DE102010023387A1 - Method for compensating sensing error of sensor, involves correcting output signal of rotor position detecting sensor by using correction signal - Google Patents

Abstract

The method involves correcting an output signal (S1,S2) of a rotor position detecting sensor (4) by using a correction signal (K1,K2). A parameter of a dynamically changing calculation formula is adjusted by a correction element (3). The correction element adjusts a parameter as a function of the correction signal. An independent claim is also included for an apparatus for compensating a sensing error of a sensor.

Description

The invention relates to a method and a device for compensating a sensor error, in particular for compensating for a systematic error of a rotor position sensor.

Electromechanical steering systems, in particular the so-called rack-concentric electric power steering (RC-EPS), have an electric servomotor, which is usually commuted electrically. The servo motor in this case brings a support torque on a steering column or preferably a rack of electromechanical steering. To regulate the assist torque u. a. a current rotor position of a rotor of the servomotor needed. In this case, such a rotor position is determined by means of a rotor position sensor or a so-called resolver.

Through various influences, eg. B. by temperature changes, contamination or design-related influences, the output signals of such a resolver to ideal, d. H. error-free, output signals falsified. As a result, a faulty rotor position of the rotor is determined and a regulation of the assist torque deteriorated qualitatively. The error in the output signal of the resolver is in this case of a systematic nature and thus reproducible for each individual electromechanical steering system.

The US 7,469,193 B2 discloses a method for compensating for an angle error in an output of a resolver. Initially, at least one output signal of the resolver is detected, wherein the resolver detects a current angular position of a rotor coupled to the resolver. After that, it is determined whether the current angular position of the rotor has been detected by a first pole pair of the resolver or by a second pole pair of the resolver. If the current angular position of the rotor has been detected by the first pole pair, then an output signal of the resolver is synchronized with a first resolver-dependent angular error information. If a current angular position of the rotor was detected by the second pole pair, the output signal of the resolver is synchronized with a second resolver-dependent angular error information. Here, the document discloses that a resolver-dependent angular error can be expressed in the form of a higher-order polynomial, so that a compensation information can be determined by calculation. Furthermore, the higher-order polynomial can be adapted to different speeds of the rotor.

The technical problem arises of providing a method and a device for compensating a sensor error, in particular a fault of a rotor position sensor of an electromechanical steering system, which enables an improved dynamic compensation of the sensor error.

The solution of the technical problem arises from the objects with the features of claims 1 and 9. Further advantageous embodiments of the invention will become apparent from the dependent claims.

Proposed is a method for compensating a sensor error of a sensor, wherein the sensor detects a rotor position of a rotor. In particular, the sensor can detect a rotor position of a rotor of a servomotor in an electromechanical steering system. In this case, the sensor can also be referred to as a so-called resolver. Furthermore, at least one output signal of the sensor is corrected by means of at least one correction signal. For example, the correction signal can be added to or subtracted from the at least one output signal of the sensor.

The at least one correction signal is determined as a function of at least one dynamically changeable calculation rule. In particular, parameters of the calculation rule can be changed during runtime of the method, ie, time-dependent. For example, the calculation rule can represent a mathematical model of the output signal of the sensor. The mathematical model can be z. B. consider physical measuring principles of the sensor for detecting the rotor position.

According to the invention, at least one parameter of the dynamically changeable calculation rule is set by means of a correction element, wherein the correction element adjusts the at least one parameter as a function of the at least one correction signal. As a result, a dynamic compensation of the sensor error is made possible in an advantageous manner, wherein the sensor error is a systematic error. In particular, sensor errors dependent on temperature changes, sensor errors dependent on an input voltage change of the sensor or sensor errors dependent on further changes can be compensated for at runtime with a sufficient compensation quality. In particular, the calculation rule or at least one parameter of the calculation rule is adapted for this purpose. Changes z. For example, if the correction signal is stronger than a predetermined amount within a predetermined period of time, it can be assumed that causes have occurred or changed that lead to a falsification of the output signal of the sensor. If the correction signal rises or falls z. B. within a predetermined period of time more than a predetermined percentage, so can causes for a systematic sensor errors have occurred or have changed.

