
Figure PCTCN2021083032-appb-000155
其中,
Figure PCTCN2021083032-appb-000156
分别为东、北、天方向平台失准角,δv E,δv N,δv U分别为INS在东、北、天方向的速度误差,δλ,δL,δH分别为INS的经度、纬度和高度误差,ε xyz为陀螺仪随机常值漂移在X,Y,Z轴的分量,
Figure PCTCN2021083032-appb-000157
为INS的加速度计零偏在X,Y,Z轴的分量。可以理解的是,上述X 1包含的参数项仅仅作为一个示例,不构成对X 1的限定,在一些其他的实现方式中,X 1可以包含更多的参数项或者更少的参数项,本申请实施例此处不做具体展开说明。
以及,X 2的系统状态方程为:
Figure PCTCN2021083032-appb-000158
其中,
Figure PCTCN2021083032-appb-000159
为IMU与LiDAR之间的安装偏差角X 3的导数;F 3为IMU与LiDAR之间的安装偏差角的误差模型中可以推导的系统矩阵,例如可以是该误差模型的非线性转移矩阵; G 3,W 3为IMU与LiDAR之间的安装偏差角的误差驱动噪声,其中,G 3可以是IMU与LiDAR之间的安装偏差角的误差的噪声驱动矩阵,W 3可以是IMU与LiDAR之间的安装偏差角的误差的状态噪声矩阵,可选的,G 3和W 3均为高斯白噪声。
示例地,
Figure PCTCN2021083032-appb-000160
其中,
Figure PCTCN2021083032-appb-000161
为IMU与LiDAR之间的X轴方向的安装偏差角、
Figure PCTCN2021083032-appb-000162
为IMU与LiDAR之间的Y轴方向的安装偏差角、
Figure PCTCN2021083032-appb-000163
为IMU与LiDAR之间的Z轴方向的安装偏差角。可以理解的是,上述X 3包含的参数项仅仅作为一个示例,不构成对X 3的限定,在一些其他的实现方式中,X 3可以包含更多的参数项或者更少的参数项,本申请实施例此处不做具体展开说明。
另设,IMU与LiDAR之间的安装偏差角的误差为
Figure PCTCN2021083032-appb-000164
Figure PCTCN2021083032-appb-000165
对应的系统状态方程为:
Figure PCTCN2021083032-appb-000166
其中,
Figure PCTCN2021083032-appb-000167
Figure PCTCN2021083032-appb-000168
的导数。
以上,公式(16)-(18)即构成了本申请实施例的组合定位系统的系统状态方程,可以看出,该系统状态方程的状态变量不仅包含了惯导误差,还包含了IMU与LiDAR之间的安装偏差角,由此,在后续使用组合定位滤波器对系统状态方程的状态变量进行在线估计时,就能够同时对IMU与LiDAR之间的安装偏差角进行在线标定。
步骤S205,构建系统观测方程,系统观测方程的观测量包括根据误差模型确定的INS与LiDAR之间的安装偏差角对应的观测矩阵。
具体来说,该系观测方程的观测量为惯导误差对应的观测矩阵和IMU与LiDAR之间的安装偏差角的误差对应的观测矩阵。
在一个实施例中,该系统观测方程可以为:
Z=HX+V=[H 1 H 2 H 3]
其中,Z为系统观测量,H为系统观测矩阵,X为本申请实施例在步骤S205中建立的新的系统状态方程,V为系统观测噪声矩阵。进一步地,H 1为姿态误差对应的观测矩阵,H 2为误差模型中除了姿态误差以外的其他状态变量对应的观测矩阵,H 3为IMU与LiDAR之间的安装偏差角对应的观测矩阵。H 2可根据惯导误差模型得到,H 1和H 3如下所示:
Figure PCTCN2021083032-appb-000169
步骤S206,使用组合定位滤波器对系统状态方程和系统观测方程进行滤波解算,以得到IMU与LiDAR之间的安装偏差角的最优估计结果。
具体实现中,组合定位滤波器可以是卡尔曼滤波器,即对系统状态方程和系统观测方程 进行卡尔曼滤波,以得到IMU与LiDAR之间的安装偏差角的最优估计结果。
步骤S207,根据IMU与LiDAR之间的安装偏差角的最优估计结果对LiDAR定位结果进行补偿。
这样,LiDAR与IMU之间消除了由于安装偏差角而产生的实际安装位置关系与其坐标系位置关系的不一致性,从而消除了由安装偏差角造成的组合定位误差,提高组合定位系统的定位精度。
另外,本申请实施例还可以根据卡尔曼滤波得到的状态变量最优估计结果对惯导解算结果,例如位置、速度和姿态等进行校正,以修正惯导解算结果的累积误差,提高组合定位系统的定位精度。
步骤S208,当组合定位滤波器对的滤波解算结果收敛时,将IMU与LiDAR之间的安装偏差角的最优估计结果作为标定结果保存至配置文件中。
