Method and detection system for detecting angular position

Abstract

The invention relates to a method (100) for detecting an angular position (α) of a rotating component rotatable about a rotation axis via a sensor unit (102), the sensor unit having a fixed sensor element and a rotating element rotatable relative to the sensor element and together with the rotating component, wherein the sensor element generates a first sensor signal (S) which is correlated with the angular position (α). 1，a ) and the second sensor signal (S 2，a ) are output to an evaluation unit (104), the first sensor signal being associated with a first detection position, the second sensor signal being associated with a second detection position perpendicular to the first detection position about the rotation axis, the evaluation unit calculating the angular position (α) in an evaluation step (118) via the atan2 function in relation to the first detection position and the second detection position, wherein in a processing step (110) preceding the evaluation step (118), the first sensor signal and the second sensor signal (S 1，a , S 2，a ) is analyzed by principal component analysis, and the first sensor signal and the second sensor signal (S 1，a , S 2，a The invention also relates to a detection system for detecting the angular position (α) of a rotating component using this method (100).

Description

Technical Field

[0001] The invention relates to a method for detecting an angular position of a rotating component rotatable about a rotation axis via a sensor unit. Furthermore, the invention relates to a detection system for detecting an angular position. Background Art

[0002] For example, a method for detecting the angular position of a rotating component is known from WO 2018 / 219388 A1. This document describes a method for detecting the angular position of a rotating component rotatable about an axis of rotation, wherein the angular position of the rotating component is received by a sensor system arranged radially spaced apart from the axis of rotation. A magnetic ring fixedly and concentrically arranged on the rotating component causes a magnetic field that changes relative to the sensor system, which is detected by the sensor system, wherein the signal received from the sensor system is evaluated with respect to the angular position. The amplitude information of the signal received from the sensor system with respect to the magnetic field is evaluated and a correction parameter is obtained from the amplitude information, and the angular error of the angular position received from the signal of the sensor system is determined with the aid of the correction parameter. The angular error is then used to correct the angular position obtained based on the signal output by the sensor system. Summary of the Invention

[0003] The object of the present invention is to detect the angular position of a rotating component more accurately and more quickly. The angular position of the rotating component should be determined with as little computational effort as possible. Furthermore, the angular position should be detected at a lower cost.

[0004] At least one of these objects is achieved by a method for detecting the angular position of a rotating component that can rotate about a rotation axis via a sensor unit, the sensor unit having a fixed sensor element and a rotating element that can rotate relative to the sensor element and together with the rotating component, wherein the sensor element outputs a first sensor signal and a second sensor signal that are related to the angular position to an evaluation unit, the first sensor signal being associated with a first detection position and the second sensor signal being associated with a second detection position that is perpendicular to the first detection position about the rotation axis, the evaluation unit calculating the angular position in an evaluation step via the atan2 function as a function of the first detection position and the second detection position, wherein in a processing step preceding the evaluation step, possible amplitude errors and possible orthogonal errors of the first sensor signal and the second sensor signal are converted as a total error into an amplitude difference of the first sensor signal and the second sensor signal via a principal component analysis of the first sensor signal and the second sensor signal.

[0005] This makes it possible to more accurately detect the angular position of the rotating member and further reduce angular errors.

[0006] The sensor unit and the rotating member may be disposed in a vehicle. The rotating member may be associated with a parking lock device. The rotating member may be disposed on an actuator for adjusting the parking lock device or for operating a clutch of the vehicle. The rotating member and the rotating element may be disposed concentrically and rotatably.

[0007] The sensor unit can be designed as an angle sensor. The sensor unit can be arranged radially spaced apart from the rotating component. The sensor unit can be arranged axially spaced apart from the rotating component.

[0008] The sensor element can be a Hall sensor. The rotating element can be a magnetic ring. The rotating element can be a permanent magnet. The rotating element can be magnetized in opposite directions.

[0009] The first sensor signal and / or the second sensor signal can be a periodic signal, in particular a periodic signal caused by a rotation of the rotating element. The first sensor signal can be a cosine signal and the second sensor signal can be a sine signal.

[0010] The first sensor signal and the second sensor signal can have a possible amplitude error and a possible quadrature error. The first sensor signal detecting the angular position α and the second sensor signal The amplitude error and quadrature error in can be described as follows:

[0011] (1)

[0012] in , are the corresponding amplitudes and φ is the quadrature error. The quadrature error φ describes the first and second sensor signals , The first sensor signal and the second sensor signal should theoretically extend perpendicularly to each other due to the vertical orientation of the first and second detection positions. , The deviation in magnitude between them.