In a further embodiment, the at least one correction signal is calculated as the difference between the at least one output signal of the sensor and at least one output signal of a unit for determining an estimated value, wherein the unit for determining an estimated value estimates the at least one output signal of the sensor by means of the dynamically changeable calculation rule. In this case, the value of the output signal is estimated by means of the unit for determining an estimated value at runtime of the method. In this case, the calculation rule can be, for example, a model-based calculation rule, the calculation rule being based on a dynamic model of the sensor. In this case, the calculation rule can estimate the estimated value as a function of a time course. The unit for determining an estimated value can also be referred to as a so-called disturbance observer. For example, the output signal of the sensor may have a sinusoidal profile over a running time if the rotor rotates. A dynamic model of the output signal of the sensor is then z. B. by means M1 = A · sin (ω · t + δ) Formula 1 given. Here, A denotes an amplitude, ω an angular frequency and δ a phase angle of a sine wave. The amplitude A, the angular frequency ω and the phase offset δ form parameters of the calculation rule. The angular frequency ω is dependent on a rotational speed of the rotor. In this case, at least one of the aforementioned parameters can be set as a function of the at least one correction signal, the correction signal K1 being a difference K1 = M1 - S1 Formula 2 calculated. Here K1 designates the correction signal, M1 the output signal calculated or estimated by means of the calculation rule, and S1 the output signal of the sensor. If all the parameters of the calculation rule are known at a first time t ₁ , the estimated value M1 can be calculated for this first time t ₁ . Similarly, the estimated value may be determined for M1 further time points t _1, for example, follow the time.

The calculation rule can be based, for example, on a so-called Kalman filter, in particular a Kalman non-linear filter (extended Kalman filter, unscented Kalman filter). By means of the Kalman filter, an estimated value M1 of the output signal S1 of the sensor is estimated. Depending on the difference between the estimated value M1 and the output signal S1, that is, the correction signal K1, the parameters of the calculation rule are adapted in this case. The model-based calculation of an estimated value and thus of a correction signal advantageously results in an improved dynamic adaptation of the compensation of a sensor error.

In a further embodiment, an input signal of the unit for determining a correction signal is the at least one output signal of the sensor. Another input to the unit for determining an estimate may be a current time.

In a further embodiment, a first sinusoidal output signal of the sensor is corrected by means of a first correction signal and a second sinusoidal output signal of the sensor by means of a second correction signal, wherein the first output signal leads the second output signal by 90 ° out of phase. If the first output signal is sinusoidal, it results in an advantageous manner that a corrected rotor position of the rotor can be calculated by means of an arctan function, wherein the argument of the arctan function is determined as the quotient of the first sensor signal and the second sensor signal. As a result, an unambiguous determination of the rotor position can be carried out in an advantageous manner.

For this purpose, for. B. known sensors or resolvers are used, the z. B. two offset by 90 ° spatially arranged windings, which generate the first and the second sensor signal in response to rotation of the rotor. Here, the first winding may be referred to as a so-called sine winding and the second winding as a so-called cosine winding. Ideally, the first sensor signal thus has a purely sinusoidal course and the second sensor signal has a purely cosinusoidal course.

In a further embodiment, a first output signal of the unit for determining an estimated value is determined as the X-coordinate of a point on an idealized circle and a second output signal of the unit for determining an estimated value as the Y-coordinate of the point on the idealized circle. In this case, the point on the idealized circle is determined as a function of a minimum geometric difference between a measuring point and the point of the idealized circle, wherein the measuring point has the first output signal of the sensor as X-coordinate and the second output signal of the sensor as Y-coordinate. Is the second output signal of the sensor (Y-value) above the first output signal of the sensor (X-value) in a Cartesian coordinate system with an X-axis and a Y-axis applied, results in an error-free detection of the rotor position of the idealized circle. However, if there are systematic errors in the detection of the rotor position, a geometric shape of a temporal course of the measuring points deviates from a shape deviating from the idealized circle, in particular an ellipsoidal shape. The size of the systematic error is dependent on a deviation or a geometric difference between the measuring point and a point of the idealized circle. The first and second correction signals are then determined as the difference between the first output signal of the sensor and the first output signal of the unit for determining an estimate or the difference between the second output signal of the sensor and the second output signal of the unit for determining an estimated value. The dependence of the determination of the first and second output signals of the unit for determining an estimated value of an idealized circuit advantageously results in a robust, easily implemented and reliable determination of the correction signal.

Here, the minimum geometric difference may be a minimum difference along an X-axis of the aforementioned Cartesian coordinate system or a minimum difference along a Y-axis or a smallest quadratic difference between the coordinates of the measurement point and the point of the idealized circle. The way of determining the minimum geometric difference determines which point of the idealized circle should correspond to the measuring point. For example, if it is known that only the first output signal of the sensor is subject to a systematic error, the first output signal of the unit for determining an estimate can be determined as an X coordinate of the point of the idealized circle whose geometric distance from the X coordinate of the measurement point is minimal is. Similarly, determining the second output of the estimated estimate unit is a Y-coordinate of the point of the idealized circle whose geometric distance from the Y-coordinate of the measurement point is minimal, if it is known that a systematic error occurs only in the second output of the sensor is present. If there is a systematic error for both output signals of the sensor, the difference is preferably determined as the smallest quadratic difference of the coordinates of the measuring point and the point of the idealized circle.