在此之后,如果IMU与LiDAR之间的安装位置关系没有发生变化,那么IMU与LiDAR之间的安装偏差角就不会改变,那么卡尔曼滤波器对的滤波解算结果将一直处于收敛状态;如果车辆在行驶过程中发生振动或者车辆维修等导致IMU与LiDAR之间的安装位置关系,那么当卡尔曼滤波器对的滤波解算结果经过短暂的不收敛状态之后也会再次进入到收敛状态,这时,IMU与LiDAR之间的安装偏差角的最优估计结果会发生变化,组合定位系统可以在卡尔曼滤波器重新收敛之后,将最新的标定结果保存至配置文件中。
由此可见,本申请实施例的方法将IMU与LiDAR之间的安装偏差角整合到组合定位系统的状态变量中,实现了在车辆行驶过程中对组合定位系统的IMU与LiDAR之间的安装偏差角进行标定(即在线标定),标定过程不需要考虑IMU和LiDAR的物理结构,标定精度高。并且,本申请实施例的方法不需要借助用的标定工具,仅需要车辆产生机动即可完成标定,操作简单、快速。另外,当IMU与LiDAR之间的安装位置关系发生变化时,本申请实施例的方法能够在线更新标定结果,从而始终保证组合定位系统的定位精度。
需要补充说明的是,本申请实施例的方法可以应用到包含IMU和GNSS的组合定位系统中,以标定IMU和GNSS双天线之间的安装偏差角;本申请实施例的方法可以应用到包含IMU和LiDAR的组合定位系统中,以标定IMU和LiDAR之间的安装偏差角;本申请实施例的方法可以应用到包含IMU、GNSS和LiDAR的组合定位系统中,以标定IMU和GNSS双天线之间的安装偏差角和IMU和LiDAR之间的安装偏差角,其中X=[X 1,X 2,X 3]。
上述本申请提供的实施例对组合定位系统实现传感器安装偏角标定方法的各方案进行了介绍。可以理解的是,组合定位系统为了实现上述功能,可以包含执行各个功能相应的硬件结构和/或软件模块。本领域技术人员应该很容易意识到,结合本文中所公开的实施例描述的各示例的单元及算法步骤,本申请能够以硬件或硬件和计算机软件的结合形式来实现。某个功能究竟以硬件还是计算机软件驱动硬件的方式来执行,取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能,但是这种实现不应认为超出本申请的范围。
在一个实施例中,组合定位系统可以通过如4所示的硬件结构实现相应的功能。其中,各个传感器可以安装到车辆的不同位置,存储器和处理器例如可以是车辆的行车电脑,例如电子控制单元(electronic control unit,ECU),或者自动驾驶计算平台,例如移动数据中心(mobile data center,MDC)中的存储器和处理器。存储器包括有用于实现相应的功能的程序指令,程序指令被处理器运行时,使得组合定位系统用于执行如下步骤:根据各个传感器 输出的姿态信息,构建各个传感器之间的安装偏差角的误差模型;构建系统状态方程,系统状态方程的状态变量包括安装偏差角;构建系统观测方程,系统观测方程的观测量包括根据误差模型确定的安装偏差角对应的观测矩阵;对系统状态方程和系统观测方程进行滤波解算,将滤波解算结果收敛时得到的安装偏差角的估计结果作为标定结果。
在一个实施例中,组合定位系统用于标定IMU和GNSS天线之间的安装偏差角,因此多个传感器可以至少包括IMU和GNSS接收机,GNSS接收机包括GNSS天线,例如主天线和从天线,简称GNSS双天线,IMU和GNSS的每根天线安装在车辆的不同位置;程序指令被处理器运行时,使得组合定位系统用于执行如下步骤,以实现根据各个传感器输出的姿态信息,构建各个传感器之间的安装偏差角的误差模型:根据INS和GNSS输出的姿态信息,构造误差模型的观测量,观测量为IMU和GNSS天线之间的姿态偏差;构建从INS对应的姿态矩阵到GNSS对应的姿态矩阵的转移关系方程;联立误差模型的观测量和转移关系方程构建误差模型,误差模型为IMU和GNSS天线之间的安装偏差角的误差模型;其中,姿态信息包括滚转角、俯仰角和航向角中的一个或者多个,转移关系方程包含INS的姿态误差和IMU与GNSS天线之间的安装偏差角。
在一个实施例中,程序指令被处理器运行时,使得组合定位系统用于执行如下步骤,以实现对系统状态方程和系统观测方程进行滤波解算,将滤波解算结果收敛时得到的安装偏差角的最优估计结果作为标定结果:对系统状态方程和系统观测方程进行卡尔曼滤波解算,得到IMU与GNSS天线之间的安装偏差角的最优估计结果;根据最优估计结果对GNSS的定位结果进行补偿,以及对INS的惯导解算结果进行修正;当滤波解算结果收敛时,将IMU与GNSS天线之间的安装偏差角的最优估计结果作为标定结果保存至配置文件中。
在一个实施例中,组合定位系统用于标定IMU和LiDAR之间的安装偏差角,因此多个传感器至少包括IMU和LiDAR,IMU和LiDAR安装在车辆的不同位置;程序指令被处理器运行时,使得组合定位系统用于执行如下步骤,以实现根据各个传感器输出的姿态信息,构建各个传感器之间的安装偏差角的误差模型:根据INS和LiDAR输出的姿态信息,构造误差模型的观测量,观测量为IMU和LiDAR之间的姿态偏差;构建从INS对应的姿态矩阵到LiDAR对应的姿态矩阵的转移关系方程;联立误差模型的观测量和转移关系方程构建误差模型,误差模型为IMU和LiDAR之间的安装偏差角的误差模型;其中,姿态信息包括滚转角、俯仰角和航向角中的一个或者多个,转移关系方程包含INS的姿态误差和IMU与LiDAR之间的安装偏差角。