[0013] It has been found that the amplitude error and the quadrature error φ are of the same type and can be compensated more effectively than before. The method used for this will be described in detail below:

[0014] The first sensor signal output by the sensor unit and the second sensor signal Before being forwarded to the evaluation step, it can first be processed in a preparation step. In this case, the first and second sensor signals can be compensated. , The superimposed offset error is removed and the first pre-processed sensor signal corrected in this way can be output. and the second preprocessed sensor signal thus corrected For example, the offset error can be determined by applying the maximum-minimum method. Here, during at least one revolution of the rotating component, the maximum and minimum sensor values ​​of the corresponding sensor signal can be detected, and then the offset of the corresponding sensor signal can be determined. The calculated offset can then be compensated.

[0015] First and second preprocessed sensor signals , The processing step can be input to a processing step following the preparation step. The processing step can include multiple individual steps. The processing step can include an analysis step, a conversion step, and an adjustment step. The analysis step and the conversion step can end before the adjustment step. The analysis step can be performed before the conversion step.

[0016] The first and second pre-processed sensor signals can be processed in an analysis step , In the analysis step, a coordinate rotation can be assumed which rewrites the first and second preprocessed sensor signals according to the following relationship , :

[0017] (2).

[0018] Here, A and B are the coefficients of the rotation matrix, and is the amplitude of the rewritten sensor signal and is the phase shift. It is found that the first and second pre-processed sensor signals , Possible amplitude errors and possible quadrature errors of can be converted together into an amplitude error of the rewritten sensor signal. The rewritten sensor signal is thus converted relative to the first and second preprocessed sensor signals. , With phase shift .

[0019] Ability to calculate parameters and magnitudes of rotation matrices during analysis step and For this purpose, a principal component analysis can be used. This in turn can be performed by means of a singular value decomposition or a principal axis transformation. In the following, the principal axis transformation will be used, which requires particularly little computing power.

[0020] The rotation matrix is ​​given by the correlation matrix The correlation matrix is ​​calculated as follows:

[0021] (3)

[0022] Where n is the number of measurement points. Then the correlation matrix is ​​obtained The eigenvalues ​​λ1 and λ2 and the associated eigenvectors and , whereby the rotation matrix is ​​described as follows:

[0023] .

[0024] The amplitude from (2) and It can then be calculated via the characteristic values ​​λ1 and λ2 according to the following relationship, where λ1>λ2:

[0025] (4)

[0026] (5)

[0027] From this, the phase shift can be calculated using (2) as follows :

[0028] (6)

[0029] In a subsequent transformation step, the first preprocessed sensor signal and the second preprocessed sensor signal can be rotated using a rotation matrix, whereby the first preprocessed sensor signal is transformed into Converted into a first converted sensor signal And the second pre-processed sensor signal Converted into a second converted sensor signal :

[0030] (7)

[0031] In a subsequent adjustment step, the first converted sensor signal and the second converted sensor signal The amplitude from (4) can be used and the magnitude from (5) Normalization. In particular, the first converted sensor signal Can be multiplied by the first analysis parameter And the first converted sensor signal Ability to multiply by a second analysis parameter And thus normalized.

[0032] The first processed sensor signal thus output and a second processed sensor signal can be forwarded to a subsequent evaluation step which calculates the output angular position from the two sensor signals by applying the atan2 function .

[0033] The angular position output by the evaluation step is changed via the correction step , in such a way that, from the output angular position The constant phase shift as the third analysis parameter calculated in the analysis step is subtracted from And output the angular position processed in this way .

[0034] In a preferred embodiment of the present invention, the processing step advantageously includes an analysis step that performs a principal component analysis of the first and second sensor signals and thereby determines at least one analysis parameter, which is used in at least one subsequent processing step to modify the first and / or second sensor signal. Consequently, the first and second sensor signals can subsequently be modified in accordance with the at least one analysis parameter to reduce the angular error.

[0035] In a specific embodiment of the present invention, it is advantageous if the principal component analysis uses a principal axis transformation with a rotation matrix. This allows the analysis parameters to be calculated with less computing power. The analysis parameters can be directly related to at least one eigenvalue of the principal axis transformation.

[0036] In a preferred embodiment of the present invention, the processing step includes a conversion step that converts the first and second sensor signals into corresponding converted first and second sensor signals by performing a coordinate rotation using a rotation matrix. This allows any amplitude errors and any quadrature errors in the first and second sensor signals to be converted into a common amplitude error for the first and second converted sensor signals. In addition to the common amplitude error, the first and second converted sensor signals can also have a phase shift relative to the first and second sensor signals.

