Method for detecting attitude deviation of a vehicle's inertial measurement unit and a vehicle equipped with an inertial measurement unit

Abstract

The present invention relates to a method for determining the attitude shift of the sensor coordinate system of a vehicle (1) with respect to the vehicle coordinate system during the vehicle (1)'s driving operation in multiple iterations (IS1 to ISn). Each iteration (IS1 to ISn) includes actual measured values ​​(MW) of longitudinal acceleration and lateral acceleration determined by the inertial measurement unit (2), gravitational acceleration, state data necessary for calculating inertial force, and coefficients of a predetermined polynomial that represent the approximate relationship between the measured longitudinal acceleration and pitch angle (Φ), or between the measured lateral acceleration and roll angle (Θ). The inertial force is calculated in each iteration (IS1 to ISn), the measured values ​​(MW) of longitudinal acceleration and lateral acceleration are adjusted for contributions due to gravitational acceleration and inertial force, and the coefficients of the polynomial are updated with the adjusted measured values ​​of longitudinal acceleration and lateral acceleration. The iteration is terminated when each of the zero-order coefficients of the polynomial satisfies a predetermined convergence criterion, and the zero-order coefficients of the polynomial are output as a result of determining the attitude shift of the sensor coordinate system.

Description

[Technical Field]

[0001] The present invention relates to a method for determining the attitude deviation of a vehicle's inertial measurement unit based on the features of the higher-level conceptual part described in claim 1, and to a vehicle. [Background technology]

[0002] As described in DE102005033237A1, methods for determining and correcting the attitude deviation and offset of sensors in an inertial measurement unit of a land vehicle are known from the prior art. The inertial measurement unit includes three linear acceleration sensors and three angular velocity sensors. The desired orientation of the sensors is parallel to the coordinate axes of a Cartesian coordinate system fixed to the vehicle. The actual orientation of the sensors may deviate from the desired orientation due to incorrect attitude. The actual orientation of the linear acceleration sensors is determined by comparing the acceleration measured by the linear acceleration sensors at different locations on the vehicle with known acceleration values ​​for these different locations in a Cartesian coordinate system fixed to the vehicle. The measured acceleration can then be converted to the actual acceleration using coordinate transformations. [Overview of the project]

[0003] The object of the present invention is to provide a vehicle that is improved compared to the prior art and a method for determining the attitude deviation of an inertial measurement unit of a vehicle that is improved compared to the prior art.

[0004] The object of the present invention is achieved by a method for determining the attitude deviation of an inertial measuring unit of a vehicle having the features described in claim 1, and by a vehicle having the features described in claim 6.

[0005] Advantageous embodiments of the present invention are the subject matter of the dependent claims.

[0006] In a method for determining the misorientation of a vehicle's inertial measurement unit, the inertial measurement unit is provided specifically for measuring the vehicle's acceleration, and the three-dimensional components of the vehicle's acceleration are measured by the inertial measurement unit in a sensor coordinate system, thereby identifying a vehicle coordinate system having an x-axis, also called the roll axis in the longitudinal direction of the vehicle, a y-axis, also called the pitch axis in the lateral direction of the vehicle, and a z-axis, also called the yaw axis in the vertical direction of the vehicle. In this method, the misorientation of the plane defined by the x-axis and y-axis of the sensor coordinate system with respect to the plane defined by the x-axis and y-axis of the vehicle coordinate system is determined. That is, the misorientation of the sensor coordinate system with respect to the roll axis, i.e., the x-axis, and the pitch axis, i.e., the y-axis, of the vehicle coordinate system is determined.

[0007] According to the present invention, attitude deviation is determined iteratively in several stages while the vehicle is moving, i.e., while the vehicle is running. Each stage receives the following inputs, i.e., input values: current measurements determined by the inertial measurement unit, in particular measured longitudinal and lateral accelerations of the vehicle; gravitational acceleration; state data of the vehicle necessary for calculating inertial forces, in particular Coriolis force, centripetal force and / or Euler force, in particular the rotational speed of the vehicle, the rotation rate of the vehicle, the vehicle speed and / or the wheel rotation speed of the vehicle's wheels; and coefficients of predetermined polynomials representing approximate relationships between the measured longitudinal acceleration and the pitch angle of the vehicle, or between the measured lateral acceleration and the roll angle of the vehicle, respectively. Here, one polynomial represents the approximate relationship between the measured longitudinal acceleration and pitch angle of the vehicle, and the other polynomial represents the approximate relationship between the measured lateral acceleration and roll angle of the vehicle. Certain polynomials are, in particular, first- or second-order polynomials. In particular, these coefficients are parameters that are iteratively refined at each stage based on a predetermined starting value.