在一个实施例中,程序指令被处理器运行时,使得组合定位系统用于执行如下步骤,以实现对系统状态方程和系统观测方程进行滤波解算,将滤波解算结果收敛时得到的安装偏差角的最优估计结果作为标定结果:对系统状态方程和系统观测方程进行卡尔曼滤波解算,得到IMU与LiDAR之间的安装偏差角的最优估计结果;根据最优估计结果对LiDAR的定位结果进行补偿,以及对INS的惯导解算结果进行修正;当滤波解算结果收敛时,将IMU与LiDAR之间的安装偏差角的最优估计结果作为标定结果保存至配置文件中。
在另一个实施例中,组合定位系统可以通过图8所示的软件模块实现相应的功能。如图8所示,组合定位系统可以包括惯导解算模块310、误差模型构建模块320、系统状态方程构建模块330、系统状态方程构建模块340和滤波解算模块350。下面对上述模块的功能进行具体说明:
如图9所示,当组合定位系统用于标定IMU和GNSS天线之间的安装偏差角时:
惯导解算模块310,用于对IMU的测量结果进行惯导解算,以得到INS的姿态信息。
误差模型构建模块320,用于根据INS的姿态信息和GNSS姿态信息构建INS与GNSS天线之间的安装偏差角的误差模型;
系统状态方程构建模块330,用于构建系统状态方程,系统状态方程的状态变量包括INS与GNSS天线之间的安装偏差角;
系统观测方程构建模块340,用于构建系统观测方程,系统观测方程的观测量包括根据误差模型确定的INS与GNSS天线之间的安装偏差角对应的观测矩阵。
滤波解算模块350,用于对系统状态方程和系统观测方程进行滤波解算,将滤波解算结果收敛时得到的INS与GNSS天线之间的安装偏差角的估计结果作为标定结果。
在一个实施例中,误差模型构建模块320,具体可以用于:根据INS和GNSS输出的姿态信息,构造误差模型的观测量;构建从INS对应的姿态矩阵到GNSS对应的姿态矩阵的转移关系方程;联立误差模型的观测量和转移关系方程构建误差模型;其中,姿态信息包括滚转角、俯仰角和航向角中的一个或者多个,转移关系方程包含INS的姿态误差和IMU与GNSS天线之间的安装偏差角。
在一个实施例中,滤波解算模块350,具体可以用于:对系统状态方程和系统观测方程进行卡尔曼滤波解算,得到IMU与GNSS天线之间的安装偏差角的最优估计结果;根据最优估计结果对GNSS的定位结果进行补偿,以及对INS的惯导解算结果进行修正;当滤波解算结果收敛时,将IMU与GNSS天线之间的安装偏差角的最优估计结果作为标定结果保存至配置文件中。
进一步如图10所示,当组合定位系统用于标定IMU和LiDAR之间的安装偏差角时:
惯导解算模块310,用于对IMU的测量结果进行惯导解算,以得到INS的姿态信息。
误差模型构建模块320,用于根据INS的姿态信息和LiDAR姿态信息构建INS与LiDAR之间的安装偏差角的误差模型;
系统状态方程构建模块330,用于构建系统状态方程,系统状态方程的状态变量包括INS与LiDAR之间的安装偏差角;
系统观测方程构建模块340,用于构建系统观测方程,系统观测方程的观测量包括根据误差模型确定的INS与LiDAR之间的安装偏差角对应的观测矩阵。
滤波解算模块350,用于对系统状态方程和系统观测方程进行滤波解算,将滤波解算结果收敛时得到的INS与LiDAR之间的安装偏差角的估计结果作为标定结果。
在一个实施例中,误差模型构建模块320,具体可以用于:根据INS和LiDAR输出的姿态信息,构造误差模型的观测量;构建从INS对应的姿态矩阵到LiDAR对应的姿态矩阵的转移关系方程;联立误差模型的观测量和转移关系方程构建误差模型;其中,姿态信息包括滚转角、俯仰角和航向角中的一个或者多个,转移关系方程包含INS的姿态误差和IMU与LiDAR之间的安装偏差角。
在一个实施例中,滤波解算模块350,具体可以用于:对系统状态方程和系统观测方程进行卡尔曼滤波解算,得到IMU与LiDAR之间的安装偏差角的最优估计结果;根据最优估计结果对LiDAR的定位结果进行补偿,以及对INS的惯导解算结果进行修正;当滤波解算结果收敛时,将IMU与LiDAR之间的安装偏差角的最优估计结果作为标定结果保存至配置文件中。
本申请实施例还提供了一种车辆,该车辆可以包含前述各实施例提供的组合定位系统, 并且用户执行前述各个实施例提供的传感器安装偏差角的标定方法。
本申请实施例还提供一种计算机可读存储介质,该计算机可读存储介质中存储有指令,当其在计算机上运行时,使得计算机执行上述各方面的方法。
本申请实施例还提供了一种包含指令的计算机程序产品,当其在计算机上运行时,使得计算机执行上述各方面的方法。
以上的具体实施方式,对本申请实施例的目的、技术方案和有益效果进行了进一步详细说明,所应理解的是,以上仅为本申请实施例的具体实施方式而已,并不用于限定本申请实施例的保护范围,凡在本申请实施例的技术方案的基础之上,所做的任何修改、等同替换、改进等,均应包括在本申请实施例的保护范围之内。