[0037] In a preferred embodiment of the present invention, in a conditioning step following the conversion step, the first converted sensor signal is normalized using the first analysis parameter determined in the analysis step, and the second converted sensor signal is normalized using the second analysis parameter determined in the analysis step, and these are output as first and second processed sensor signals. This allows for compensation of a common amplitude error between the first and second converted sensor signals.

[0038] In a specific embodiment of the present invention, it is advantageous if the first and second analysis parameters are related to eigenvalues ​​of the principal axis transformation. The first analysis parameter can be related only to the first eigenvalue determined in the principal axis transformation. The second analysis parameter can be related only to the second eigenvalue determined in the principal axis transformation.

[0039] In one advantageous embodiment of the present invention, the adjustment step is followed by an evaluation step, which calculates the angular position as the output angular position based on the first and second processed sensor signals by applying the atan2 function. The atan2 function is an extension of the inverse angular function arctangent and, as such, the inverse of the angular function tangent. The atan2 function takes two real numbers as arguments, in contrast to the normal inverse tangent function, which only has one real number as an argument. As a result, the atan2 function has sufficient information to output function values ​​across a 360° value range (i.e., all four quadrants) and does not need to be restricted to two quadrants (as is the case with the normal inverse tangent function).

[0040] In a preferred embodiment of the present invention, the analysis step is followed by a correction step, in which the angular position is corrected using the third analysis parameter determined in the analysis step. The first, second, and / or third analysis parameter can be stored in a retrievable memory after calculation in the analysis step.

[0041] A preferred embodiment of the invention is advantageous in which the third evaluation parameter corresponds to a phase shift of the respective converted sensor signal relative to the respective preprocessed sensor signal. The output angular position can be output as a processed angular position corrected by the phase shift.

[0042] Furthermore, at least one of the aforementioned objects is achieved by a detection system for detecting the angular position of a rotating component by means of a method having at least one of the aforementioned features, the detection system having an evaluation unit and a sensor unit having a fixed sensor element and a rotating element rotatable relative to the sensor element and together with the rotating component.

[0043] Further advantages and advantageous embodiments of the invention are apparent from the description and the figures of the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The present invention will be described in detail below with reference to the accompanying drawings.

[0045] Figure 1 A flow chart illustrating a method in a specific embodiment of the present invention.

[0046] Figure 2 Shown in accordance with Figure 1 Corresponding graphs of the sensor signal after the individual processing steps of the method.

[0047] Figure 3 Shown in accordance with Figure 1 The corresponding angular error curves of the sensor signal after the individual processing steps of the method.

[0048] Figure 4 The curve showing the change in calculation accuracy as a function of the number of measuring points is shown.

[0049] Figure 5 The curve of the maximum angular error as a function of the signal quality in a comparison between a plurality of methods is shown. DETAILED DESCRIPTION

[0050] Figure 1 A flow chart of a method 100 in a specific embodiment of the present invention is shown. The method 100 detects the angular position of a rotating component rotatable about a rotation axis via a sensor unit 102. , the sensor unit will be with the angular position Related first sensor signal and the second sensor signal The first sensor signal is output to the evaluation unit 104 , being associated with a first detection position, and the second sensor signal is associated with a second detection position perpendicular to the first detection position about the rotation axis.

[0051] The evaluation unit 104 processes the first and second sensor signals from the , Calculating angular position .exist Figure 2 The corresponding graphs of the sensor signals output after the various processing steps of the method 100 in xy coordinates are depicted in FIG. Figure 3 The angular error of the sensor signal output after the corresponding processing steps of the method 100 is depicted in FIG. Unless otherwise specified, the following descriptions refer to the corresponding angular error variation curve. Figure 1 Method 100, however some places explicitly refer to Figure 2 and Figure 3 .

[0052] The first and second sensor signals , Measuring angular position When Figure 3 The angular error depicted in Can occur via the angular position α. ​​Angular error corresponds to a measurement inaccuracy and, if not taken into account, would result in an inaccurate measured angular position α. This can be taken into account and compensated for by subsequent further processing steps of method 100 .

[0053] First sensor signal and the second sensor signal The output from the sensor unit 102 is sent to the evaluation unit 104. The evaluation unit 104 has a preparation step 106 in which the first and second sensor signals are compared with each other. , The superimposed offset error is compensated and the thus corrected first and second preprocessed sensor signals are output , For example, the offset error can be determined by applying the maximum-minimum method 108. In this case, the corresponding sensor signal can be detected during at least one revolution of the rotating element. , The maximum and minimum sensor values ​​and the corresponding sensor signals can then be determined , Subsequently, it is possible to compensate for the offset between the first and second sensor signals , This calculated offset. Figure 2 The corresponding prepared sensor signal in is the zero-centered ellipse after the preparation steps. Figure 3 The angular error contained in the first and second preprocessed sensor signals is shown in The relevant change curve.