[0008] Gravitational acceleration is calculated, in particular, using the other inputs mentioned above, such as the Kalman filter.

[0009] At each stage, inertial forces, particularly the Coriolis force, centripetal force, and / or Euler force, are calculated, and the measured values ​​of longitudinal and lateral acceleration are adjusted for the contributing components due to gravitational acceleration and inertial forces, i.e., the contributing components caused by them. Furthermore, the coefficients of the polynomial are updated with the adjusted measured values ​​of longitudinal and lateral acceleration at each stage.

[0010] At each stage, the system checks whether the zeroth-order coefficients of the polynomial satisfy a specific convergence criterion.

[0011] If the zero-order coefficients of a polynomial do not satisfy a specific convergence criterion, the coefficients of the polynomial are output to the next stage, where they are used as input values.

[0012] Otherwise, that is, if the zero-order coefficients of the polynomial each satisfy a specific convergence criterion, the iteration terminates, and no further iteration steps are performed.

[0013] At the end of the iteration, the zero-order coefficients of the polynomials are output as a result of the determined attitude shift of the sensor coordinate system, that is, as a result of the determined attitude shift of the plane defined by the x and y axes of the sensor coordinate system relative to the plane defined by the x and y axes of the vehicle coordinate system, that is, as a result of the determined attitude shift of the sensor coordinate system relative to the roll axis, i.e., the x-axis, and the pitch axis, i.e., the y-axis, of the vehicle coordinate system. The zero-order coefficients correspond to the static pitch angle or static roll angle and represent the attitude shift of the sensor coordinate system relative to the pitch axis, i.e., the y-axis, or the roll axis, i.e., the x-axis, of the vehicle coordinate system. That is, the zero-order coefficient of one polynomial corresponds to the static pitch angle and represents the attitude shift of the sensor coordinate system relative to the pitch axis of the vehicle coordinate system, and the zero-order coefficient of the other polynomials corresponds to the static roll angle and represents the attitude shift of the sensor coordinate system relative to the roll axis of the vehicle coordinate system.

[0014] This misalignment of the sensor coordinate system of the inertial measurement unit with respect to a specific vehicle coordinate system can be caused, for example, by installation tolerances when installing the inertial measurement unit on the vehicle and / or manufacturing tolerances of the inertial measurement unit.

[0015] In one possible embodiment of the method, the results of the attitude misalignment determined in the sensor coordinate system (static pitch angle and static roll angle) are used to create a coordinate transformation matrix that can transform or transform the three-dimensional components of the acceleration determined in the sensor coordinate system into the vehicle coordinate system. In this way, the inertial measurement unit is calibrated, that is, the measurement error caused by the attitude misalignment of the sensor coordinate system is compensated. Therefore, this embodiment of the method is a method for calibrating, in particular, the inertial measurement unit of a vehicle.

[0016] Using the described method, only the misalignments regarding the roll axis and the pitch axis are determined. The misalignment regarding the yaw axis is not determined. It can be determined, if necessary, using another method, especially retrospectively. Therefore, the calibration corrects only the misalignments of the inertial measurement unit with respect to the roll axis and the pitch axis of the vehicle. However, this is already sufficient for some applications, for example, for the headlight range control of the vehicle headlights. Because for this headlight range control, it is sufficient that the x-y plane of the sensor coordinate system, that is, the plane defined by the x-axis and the y-axis of the sensor coordinate system, is aligned parallel to the x-y plane of the vehicle coordinate system, that is, the plane defined by the x-axis and the y-axis of the vehicle coordinate system. For headlight range control, it is particularly important that the illumination of the vehicle headlights is not too low or too high. This is ensured by the described solution. Since the rotation of the sensor coordinate system with respect to the vehicle coordinate system about the yaw axis has no relation to the headlight range control because the illumination of the vehicle headlights is not too much to the left or right, either. <U+