Claims (21)

  1. 一种传感器安装偏差角的标定方法,其特征在于,应用于组合定位系统,所述组合定位系统包括多个传感器,所述多个传感器的安装位置不同;
    所述方法包括:
    根据各个传感器输出的姿态信息,构建所述各个传感器之间的安装偏差角的误差模型;
    构建系统状态方程,所述系统状态方程的状态变量包括所述安装偏差角;
    构建系统观测方程,所述系统观测方程的观测量包括根据所述误差模型确定的所述安装偏差角对应的观测矩阵;
    对所述系统状态方程和所述系统观测方程进行滤波解算,将滤波解算结果收敛时得到的所述安装偏差角的估计结果作为标定结果。
  2. 根据权利要求1所述的方法,其特征在于,所述多个传感器包括惯性导航系统INS的惯性测量单元IMU和全球导航卫星系统GNSS接收机,所述GNSS接收机包括GNSS天线,所述IMU和所述GNSS天线的安装位置不同,
    所述根据各个传感器输出的姿态信息,构建所述各个传感器之间的安装偏差角的误差模型,包括:
    构造所述误差模型的观测量,所述观测量为所述IMU和所述GNSS天线之间的姿态偏差;
    构建从所述INS对应的姿态矩阵到所述GNSS对应的姿态矩阵的转移关系方程;
    联立所述误差模型的观测量和所述转移关系方程构建所述误差模型,所述误差模型为所述IMU和所述GNSS天线之间的安装偏差角的误差模型;
    其中,所述姿态信息包括滚转角、俯仰角和航向角中的一个或者多个,所述转移关系方程包含所述INS的姿态误差和所述IMU与所述GNSS天线之间的安装偏差角。
  3. 根据权利要求2所述的方法,其特征在于,所述误差模型的观测量包括:
    Figure PCTCN2021083032-appb-100001
    其中,γ m、θ m
    Figure PCTCN2021083032-appb-100002
    分别为INS输出的滚转角、俯仰角和航向角;γ g、θ g
    Figure PCTCN2021083032-appb-100003
    分别为GNSS输出的滚转角、俯仰角和航向角。
  4. 根据权利要求2或3所述的方法,其特征在于,所述转移关系方程包括:
    Figure PCTCN2021083032-appb-100004
    其中,
    Figure PCTCN2021083032-appb-100005
    为所述INS对应的姿态矩阵,
    Figure PCTCN2021083032-appb-100006
    为所述GNSS对应的姿态矩阵,
    Figure PCTCN2021083032-appb-100007
    为所述INS的姿态误差,λ 0为所述IMU与所述GNSS天线之间的安装偏差角,λ 0×表示λ 0的叉乘。
  5. 根据权利要求2-4任一项所述的标定方法,其特征在于,所述系统状态方程包括:
    X=[X 1,X 2]
    Figure PCTCN2021083032-appb-100008
    Figure PCTCN2021083032-appb-100009
    其中,X 1和X 2为状态变量,X 1为惯导误差,X 2为所述IMU与所述GNSS天线之间的安 装偏差角;
    Figure PCTCN2021083032-appb-100010
    为X 1的导数;F 1为所述组合定位系统的惯导误差模型的系统矩阵,G 1、W 1为惯导误差驱动噪声;
    Figure PCTCN2021083032-appb-100011
    为X 2的导数;F 2为所述IMU与所述GNSS天线之间的安装偏差角的误差模型的系统矩阵,G 2、W 2为所述IMU与所述GNSS天线之间的安装偏差角的误差驱动噪声。
  6. 根据权利要求5所述的标定方法,其特征在于,所述系统状态方程包括:
    Z=HX+V=[H 1 H 2 H 3]
    其中,Z为系统观测量,H为系统观测矩阵,X为所述系统状态方程,V为系统观测噪声矩阵,H 1为INS的姿态误差对应的观测矩阵,H 2为所述惯导误差模型中除了所述姿态误差以外的其他状态变量对应的观测矩阵,H 3为所述IMU与所述GNSS天线之间的安装偏差角对应的观测矩阵。
  7. 根据权利要求2-6任一项所述的标定方法,其特征在于,所述对所述系统状态方程和所述系统观测方程进行滤波解算,将滤波解算结果收敛时得到的所述安装偏差角的最优估计结果作为标定结果,包括:
    对所述系统状态方程和所述系统观测方程进行卡尔曼滤波解算,得到所述IMU与所述GNSS天线之间的安装偏差角的最优估计结果;
    根据所述最优估计结果对所述GNSS的定位结果进行补偿,以及对所述INS的惯导解算结果进行修正;