[0054] First and second preprocessed sensor signals , It is then fed into a processing step 110 following the preparation step 106. In the processing step 110, a plurality of individual processing steps can be carried out, here an analysis step 112, a conversion step 114 and a conditioning step 116. The first and second pre-processed sensor signals , First, it is processed in an analysis step 112. In the analysis step 112, it is assumed that based on the first and second pre-processed sensor signals , The coordinate rotation of (2) is calculated via the rotation matrix. Here, it has been found that the first and second preprocessed sensor signals , The possible amplitude error and the possible quadrature error of can be jointly converted into an amplitude error of the sensor signal obtained via the coordinate rotation.

[0055] In the analysis step 112, the rotation matrix is ​​calculated by applying the principal component analysis according to (3). This is in turn performed using the principal axis transformation, which is used to determine the eigenvalues ​​of the correlation matrix and the associated eigenvectors, based on which the rotation matrix is ​​constructed. In the analysis step 112, the first analysis parameter associated with the first eigenvalue is calculated. and the second analysis parameter associated with the second eigenvalue , and store it for later access.

[0056] The third analysis parameter is obtained by the phase shift according to (6) form.

[0057] In a conversion step 114 following the analysis step 112, the first and second pre-processed sensor signals , is rotated by the rotation matrix according to (2), whereby according to (7), the first preprocessed sensor signal is converted into a first converted sensor signal And the second pre-processed sensor signal is converted into a second converted sensor signal In this first and second converted sensor signals , The first and second sensor signals , The possible amplitude error and orthogonality error in are converted into amplitude difference as the total error.

[0058] exist Figure 2 In the example, the converted sensor signal From the prepared sensor signal , whereby the major axis of the ellipse lies on the axis of the coordinate system defined by the zero point. Figure 3 The relevant angular error F in r Shift by coordinate rotation.

[0059] In the subsequent adjustment step 116 , the first converted sensor signal Use the first analysis parameter Normalized, where calculated according to (4) , and the second converted sensor signal Use the second analysis parameter Normalized, where calculated according to (5) . Thus, the first and second converted sensor signals are compensated , and the amplitude difference between them as the first and second processed sensor signals, respectively. , Output.

[0060] Figure 2 The processed sensor signal S in C It is a circle with its center at zero. Figure 3 The corresponding angular error F C and angular position Move forward irrelevantly.

[0061] The first and second processed sensor signals thus output , It is then forwarded to a subsequent evaluation step 118, which evaluates the first and second processed sensor signals by applying the atan2 function. , The evaluation step 118 uses the calculated angular position as the output angular position Output to the correction step 120.

[0062] In the correction step 120, the output angular position is changed , in such a way that, from the output angular position The phase shift α0 as the third analysis parameter calculated in the analysis step 112 is subtracted from the angular position processed in this way and outputted. .

[0063] Therefore, compensation Figure 3 The angular error calculated in And can accurately calculate the angular position of the rotating component .

[0064] Figure 4 Shows calculation accuracy The number of measurement points In order to use the computing power as efficiently as possible when determining the angular position by applying the above-described principal component analysis, it is necessary to determine how many measuring points are required in order to be able to carry out the method cost-effectively and quickly.

[0065] If the calculation accuracy is preset, the required number of measuring points can be determined The number of measurement points , calculation accuracy and measurement resolution The relationship between them is:

[0066] (8)

[0067] Here, and Greater than zero.

[0068] For example, if according to (8), at a resolution of, for example, 0.5° In the case of 1° calculation accuracy is required , then there should be at least 58 measurement points in the corresponding sensor signal.

[0069] Figure 5 Shows the largest angular error in the comparison between multiple methods and signal quality The relevant change curve of the maximum angular error The signal quality is related to the signal-to-noise ratio. When the maximum-minimum method Ma is applied, the maximum angular error in detecting the angular position is The highest, where the orthogonality error is not compensated. In the case of the method Ms according to the prior art, the maximum angular error Compared to this, the method Mc described in DE 10 2020102 064.3 is designed to further reduce the maximum angular error Furthermore, the computing power is significantly reduced here compared to the method Ms according to the prior art.