[0017] In particular, the described solution enables automatic self-calibration of the inertial measurement unit. <U+ <U+

[0018] An inertial measurement unit can be used in a vehicle for a number of functions, such as the aforementioned headlight range control, airbag control, electronic stability program ESP, dead reckoning, i.e., inferential navigation, and / or ego motion estimation, i.e., estimation of the movement of the vehicle. In these applications, it is important that the orientation of the inertial measurement unit, particularly the orientation of its sensor coordinate system with respect to the vehicle coordinate system, be as accurate as possible. Therefore, when manufacturing and installing the inertial measurement unit in a vehicle, attention is paid to minimizing a series of errors caused by the arrangement of sensors, particularly acceleration sensors, on the circuit board of the inertial measurement unit, the arrangement of the circuit board within the housing of the inertial measurement unit, the attachment of the inertial measurement unit to the vehicle body, and vehicle body deformation. Nevertheless, until now, a misalignment of +-3°, i.e., a posture deviation of the inertial measurement unit, has occurred. The described method enables simple and automatic calibration of an inertial measurement unit already installed in a vehicle. The described method does not require additional hardware components. Furthermore, the described method does not require additional processing steps on the production line during vehicle manufacturing. This method enables significantly better calibration and thus significantly higher accuracy.

[0019] In particular, the offset of the sensor is not important for calibration accuracy.

[0020] A vehicle according to the invention has an inertial measurement unit and a device designed and configured to carry out the method. Therefore, a vehicle according to the invention has the same advantages as the method described above.

[0021] In one possible embodiment, the device or at least one processing unit of the device is a component of the inertial measurement unit. This particularly enables the automatic self-calibration of the inertial measurement unit described above. In particular, the processing unit of the device is designed and configured to carry out the processing operations described above, particularly the calculations and decisions.

[0022] In one possible embodiment, the device has sensors for determining gravitational acceleration and / or vehicle state data necessary for calculating inertial forces. For example, each sensor is a sensor already installed on the vehicle for other purposes. In this embodiment, the device has, in particular, sensors and a processing unit, where the sensors provide corresponding sensor data to the processing unit, which is processed by the processing unit in the manner described above.

[0023] Examples of embodiments of the present invention will be described in more detail below with reference to the drawings. [Brief explanation of the drawing]

[0024] [Figure 1] A vehicle equipped with an inertial measurement unit is shown schematically.

[0025] [Figure 2] The procedure for determining mismatches in a vehicle's inertial measurement unit is outlined below. [Modes for carrying out the invention]

[0026] Corresponding parts are given the same reference numerals in all figures.

[0027] Figure 1 shows a schematic diagram of a vehicle 1 having an inertial measurement unit 2, also known as an IMU (Inertial Measurement Unit), acceleration sensor, or acceleration sensor unit. The inertial measurement unit 2 is specifically intended to measure the acceleration of the vehicle 1. The vehicle 1 also has a device designed and configured to perform a method for determining mismatches in the inertial measurement unit 2. This device, or at least one processing unit of the device, is, for example, a component of the inertial measurement unit 2.

[0028] Figure 2 shows a schematic diagram of the procedure for determining the mismatch in the inertial measurement unit 2.

[0029] The inertial measurement unit 2 is intended to be used to determine the three-dimensional components of the acceleration of the vehicle 1 in a vehicle coordinate system fixed to the vehicle. This is necessary for various vehicle functions, for example, the ABS function (ABS = anti-lock braking system), the ESP function (ESP = electronic stability program) and / or the headlamp range control.