Here, the measuring point on the coordinates (S1, S2), wherein S1 denotes the first output signal of the sensor and S2, the second output signal of the sensor. The coordinates of the idealized circle are given by M1 = A · sin (ω · t + δ) Formula 3, M2 = A · cos (ω · t + δ) Formula 4.

In a further embodiment, the correction element may set the amplitude A and / or the phase offset δ and / or the angular frequency ω in the first and second coordinate equations.

Further proposed is a device for compensating a sensor error of a sensor, wherein by means of the sensor, a rotor position of a rotor can be detected, wherein at least one output signal of the sensor is correctable by means of a correction signal, wherein the at least one correction signal in response to at least one dynamically changeable calculation rule can be determined. According to the invention, the at least one parameter of the dynamically changeable calculation rule can be set by means of a correction element, the correction element setting the at least one parameter as a function of the at least one correction signal. By means of the proposed device, one of the previously explained methods is feasible.

The invention will be explained in more detail with reference to an embodiment. The figures show:

1 a schematic block diagram of a device for compensating a sensor error and

2 a schematic representation of a calculation of a first and second output of a unit for determining an estimate.

1 shows a schematic block diagram of a device 1 for compensation of a sensor error of a rotor position sensor 4 in an electromechanical steering system, not shown. In this case, the device comprises 1 one unity 2 for determining an estimated value and a correction term 3 , The rotor position sensor generates a first output signal S1 and a second output signal S2. A time profile of the first output signal S1 of the rotor position sensor 4 Here is a sinusoidal curve, wherein the second output signal S2 of the rotor position sensor 4 the first output signal S1 of the rotor position sensor 4 out of phase by 90 °, thus having a cosinusoidal course. The first and the second output signal S1, S2 of the rotor position sensor 4 form here input signals of the unit 2 to determine an estimate. Based on idealized coordinate equations defined in Formula 3 and Formula 4 of the general description, the unit determines 2 for determining an estimated value, a first output signal M1 and a second output signal M2 of the unit 2 to determine an estimate. Here, the first output signal M1 is an estimated value of the first output signal S1 of the rotor position sensor 4 , The second output M2 of the unit 2 for determining an estimated value is an estimated value of the second output signal S2 of the rotor position sensor 4 , A first correction signal K1 is calculated as the difference between the first output signal S1 of the rotor position sensor 4 and the first output M1 of the unit 2 calculated to determine an estimate. A second correction signal K2 is a difference of the second output signal S2 of the rotor position sensor 4 and the second output M2 of the unit 2 calculated to determine an estimate. A corrected first output signal S1 _K is then used as a difference between the first output signal S1 of the rotor position sensor 4 and the first correction signal K1 and a second corrected output signal S2 _{K of} the rotor position sensor 4 as the difference between the second output signal S2 of the rotor position sensor 4 and the second correction signal K2. Further, the first and the second correction signal K1, K2 serve as input variables for the correction element 3 , Depending on a time profile of the first and the second correction signal K1, K2, the correction element 3 an amplitude A, an angular frequency ω and a phase offset δ of the idealized coordinate equations (Formula 3, Formula 4).

In 2 is a schematic determination of the first output signal M1 and the second output signal M2 of the unit 2 to determine an estimate. In this case, a time course of the second output signal S2 of the rotor position sensor 4 above the first output signal S1 of the rotor position sensor 4 in a Cartesian coordinate system described by the X-axis and the Y-axis. If there is a systematic error, then the geometric course describes an ellipsoidal course 5 , An exemplary measuring point P _S has the coordinates S1, S2, wherein a value or an amplitude of the first output signal S1 of the rotor position sensor 4 as an X value and a value or an amplitude of the second output signal S2 of the rotor position sensor 4 is interpreted as a Y value in the Cartesian coordinate system. Next is an idealized circle 6 which shows a time profile of the first output signal S1 and the second output signal S2 of the rotor position sensor 4 in error-free case represents. The first output M1 of the unit 2 for the determination of an estimated value, this is called the X-coordinate of a point P _{M of} the idealized circle 6 certainly. Similarly, the second output M2 is calculated as the Y coordinate of the point P _M. In this case, the point P _{M of} the idealized circle has the smallest geometric distance to the measuring point P _s . It is assumed that only the first output signal S1 of the rotor position sensor 4 is defective, a point P1 _{M may} be a point of the idealized circle 6 determine that has the smallest distance in the X direction from the measuring point P _s . Analogously, a point P2 _{M of} the idealized circle 6 determine as a point having the least geometric distance in the Y direction from the measuring point P _S , in which case it is assumed that only the second output signal S2 of the rotor position sensor 4 is flawed.

