DE102018209038A1 - METHOD AND CONTROL DEVICE FOR AUTOMATED LEARNING OF A STEERING WHEEL ANGLE SET IN THE OPERATION OF A VEHICLE - Google Patents

Abstract

The present invention relates to a method for automated learning of a steering wheel angle offset in the operation of a vehicle with measuring a current steering wheel angle of the vehicle with a steering wheel angle sensor; Detecting a current driving situation of the vehicle on a lane for the current steering wheel angle by a camera of the vehicle, wherein the current driving situation is detected in the form of camera data; Calculating an actual self-motion of the vehicle from the camera data for the current driving situation of the vehicle; Determining a modeled intrinsic motion of the vehicle based on the current steering wheel angle and a lateral dynamics vehicle model of the vehicle, the lateral dynamics vehicle model comprising a steering wheel angle offset state quantity as a degree of freedom; Comparing the actual proper motion of the vehicle with the modeled self-motion of the vehicle, determining a self-motion difference between the actual self-motion of the vehicle and the modeled self-motion of the vehicle; and returning the determined self-motion difference to the vehicle model in a closed loop such that the steering wheel angle offset state quantity converges to the steering wheel angle offset.

Description

The invention relates to a method for automated learning of a Lenkradwinkeloffsets in the operation of a vehicle. Furthermore, the invention relates to a control device for carrying out such a method.

Today's motor vehicles typically have a steering wheel angle sensor for measuring a steering wheel angle of a steering wheel, which may for example be formed directly on the steering wheel and / or on a lower end of a steering column of a power steering. For example, the measured steering wheel angle may be used in ESP systems to calculate a reference yaw rate of the vehicle to compare it to an actual yaw rate of the vehicle. For example, it can be determined whether a vehicle is under- or overpriced. This can in turn be used in assistance systems and / or automated driving to guide the vehicle in a lane and / or on a given ideal trajectory. However, if this measured steering wheel angle is subject to an offset, a lateral deviation of the vehicle relative to the orientation set via the steering wheel can generally result without additional counter-control.

Modern motor vehicles often include one or more cameras that may be configured to detect the current driving situation. Camera data recorded by the camera can be used to determine an actual proper motion of the vehicle relative to a lane.

Against this background, the present invention has the object to find automated solutions for determining a Lenkradwinkeloffsets.

According to the invention, this object is achieved by a method having the features of patent claim 1 and by a regulating device having the features of patent claim 8.

Accordingly, a method for automated learning of a steering wheel angle offset is provided. The method includes measuring a current steering wheel angle of the vehicle with a steering wheel angle sensor; Detecting a current driving situation of the vehicle on a lane for the current steering wheel angle by a camera of the vehicle, wherein the current driving situation is detected in the form of camera data; Calculating an actual self-motion of the vehicle from the camera data for the current driving situation of the vehicle; Determining a modeled intrinsic motion of the vehicle based on the current steering wheel angle and a lateral dynamics vehicle model of the vehicle, the lateral dynamics vehicle model comprising a steering wheel angle offset state quantity as a degree of freedom; Comparing the actual proper motion of the vehicle with the modeled self-motion of the vehicle, determining a self-motion difference between the actual self-motion of the vehicle and the modeled self-motion of the vehicle; and returning the determined self-motion difference to the vehicle model in a closed loop such that the steering wheel angle offset state quantity converges to the steering wheel angle offset.

Furthermore, a control device for carrying out a method according to the invention is provided.

An idea on which the invention is based is to simulate the steering system of a vehicle via a model and to continuously adapt the model to the actual conditions via a control loop. The vehicle model uses the measured and thus potentially offset steering wheel angle as input. Such an offset of the steering wheel angle is not directly measurable easily and is for this reason as a degree of freedom in the model. If the model is compared with the actually traveled lane of the vehicle, the steering wheel angle offset can be reconstructed from a deviation of the model from the measured ratios. In this sense, the system learns the steering wheel angle offset to a certain extent independently. In this case, the learning speed can be specified via the specific design of the control loop. Depending on a specific application, it is thus possible to specify how quickly the steering wheel angle offset state variable converges to the actual steering wheel angle offset.

Assume in one example that a steering wheel angle offset of 2 ° exists. Assuming a static steering ratio of, for example, 18, this would result in an offset of the effective steering angle of the vehicle of about 2 ° / 18 = 0.11 °. A lateral deviation controller of an assistance system or the like would thus cause a constant lateral deviation of, for example, a few centimeters when controlling the steering wheel angle without knowing this offset. This is prevented by the present invention.

Advantageous embodiments and further developments will become apparent from the other dependent claims and from the description with reference to the figures.