[0030] As shown in FIG. 1, a predetermined vehicle coordinate system has a roll axis X V also known as the x-axis, a pitch axis Y V also known as the y-axis, and a yaw axis Z V also known as the z-axis in the vertical direction of the vehicle. The origin of the predetermined vehicle coordinate system is located, in particular, at the center of gravity of the vehicle 1. The predetermined vehicle coordinate system is fixed, in particular, to the vehicle body of the vehicle 1. Specifically, the yaw axis Z V extends upward parallel to the normal vector of the cab floor and the cab roof of the vehicle 1. The roll axis X V extends, in particular, perpendicular to the normal vector of the cab floor and the cab roof of the vehicle 1 and parallel to the longitudinal axis of the vehicle. The pitch axis Y V extends, in particular, parallel to the transverse axis of the vehicle 1 and perpendicular to the normal vector of the cab floor and the cab roof of the vehicle 1.

[0031] In fact, the inertial measurement unit 2 measures the three-dimensional components of the acceleration of the vehicle 1 in the sensor coordinate system of the inertial measurement unit 2. The sensor coordinate system has an x-axis X, a y-axis Y and a z-axis Z. The sensor coordinate system has a misalignment in attitude with respect to a specific vehicle coordinate system, for example, due to installation tolerances and / or manufacturing tolerances. That is, it is rotated with respect to the vehicle coordinate system. In other words, the sensor coordinate system is rotated by a static pitch angle Φ with respect to the pitch axis Y V of the vehicle 1, rotated by a static roll angle Θ with respect to the roll axis X V of the vehicle 1, and rotated by a static yaw angle Ψ with respect to the yaw axis Z V of the vehicle 1.

[0032] In this method, the roll axis X V and the pitch axis Y VThe attitude displacement of the plane, defined by the x-axis X and y-axis Y of the sensor coordinate system relative to the plane defined by the system, is determined by the device, in particular by the processing unit of the device. That is, the roll axis X of the vehicle coordinate system V And with respect to the pitch axis Y V The attitude shift of the sensor coordinate system relative to the vehicle is determined, and therefore the static roll angle Θ and static pitch angle Φ are determined. Using these angles Θ and Φ, the measured value MW of the inertial measurement unit 2 recorded in the sensor coordinate system is transformed into the vehicle coordinate system by coordinate transformation. This transformation calibrates the inertial measurement unit 2 to the vehicle coordinate system.

[0033] As shown in Figure 2, the mismatch, i.e., the static pitch angle Φ and static roll angle Θ, is determined iteratively in several iterative stages IS1 to ISn during travel.

[0034] Each stage IS1 to ISn represents the current measurement MW of the inertial measurement unit 2, i.e., the longitudinal acceleration a of the vehicle 1 determined by the inertial measurement unit 2. x,roh and lateral acceleration a x,roh The device receives, as input values, the current measured value MW, and gravitational acceleration g and state data of vehicle 1, which are necessary to calculate inertial forces, particularly the Coriolis force, centripetal force and / or Euler force, particularly the rotational speed of vehicle 1, the rotation rate of vehicle 1, the speed of vehicle 1 and / or the wheel rotation speed of vehicle 1. It is advantageous for the device to have corresponding sensors for this purpose. These sensors may be, for example, sensors already present in vehicle 1, and in particular, sensors intended for other vehicle applications.

[0035] Furthermore, each stage IS1 to ISn receives, as input values, the coefficients of a predetermined polynomial, particularly a first- or second-order polynomial. These coefficients, in each case, are the measured longitudinal acceleration a of the vehicle 1. x,roh The distance between the pitch angle Φ or the measured lateral acceleration a of the vehicle 1 y,rohThis represents an approximate relationship between and the roll angle Θ. These coefficients are parameters P, which are iteratively refined in each stage IS1 to ISn, starting from a predetermined initial value. Thus, the starting values ​​for the first stage IS1 are the parameters P, i.e., the coefficients Φ0, Φ1, Φ2, Θ0, Θ1, Θ2.

[0036] In each stage IS1 to ISn, the inertial force is calculated, and the longitudinal acceleration a x,roh and lateral acceleration a y,roh The measured MW is adjusted for contributions from gravitational acceleration and inertial forces.