[0070] About the maximum angular error In a particular embodiment of the present invention, the proposed method Me is the best. In the case of can be reduced, for example, to about 50% compared to the Mc method.

[0071] Reference Signs List

[0072] 100 methods

[0073] 102 sensor unit

[0074] 104 evaluation units

[0075] 106 Preparatory Steps

[0076] 108 Max-Min method

[0077] 110 Processing Steps

[0078] 112 Analysis Steps

[0079] 114 Conversion Steps

[0080] 116 Adjustment Steps

[0081] 118 Evaluation Steps

[0082] 120 Calibration Steps

[0083] α Angular position

[0084] α c The output angular position

[0085] α 0 Phase Shift

[0086] F Angular error

[0087] F p Angular error

[0088] F r Angular error

[0089] F c Angular error

[0090] F s Angular error

[0091] F max Maximum angular error

[0092] λ1 first eigenvalue

[0093] λ2 second eigenvalue

[0094] S 1,a First sensor signal

[0095] S 2,a Second sensor signal

[0096] S 1,p First ready sensor signal

[0097] S 2,p Second prepared sensor signal

[0098] S p Prepared sensor signal

[0099] S 1,r First converted sensor signal

[0100] S 2,r Second converted sensor signal

[0101] S r Converted sensor signal

[0102] S 1,c First processed sensor signal

[0103] S 2,c Second processed sensor signal

[0104] S c Processed sensor signal

[0105] 1 / a 1. First analysis parameter

[0106] 1 / b 1 Second analysis parameter

Claims

1. A method for detecting an angular position of a rotating member rotatable about a rotation axis via a sensor unit (102) ) method (100), the sensor unit having: fixed sensor element, and a rotating element rotatable relative to the sensor element and together with the rotating member, wherein The sensor element will be aligned with the angular position ( ) related to the first sensor signal ( ) and the second sensor signal ( ) are output to an evaluation unit (104), the first sensor signal being associated with a first detection position, the second sensor signal being associated with a second detection position perpendicular to the first detection position around the rotation axis, The evaluation unit calculates the angular position ( ) in an evaluation step ( 118 ) via the atan2 function in relation to the first and second detection positions. ), It is characterized by: In a processing step (110) preceding the evaluation step (118), the first sensor signal and the second sensor signal ( , ) of the principal component analysis, the first sensor signal and the second sensor signal ( , ) is converted into an amplitude difference between the first sensor signal and the second sensor signal as a total error.

2. The method for detecting an angular position according to claim 1 ) method (100), characterized in that, The processing step (110) comprises an analyzing step (112) for analyzing the first sensor signal and the second sensor signal ( , ) performs a principal component analysis and thereby determines at least one analysis parameter, which is used in at least one subsequent processing step to modify the first sensor signal and / or the second sensor signal ( , ).

3. The method for detecting an angular position according to claim 2 ) method (100), characterized in that, The principal component analysis applies a principal axis transformation using a rotation matrix.

4. The method for detecting an angular position according to claim 3 ) method (100), characterized in that, The processing step (110) has a conversion step (114) of converting the first sensor signal and the second sensor signal ( , ) into corresponding converted first and second sensor signals ( , ).

5. The method for detecting an angular position according to claim 4 ) method (100), characterized in that, In a conditioning step (116) following the conversion step (114), the first converted sensor signal ( ) and the second converted sensor signal ( ) are normalized and used as the first and second processed sensor signals ( , ) output.

6. The method for detecting an angular position according to claim 5 ) method (100), characterized in that, The first analysis parameter and the second analysis parameter are related to the eigenvalue of the principal axis transformation ( , ) related.

7. The method for detecting an angular position according to claim 6 ) method (100), characterized in that, The adjusting step (116) is followed by an evaluating step (118) which is based on the first processed sensor signal and the second processed sensor signal ( , ) calculates the angular position by applying the atan2 function as the output angular position ( ).

8. The method for detecting an angular position according to any one of claims 2 to 7. ) method (100), characterized in that, The evaluation step (118) is followed by a correction step (120) in which the angular position ( ) by the third analysis parameter ( ) is corrected.

9. The method for detecting an angular position according to claim 8 ) method (100), characterized in that, The third analysis parameter ( ) corresponds to the converted sensor signal ( , ) relative to the corresponding pre-processed sensor signal ( , ) phase shift.

10. A detection system for detecting the angular position of a rotating component by a method (100) according to any one of the preceding claims. ) and has an evaluation unit (104) and a sensor unit (102), the sensor unit having a fixed sensor element and a rotating element rotatable relative to the sensor element and together with the rotating component.

Images