    当所述滤波解算结果收敛时,将所述IMU与所述GNSS天线之间的安装偏差角的最优估计结果作为标定结果保存至配置文件中。
  8. 根据权利要求1所述的标定方法,其特征在于,所述多个传感器包括IMU和激光雷达LiDAR,所述IMU和所述LiDAR的安装位置不同;
    所述根据各个传感器输出的姿态信息,构建所述各个传感器之间的安装偏差角的误差模型,包括:
    构造所述误差模型的观测量,所述观测量为所述IMU和所述LiDAR之间的姿态偏差;
    构建从所述INS对应的姿态矩阵到所述LiDAR对应的姿态矩阵的转移关系方程;
    联立所述误差模型的观测量和所述转移关系方程构建所述误差模型,所述误差模型为所述IMU和所述LiDAR之间的安装偏差角的误差模型;
    其中,所述姿态信息包括滚转角、俯仰角和航向角中的一个或者多个,所述转移关系方程包含所述INS的姿态误差和所述IMU与所述LiDAR之间的安装偏差角。
  9. 根据权利要求8所述的方法,其特征在于,所述误差模型的观测量包括:
    Figure PCTCN2021083032-appb-100012
    其中,γ m、θ m
    Figure PCTCN2021083032-appb-100013
    分别为INS输出的滚转角、俯仰角和航向角;γ l、θ l
    Figure PCTCN2021083032-appb-100014
    分别为LiDAR输出的滚转角、俯仰角和航向角。
  10. 根据权利要求8或9所述的方法,其特征在于,所述转移关系方程包括:
    Figure PCTCN2021083032-appb-100015
    其中,
    Figure PCTCN2021083032-appb-100016
    为所述INS对应的姿态矩阵,
    Figure PCTCN2021083032-appb-100017
    为所述LiDAR对应的姿态矩阵,
    Figure PCTCN2021083032-appb-100018
    为所述INS的姿态误差,λ 1为所述IMU与所述LiDAR天线之间的安装偏差角,λ 1×表示λ 1的叉乘。
  11. 根据权利要求8-10任一项所述的标定方法,其特征在于,所述系统状态方程包括:
    X=[X 1,X 3]
    Figure PCTCN2021083032-appb-100019
    Figure PCTCN2021083032-appb-100020
    其中,X 1和X 3为状态变量,X 1为惯导误差,X 3为所述IMU与所述LiDAR之间的安装偏差角;
    Figure PCTCN2021083032-appb-100021
    为X 1的导数;F 1为所述组合定位系统的惯导误差模型的系统矩阵,G 1、W 1为惯导误差驱动噪声;
    Figure PCTCN2021083032-appb-100022
    为X 3的导数;F 3为所述IMU与所述LiDAR之间的安装偏差角的误差模型的系统矩阵,G 3、W 3为所述IMU与所述LiDAR之间的安装偏差角的误差驱动噪声。
  12. 根据权利要求11所述的标定方法,其特征在于,所述系统状态方程包括:
    Z=HX+V=[H 1 H 2 H 3]
    其中,Z为系统观测量,H为系统观测矩阵,X为所述系统状态方程,V为系统观测噪声矩阵,H 1为INS的姿态误差对应的观测矩阵,H 2为所述惯导误差模型中除了所述姿态误差以外的其他状态变量对应的观测矩阵,H 3为所述IMU与所述LiDAR之间的安装偏差角对应的观测矩阵。
  13. 根据权利要求8-12任一项所述的标定方法,其特征在于,所述对所述系统状态方程和所述系统观测方程进行滤波解算,将滤波解算结果收敛时得到的所述安装偏差角的最优估计结果作为标定结果,包括:
    对所述系统状态方程和所述系统观测方程进行卡尔曼滤波解算,得到所述IMU与所述LiDAR之间的安装偏差角的最优估计结果;
    根据所述最优估计结果对所述LiDAR的定位结果进行补偿,以及对所述INS的惯导解算结果进行修正;
    当所述滤波解算结果收敛时,将所述IMU与所述LiDAR之间的安装偏差角的最优估计结果作为标定结果保存至配置文件中。
  14. 一种组合定位系统,其特征在于,包括:处理器、存储器和多个传感器,所述多个传感器的安装位置不同;所述存储器包括有程序指令,所述程序指令被所述处理器运行时,使得所述组合定位系统用于执行如下步骤:
    根据各个传感器输出的姿态信息,构建各个传感器之间的安装偏差角的误差模型;
    构建系统状态方程,所述系统状态方程的状态变量包括所述安装偏差角;
    构建系统观测方程,所述系统观测方程的观测量包括根据所述误差模型确定的所述安装偏差角对应的观测矩阵;