[0037] Inertial forces, including the Coriolis force, centripetal force, and Euler force, can be determined using measurements from a rotation rate sensor. Alternatively, the yaw rate, i.e., the rotation rate of vehicle 1, can be derived from the wheel speed and / or wheel rotation speed of vehicle 1. Measurements adjusted for gravitational acceleration can be obtained from the measured state data of vehicle 1, particularly speed, acceleration, and / or rotation rate, using a Kalman filter. In particular, gravitational acceleration is calculated using a Kalman filter. By subtracting this from the measurement, measurements adjusted for gravitational acceleration can be favorably obtained.

[0038] Furthermore, the coefficients Φ0, Φ1, Φ2, Θ0, Θ1, and Θ2 of the polynomial are updated by the measured values ​​MW adjusted at each stage from IS1 to ISn.

[0039] The solution described here specifically concerns the pitch angle Φ and longitudinal acceleration a of vehicle 1. x,roh Between, or between the roll angle Θ and the lateral acceleration a y,roh This is based on the idea that the relationship between them can be accurately approximated using a polynomial.

[0040]

number

[0041]

number

[0042] Here, Φ0, Φ1, Φ2, Θ0, Θ1, Θ2 are polynomials Φ(a x,roh ),Θ(a y,roh This represents the coefficient of a x,roh ,a y,roh This represents the unadjusted measured values ​​of longitudinal and lateral acceleration in MW.

[0043] For the purposes of the solution described here, it is sufficient to use a quadratic polynomial. Therefore, we will use the following polynomial for the calculation.

[0044]

number

[0045]

number

[0046] However, it is also possible to use linear polynomials or polynomials of degree 3 or higher.

[0047] The coefficients of the polynomial are iteratively optimized in each iteration stage IS1 to ISn, starting from a predetermined initial value. That is, the system learns step-by-step which coefficients are optimal. For this purpose, the error value is calculated as follows:

[0048]

number

[0049] Using a quadratic polynomial, the result is as follows:

[0050]

number

[0051] Here, a x,roh ,a y,rohrepresents the unadjusted measured value MW of longitudinal or lateral acceleration, and a x ,a y, ,a z This represents the measured acceleration in MW, adjusted for gravitational acceleration and inertial force.

[0052] And the error value a z,err The following delta coefficient is formed using this method.

[0053]

number

[0054]

number

[0055] Next, update the coefficients of the polynomial as follows.

[0056]

number

[0057]

number

[0058] Here, LR Φi ,LR Θi This is the predetermined update value (learning rate).

[0059] It is preferable to update this coefficient only within a specific speed range, for example, within a driving speed range of 0 km / h to 100 km / h.

[0060] In each iteration stage IS1 to ISn, it is checked whether the zero-order coefficients Φ0 and Θ0 of the polynomial satisfy predetermined convergence criteria. For example, it is determined whether the maximum deviation of the coefficient Φ0 obtained within a predetermined period and the maximum deviation of the coefficient Θ0 obtained within a predetermined time are both below a predetermined limit value.

[0061] If the zero-order coefficients Φ0 and Θ0 do not satisfy the respective convergence criteria, the polynomial coefficients Φ0, Φ1, Φ2, Θ0, Θ1, and Θ2 are output to the next iteration stage IS2 to ISn. Otherwise, i.e., if the zero-order coefficients Φ0 and Θ0 satisfy the respective convergence criteria, the iteration terminates.

[0062] At the end of the iteration, the zero-order coefficients Φ0 and Θ0, i.e., the last determined values ​​of the zero-order coefficients Φ0 and Θ0, are output as the result of determining the shift of the sensor coordinate system.

[0063] The determined values ​​of these zero-order coefficients Φ0 and Θ0 correspond to the static pitch angle Φ or the static roll angle Θ, and to the pitch axis Y in the vehicle coordinate system. V or roll axis X V This represents the shift in the sensor coordinate system relative to the given coordinates.

[0064] Using the determined displacements, i.e., the static pitch angle Φ and the static roll angle Θ, i.e., the last determined zero-order coefficients Φ0 and Θ0, the coordinate transformation matrix R is constructed as follows:

[0065]

number

[0066] The coordinate transformation matrix R is used to transform the three-dimensional components of the acceleration determined in the sensor coordinate system into the vehicle coordinate system. In this way, the inertial measurement unit 2 is calibrated, that is, measurement errors caused by the shift in the sensor coordinate system are compensated for.