LIST OF REFERENCE NUMBERS

1

contraption

2

Unit for determining an estimated value

3

correction term

4

Rotor position sensor

5

ellipse

6

idealized circle

S1

first output signal of the rotor position sensor

S2

second output signal of the rotor position sensor

M1

first output of the unit for determining an estimate

M2

second output of the unit for determining an estimate

K1

first correction signal

K2

second correction signal

S1 _K

first corrected output signal of the rotor position sensor

_S2K

second corrected output signal of the rotor position sensor

A

amplitude

ω

angular frequency

δ

phase displacement

P _S

measuring point

P _M

Point of the idealized circle

P1 _M

Point of the idealized circle

P2 _M

Point of the idealized circle

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• US 7469193 B2 [0004]

Claims (9)

Method for compensating a sensor error of a sensor, wherein the sensor ( 4 ) detects a rotor position of a rotor, wherein at least one output signal (S1, S2) of the sensor ( 4 ) is corrected by means of at least one correction signal (K1, K2), wherein the at least one correction signal (K1, K2) is determined as a function of at least one dynamically changeable calculation rule, characterized in that at least one parameter of the dynamically changeable calculation rule by means of a correction term ( 3 ), the correction term ( 3 ) sets the at least one parameter as a function of the at least one correction signal (K1, K2). A method according to claim 1, characterized in that the at least one correction signal (K1, K2) as a difference of the at least one output signal (S1, S2) of the sensor ( 4 ) and at least one output signal (M1, M2) of a unit ( 2 ) is calculated to determine an estimate, the unit ( 2 ) for determining an estimated value by means of the dynamic calculation rule, the at least one output signal of the sensor ( 4 ) appreciates. Method according to one of the preceding claims, characterized in that at least one input signal of the unit ( 2 ) for determining an estimated value, the at least one output signal (S1, S2) of the sensor ( 4 ). Method according to one of the preceding claims, characterized in that a first sinusoidal output signal (S1) of the sensor is corrected by means of a first correction signal (K1) and a second sinusoidal output signal (S2) of the sensor by means of a second correction signal (K2), wherein the second Output signal (S2) leads the first output signal (S1) with a phase offset of 90 degrees. Method according to claim 4, characterized in that a first output signal (M1) of the unit ( 2 ) for determining an estimated value as the x-coordinate of a point (P _M , P 1 _M , P 2 _M ) on an idealized circle ( 6 ) and a second output signal (M2) of the unit ( 2 ) For determining an estimated value (as the y-coordinate of the point P _{M, M} P1, P2 _M) (in the idealized circuit 6 ), where the point (P _M , P 1 _M , P 2 _M ) on the idealized circle ( 6 ) is determined as a function of a minimum geometric difference between a measuring point (P _S ) and the point (P _M , P 1 _M , P 2 _M ) of the idealized circle, the measuring point (P _S ) being the x-coordinate, the first output signal (S1) of the sensor ( 4 ) and as y-coordinate the second output signal (S2) of the sensor ( 4 ) having. A method according to claim 5, characterized in that the minimum geometric difference is a difference along an x-axis or a difference along a y-axis or a smallest quadratic difference between the coordinates of the measuring point (P _s ) of the point (P _M , P1 _M , P2 _M ) of the idealized circle ( 6 ). Method according to one of claims 5 or 6, characterized in that an x-coordinate M1 of an idealized circle ( 6 ) by means of a first coordinate equation of the form M1 = A · sin (ω · t + δ) and a y-coordinate M2 of the idealized circle ( 6 ) is calculated by means of a second coordinate equation of the form M2 = A * cos (ω * t + δ), the correction term ( 3 ) sets an amplitude (A) and / or a phase offset (δ) and / or an angular frequency (ω) of the first and second coordinate equations. Method according to one of the preceding claims, characterized in that the rotor position is the rotor position of a servomotor of an electromechanical steering. Device for compensating a sensor error of a sensor, wherein by means of the sensor ( 4 ), a rotor position of a rotor can be detected, wherein at least one output signal (S1, S2) of the sensor ( 4 ) can be corrected by means of a correction signal (K1, K2), wherein the at least one correction signal (K1, K2) can be determined as a function of at least one dynamically changeable calculation rule, characterized in that at least one parameter of the dynamically changeable calculation rule by means of a correction term ( 3 ) is adjustable, wherein the correction member ( 3 ) sets the at least one parameter as a function of the at least one correction signal (K1, K2).

Images