    对所述系统状态方程和所述系统观测方程进行滤波解算,将滤波解算结果收敛时得到的所述安装偏差角的估计结果作为标定结果。
  15. 根据权利要求14所述的组合定位系统,其特征在于,所述多个传感器包括惯性导航系统INS的惯性测量单元IMU和全球导航卫星系统GNSS接收机,所述GNSS接收机包括GNSS天线,所述IMU和所述GNSS天线的安装位置不同;
    所述程序指令被所述处理器运行时,使得所述组合定位系统用于执行如下步骤,以实现根据各个传感器输出的姿态信息,构建各个传感器之间的安装偏差角的误差模型:
    构造所述误差模型的观测量,所述观测量为所述IMU和所述GNSS天线之间的姿态偏差;
    构建从所述INS对应的姿态矩阵到所述GNSS对应的姿态矩阵的转移关系方程;
    联立所述误差模型的观测量和所述转移关系方程构建所述误差模型,所述误差模型为所述IMU和所述GNSS天线之间的安装偏差角的误差模型;
    其中,所述姿态信息包括滚转角、俯仰角和航向角中的一个或者多个,所述转移关系方程包含所述INS的姿态误差和所述IMU与所述GNSS天线之间的安装偏差角。
  16. 根据权利要求15所述的组合定位系统,其特征在于,所述程序指令被所述处理器运行时,使得所述组合定位系统用于执行如下步骤,以实现对所述系统状态方程和所述系统观测方程进行滤波解算,将滤波解算结果收敛时得到的所述安装偏差角的最优估计结果作为标定结果:
    对所述系统状态方程和所述系统观测方程进行卡尔曼滤波解算,得到所述IMU与所述GNSS天线之间的安装偏差角的最优估计结果;
    根据所述最优估计结果对所述GNSS的定位结果进行补偿,以及对所述INS的惯导解算结果进行修正;
    当所述滤波解算结果收敛时,将所述IMU与所述GNSS天线之间的安装偏差角的最优估计结果作为标定结果保存至配置文件中。
  17. 根据权利要求14所述的组合定位系统,其特征在于,所述多个传感器包括IMU和激光雷达LiDAR,所述IMU和所述LiDAR的安装位置不同;
    所述程序指令被所述处理器运行时,使得所述组合定位系统用于执行如下步骤,以实现根据各个传感器输出的姿态信息,构建各个传感器之间的安装偏差角的误差模型:
    构造所述误差模型的观测量,所述观测量为所述IMU和所述LiDAR之间的姿态偏差;
    构建从所述INS对应的姿态矩阵到所述LiDAR对应的姿态矩阵的转移关系方程;
    联立所述误差模型的观测量和所述转移关系方程构建所述误差模型,所述误差模型为所述IMU和所述LiDAR之间的安装偏差角的误差模型;
    其中,所述姿态信息包括滚转角、俯仰角和航向角中的一个或者多个,所述转移关系方程包含所述INS的姿态误差和所述IMU与所述LiDAR之间的安装偏差角。
  18. 根据权利要求17所述的组合定位系统,其特征在于,所述程序指令被所述处理器运行时,使得所述组合定位系统用于执行如下步骤,以实现对所述系统状态方程和所述系统观测方程进行滤波解算,将滤波解算结果收敛时得到的所述安装偏差角的最优估计结果作为标定结果:
    对所述系统状态方程和所述系统观测方程进行卡尔曼滤波解算,得到所述IMU与所述LiDAR之间的安装偏差角的最优估计结果;
    根据所述最优估计结果对所述LiDAR的定位结果进行补偿,以及对所述INS的惯导解算结果进行修正;
    当所述滤波解算结果收敛时,将所述IMU与所述LiDAR之间的安装偏差角的最优估计结果作为标定结果保存至配置文件中。
  19. 一种车辆,其特征在于,包括:组合定位系统,所述组合定位系统包括处理器、存储器和多个传感器,不同传感器安装在所述车辆的不同位置;所述存储器包括有程序指令,所述程序指令被所述处理器运行时,使得所述组合定位系统用于执行如权利要求1-13任一项所述的方法。
  20. 一种计算机可读存储介质,其特征在于,所述计算机可读存储介质中存储有指令,当所述指令在计算机上运行时,使得所述计算机执行如权利要求1-13任一项所述的方法。
  21. 一种计算机程序产品,其特征在于,当所述计算机程序产品在计算机上运行时,使得所述计算机执行如权利要求1-13任一项所述的方法。
PCT/CN2021/083032 2020-07-04 2021-03-25 传感器安装偏差角的标定方法、组合定位系统和车辆 WO2022007437A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
EP21838295.0A EP4170282A4 (en) 2020-07-04 2021-03-25 METHOD FOR CALIBRATION OF A MOUNTING DEVIATION ANGLE BETWEEN SENSORS, COMBINED POSITIONING SYSTEM AND VEHICLE