[0067] In particular, the method described above is performed only when there are no large static pitch angles Φ of vehicle 1 other than the stationary position, because this would be included in the sensor attitude learning. For example, information from the rear level sensor and / or other chassis components of vehicle 1 is used to detect a static pitching state, i.e., a static pitch angle Φ of vehicle 1 that exceeds a predetermined limit value, and the procedure is paused in such a phase. [Explanation of Symbols]

[0068] 1 vehicle 2 Inertial measurement device IS1~IS Repetition Level MW measurement P parameter X x-axis sensor coordinate system Y-axis sensor coordinate system Z-axis sensor coordinate system X V Roll axis Y V Pitch axis Z V Yaw axis Θ Roll angle Φ pitch angle Ψ Yaw angle

Claims

1. A method for determining the attitude deviation of an inertial measurement unit (2) of a vehicle (1), wherein the three-dimensional components of the acceleration of the vehicle (1) are measured in a sensor coordinate system by the inertial measurement unit (2), A vehicle coordinate system is defined for the vehicle (1) having the x-axis in the longitudinal direction of the vehicle, the y-axis in the transverse direction of the vehicle, and the z-axis in the vertical direction of the vehicle. The attitude deviation between the plane defined by the x-axis (X) and y-axis (Y) of the sensor coordinate system and the plane defined by the x-axis (X) and y-axis (Y) of the vehicle coordinate system is determined. During the driving operation of the vehicle (1), the attitude deviation is repeatedly determined in multiple iterative stages (IS1 to ISn), In each iteration (IS1 to ISn), the current measured values ​​(MW) of the longitudinal acceleration and lateral acceleration of the vehicle (1) determined by the inertial measurement unit (2), gravitational acceleration, state data of the vehicle (1) necessary for calculating inertial force, and coefficients of predetermined polynomials representing the approximate relationship between the measured longitudinal acceleration and the pitch angle (Φ) of the vehicle (1), or between the measured lateral acceleration and the roll angle (Θ) of the vehicle (1), are used as input values. The inertial force is calculated in each iteration stage (IS1 to ISn), the measured values ​​(MW) of the longitudinal acceleration and the lateral acceleration are adjusted for the contributing components due to the gravitational acceleration and the inertial force, the coefficients of the polynomial are updated with the adjusted measured values ​​of the longitudinal acceleration and the lateral acceleration, and it is confirmed whether the zero-order coefficients of the polynomial satisfy predetermined convergence criteria. If the zero-order coefficients of the polynomial do not satisfy the specified convergence criteria, the coefficients of the polynomial are output in the next iteration stage (IS2 to ISn). The iteration process is terminated when the zeroth coefficients of the polynomials satisfy the specified convergence criteria. A method wherein, at the end of the iterative process, the zero-order coefficient of the polynomial is output as a result of the determined attitude shift of the sensor coordinate system.

2. The method according to claim 1, wherein each of the iterative stages (IS1 to ISn) receives rotational speed, rotation rate, speed and / or wheel rotational speed as state data of the vehicle (1) necessary for calculating the inertial force.

3. The method according to claim 1 or 2, wherein each of the aforementioned iteration steps (IS1 to ISn) receives the coefficient of a predetermined linear or quadratic polynomial.

4. The method according to any one of claims 1 to 3, wherein the Coriolis force, centripetal force and / or Euler force are calculated as inertial forces.

5. The method according to any one of claims 1 to 4, wherein the determined attitude shift result of the sensor coordinate system is used to create a coordinate transformation matrix that transforms the three-dimensional component of the acceleration determined in the sensor coordinate system into the vehicle coordinate system.

6. Inertial measurement unit (2), and Apparatus designed and configured to perform the method described in any one of claims 1 to 5 A vehicle (1) equipped with the following:

7. The vehicle (1) according to claim 6, wherein the apparatus or at least one processing unit of the apparatus is a component of the inertial measuring unit (2).

8. The vehicle (1) according to claim 6 or 7, wherein the device has a sensor for determining state data of the vehicle (1) necessary for calculating the gravitational acceleration and / or the inertial force.

Images