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010636730.6A CN113884102A (zh) 2020-07-04 2020-07-04 传感器安装偏差角的标定方法、组合定位系统和车辆
CN202010636730.6 2020-07-04

Publications (1)

Publication Number Publication Date
WO2022007437A1 true WO2022007437A1 (zh) 2022-01-13

Family

ID=79013283

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/083032 WO2022007437A1 (zh) 2020-07-04 2021-03-25 传感器安装偏差角的标定方法、组合定位系统和车辆

Country Status (3)

Country Link
EP (1) EP4170282A4 (zh)
CN (1) CN113884102A (zh)
WO (1) WO2022007437A1 (zh)

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114234972A (zh) * 2022-03-01 2022-03-25 长沙金维信息技术有限公司 用于惯性导航系统的惯性导航方法
CN114353792A (zh) * 2022-03-16 2022-04-15 长沙金维信息技术有限公司 车载组合导航系统中惯导系统的杆臂测量方法及导航方法
CN114413934A (zh) * 2022-01-20 2022-04-29 北京经纬恒润科技股份有限公司 一种车辆定位系统校正方法和装置
CN114526734A (zh) * 2022-03-01 2022-05-24 长沙金维信息技术有限公司 用于车载组合导航的安装角补偿方法
CN115435817A (zh) * 2022-11-07 2022-12-06 开拓导航控制技术股份有限公司 Mems惯导安装误差的标定方法、存储介质和控制计算机
CN115597595A (zh) * 2022-11-28 2023-01-13 上海仙工智能科技有限公司(Cn) 多线激光定位方法及定位装置、计算机设备、存储介质
CN115790615A (zh) * 2023-01-30 2023-03-14 安徽蔚来智驾科技有限公司 基于自动驾驶的多传感器融合方法、装置、介质及车辆
CN115825999A (zh) * 2023-02-22 2023-03-21 广州导远电子科技有限公司 一种滤波器状态监测方法、装置、电子设备及存储介质
CN115900755A (zh) * 2022-08-30 2023-04-04 中国科学院上海天文台 一种靶标指向自动修正方法及实现该方法的靶标
CN117109638A (zh) * 2023-10-24 2023-11-24 山东大学 一种在线单舵轮agv参数标定方法、系统、设备及介质
CN117420570A (zh) * 2023-12-19 2024-01-19 南京牧镭激光科技股份有限公司 一种激光雷达姿态补偿测风方法、设备以及存储介质

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115388914B (zh) * 2022-10-28 2023-02-03 福思(杭州)智能科技有限公司 传感器的参数标定方法、装置、存储介质和电子装置

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130138264A1 (en) * 2011-11-30 2013-05-30 Takayuki Hoshizaki Automotive navigation system and method to utilize internal geometry of sensor position with respect to rear wheel axis
CN103217159A (zh) * 2013-03-06 2013-07-24 郭雷 一种sins/gps/偏振光组合导航系统建模及动基座初始对准方法
CN104406605A (zh) * 2014-10-13 2015-03-11 中国电子科技集团公司第十研究所 机载多导航源综合导航仿真系统
CN104977004A (zh) * 2015-07-13 2015-10-14 湖北航天技术研究院总体设计所 一种激光惯组与里程计组合导航方法及系统
CN105509738A (zh) * 2015-12-07 2016-04-20 西北工业大学 基于惯导/多普勒雷达组合的车载定位定向方法
CN109556632A (zh) * 2018-11-26 2019-04-02 北方工业大学 一种基于卡尔曼滤波的ins/gnss/偏振/地磁组合导航对准方法
CN110849364A (zh) * 2019-12-04 2020-02-28 成都国卫通信技术有限公司 基于动中通的自适应卡尔曼姿态估计方法

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5184304A (en) * 1991-04-26 1993-02-02 Litton Systems, Inc. Fault-tolerant inertial navigation system
US7979231B2 (en) * 2008-11-13 2011-07-12 Honeywell International Inc. Method and system for estimation of inertial sensor errors in remote inertial measurement unit
CN104764467B (zh) * 2015-04-08 2018-12-14 南京航空航天大学 空天飞行器惯性传感器误差在线自适应标定方法
CN106291645B (zh) * 2016-07-19 2018-08-21 东南大学 适于高维gnss/ins深耦合的容积卡尔曼滤波方法
CN107588769B (zh) * 2017-10-17 2020-01-03 北京航天发射技术研究所 一种车载捷联惯导、里程计及高程计组合导航方法
CN111323050B (zh) * 2020-03-19 2021-06-18 哈尔滨工程大学 一种捷联惯导和多普勒组合系统标定方法

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130138264A1 (en) * 2011-11-30 2013-05-30 Takayuki Hoshizaki Automotive navigation system and method to utilize internal geometry of sensor position with respect to rear wheel axis
CN103217159A (zh) * 2013-03-06 2013-07-24 郭雷 一种sins/gps/偏振光组合导航系统建模及动基座初始对准方法
CN104406605A (zh) * 2014-10-13 2015-03-11 中国电子科技集团公司第十研究所 机载多导航源综合导航仿真系统
CN104977004A (zh) * 2015-07-13 2015-10-14 湖北航天技术研究院总体设计所 一种激光惯组与里程计组合导航方法及系统
CN105509738A (zh) * 2015-12-07 2016-04-20 西北工业大学 基于惯导/多普勒雷达组合的车载定位定向方法
CN109556632A (zh) * 2018-11-26 2019-04-02 北方工业大学 一种基于卡尔曼滤波的ins/gnss/偏振/地磁组合导航对准方法
CN110849364A (zh) * 2019-12-04 2020-02-28 成都国卫通信技术有限公司 基于动中通的自适应卡尔曼姿态估计方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP4170282A4 *

Cited By (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114413934B (zh) * 2022-01-20 2024-01-26 北京经纬恒润科技股份有限公司 一种车辆定位系统校正方法和装置
CN114413934A (zh) * 2022-01-20 2022-04-29 北京经纬恒润科技股份有限公司 一种车辆定位系统校正方法和装置
CN114234972B (zh) * 2022-03-01 2022-05-24 长沙金维信息技术有限公司 用于惯性导航系统的惯性导航方法
CN114526734A (zh) * 2022-03-01 2022-05-24 长沙金维信息技术有限公司 用于车载组合导航的安装角补偿方法
CN114234972A (zh) * 2022-03-01 2022-03-25 长沙金维信息技术有限公司 用于惯性导航系统的惯性导航方法
CN114353792A (zh) * 2022-03-16 2022-04-15 长沙金维信息技术有限公司 车载组合导航系统中惯导系统的杆臂测量方法及导航方法
CN115900755B (zh) * 2022-08-30 2024-04-02 中国科学院上海天文台 一种靶标指向自动修正方法及实现该方法的靶标
CN115900755A (zh) * 2022-08-30 2023-04-04 中国科学院上海天文台 一种靶标指向自动修正方法及实现该方法的靶标
CN115435817A (zh) * 2022-11-07 2022-12-06 开拓导航控制技术股份有限公司 Mems惯导安装误差的标定方法、存储介质和控制计算机
CN115435817B (zh) * 2022-11-07 2023-03-14 开拓导航控制技术股份有限公司 Mems惯导安装误差的标定方法、存储介质和控制计算机
CN115597595A (zh) * 2022-11-28 2023-01-13 上海仙工智能科技有限公司(Cn) 多线激光定位方法及定位装置、计算机设备、存储介质
CN115790615A (zh) * 2023-01-30 2023-03-14 安徽蔚来智驾科技有限公司 基于自动驾驶的多传感器融合方法、装置、介质及车辆
CN115790615B (zh) * 2023-01-30 2023-05-05 安徽蔚来智驾科技有限公司 基于自动驾驶的多传感器融合方法、装置、介质及车辆
CN115825999B (zh) * 2023-02-22 2023-05-02 广州导远电子科技有限公司 一种滤波器状态监测方法、装置、电子设备及存储介质
CN115825999A (zh) * 2023-02-22 2023-03-21 广州导远电子科技有限公司 一种滤波器状态监测方法、装置、电子设备及存储介质
CN117109638A (zh) * 2023-10-24 2023-11-24 山东大学 一种在线单舵轮agv参数标定方法、系统、设备及介质
CN117109638B (zh) * 2023-10-24 2024-03-15 山东大学 一种在线单舵轮agv参数标定方法、系统、设备及介质
CN117420570A (zh) * 2023-12-19 2024-01-19 南京牧镭激光科技股份有限公司 一种激光雷达姿态补偿测风方法、设备以及存储介质
CN117420570B (zh) * 2023-12-19 2024-03-29 南京牧镭激光科技股份有限公司 一种激光雷达姿态补偿测风方法、设备以及存储介质

Also Published As

Publication number Publication date
CN113884102A (zh) 2022-01-04
EP4170282A1 (en) 2023-04-26
EP4170282A4 (en) 2023-12-06

Similar Documents

Publication Publication Date Title
WO2022007437A1 (zh) 传感器安装偏差角的标定方法、组合定位系统和车辆
CN109946732B (zh) 一种基于多传感器数据融合的无人车定位方法
CN110906923B (zh) 车载多传感器紧耦合融合定位方法、系统、存储介质及车辆
US11802769B2 (en) Lane line positioning method and apparatus, and storage medium thereof
CN101846734B (zh) 农用机械导航定位方法、系统及农用机械工控机
US8041472B2 (en) Positioning device, and navigation system
CN107132563B (zh) 一种里程计结合双天线差分gnss的组合导航方法
KR101765746B1 (ko) 다중 저가형 gps를 이용한 농업용 무인 트랙터의 자율주행용 위치 추정방법 및 시스템
WO2019188745A1 (ja) 情報処理装置、制御方法、プログラム及び記憶媒体
JP7245084B2 (ja) 自動運転システム
KR102331312B1 (ko) 차량 내부 센서, 카메라, 및 gnss 단말기를 이용한 3차원 차량 항법 시스템
CN112946681B (zh) 融合组合导航信息的激光雷达定位方法
Yu et al. GPS/INS/Odometer/DR integrated navigation system aided with vehicular dynamic characteristics for autonomous vehicle application
CN116047565A (zh) 一种多传感器数据融合定位系统
Niu et al. Camera-based lane-aided multi-information integration for land vehicle navigation
US11287281B2 (en) Analysis of localization errors in a mobile object
CN112859132A (zh) 导航的方法和装置
US20240077880A1 (en) Slope location correction method and apparatus, robot and readable storage medium
Chen et al. An integrated GNSS/INS/DR positioning strategy considering nonholonomic constraints for intelligent vehicle
CN113325454B (zh) 一种基于ArduRover无人车的组合定位方法
CN113030504B (zh) 车辆测速方法、装置、车载计算机设备和存储介质
CN117346768B (zh) 同时适用于室内外的多传感器融合感知定位方法
WO2021232296A1 (zh) 一种无人机的控制方法、设备、无人机及存储介质
Kang et al. Analysis of factors affecting performance of integrated INS/SPR positioning during GPS signal Blockage
TWI394944B (zh) 車輛姿態估測系統與方法

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: 21838295

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2021838295

Country of ref document: EP

Effective date: 20230119

NENP Non-entry into the national phase

Ref country code: DE