According to a development, the eigenmotion difference can be traced back via a state observer. The determined by the camera and driven by the vehicle path, ie the curvature or the driven yaw rate is thus compared via an observer structure against the vehicle model for lateral dynamics. In particular, such an observer dynamics can be used to set how fast the steering wheel angle offset is learned, for example by specifying the poles of a corresponding transmission system. The learning dynamics of the observation structure can be designed, for example, as a function of the driving speed of the vehicle.

According to a development, the transverse dynamics vehicle model may be based on a single-track model. The lateral dynamics vehicle model may include as state quantities a float angle of the vehicle, a yaw rate of the vehicle, and the steering wheel angle offset state quantity. Thus, the usual state variables of the one-track model, namely the slip angle and the yaw rate, are supplemented in this development by the steering wheel angle offset as a further state variable.

According to a development, the transverse dynamics vehicle model may include the current steering wheel angle as an input variable.

According to a further development, the transverse dynamics vehicle model may include a driven curvature of the vehicle or a driven yaw rate of the vehicle as an output variable.

According to a further development, the actual intrinsic movement of the vehicle can be calculated on the basis of a lane curvature of a lane of the vehicle on the lane, a lateral deviation of the vehicle from the lane and an orientation deviation of the vehicle from the lane. An actually driven curvature of the path of the vehicle or an actually driven yaw rate of the vehicle results here from a proportion in the event that the vehicle runs exactly parallel to the lane, and a proportion for the orientation deviation of the actual vehicle lane from the lane, i. from the direction of travel in the lane.

According to a development, the method can be applied to a longitudinal acceleration of the vehicle and / or a lateral acceleration of the vehicle less than 3 m / s ² . The method can thus be limited to areas in which the vehicle model used is valid, ie situations with relatively low longitudinal and / or transverse dynamics. Furthermore, the method may be limited to stationary situations (ie, steady cornering) in which a change in yaw rate and / or steering wheel angle is small.

The above embodiments and developments can, if appropriate, combine with each other as desired. Further possible refinements, developments and implementations of the invention also include combinations, not explicitly mentioned, of features of the invention described above or below with regard to the exemplary embodiments. In particular, the person skilled in the art will also add individual aspects as improvements or additions to the respective basic form of the present invention.

• 1 schematische Ansicht einer Regelvorrichtung gemäß einer Ausführungsform der Erfindung;
• 2 schematisches Ablaufdiagramm eines Verfahrens für die Regelvorrichtung aus 1; und
• 3 schematische Seitenansicht eines Fahrzeugs mit der Regelvorrichtung aus 1.

The present invention will be explained in more detail with reference to the exemplary embodiments given in the schematic figures. It shows:

• 1 schematic view of a control device according to an embodiment of the invention;
• 2 schematic flow diagram of a method for the control device 1 ; and
• 3 schematic side view of a vehicle with the control device off 1 ,

The accompanying figures are intended to convey a further understanding of the embodiments of the invention. They illustrate embodiments and, together with the description, serve to explain principles and concepts of the invention. Other embodiments and many of the stated advantages will become apparent with reference to the drawings. The elements of the drawings are not necessarily shown to scale to each other.

In the figures of the drawing are the same, functionally identical and same-acting elements, features and components - unless otherwise stated - each provided with the same reference numerals.

1 shows a schematic view of a control device 10 according to an embodiment of the invention. The control device is in a vehicle 1 , in particular a motor vehicle, integrated as it is in 3 is shown by way of example. 2 shows a schematic flow diagram for a method M the control device 10 for automatically learning a steering wheel angle offset during operation of the vehicle 1 ,

The procedure M includes under M1 Measuring a current steering wheel angle δ of the vehicle 1 with a steering wheel angle sensor 5 , which for example on a steering wheel and / or a Steering column of the vehicle 1 may be appropriate (not shown). The procedure M includes under M2 Further detecting a current driving situation of the vehicle 1 on a roadway 4 for the current steering wheel angle δ through a camera 2 of the vehicle 1 , The camera 2 For example, on a front of the vehicle 1 be integrated. The camera 2 is adapted to the current driving situation in the form of camera data 7 and capture this to the control device 10 to convey. The control device 10 is this with both the camera 2 as well as with the steering wheel angle sensor 5 communicatively connected.

und damit: $${\kappa }_{FahrzeugSpur}=\frac{{\kappa }_S}{1-\Delta y\cdot {\kappa }_S}$$

Under M3 includes the method M Calculate an actual proper motion 8a of the vehicle 1 from the camera data 7 for the current driving situation of the vehicle 1 , The actual proper movement 8a of the vehicle 1 can for example by an actually driven yaw rate ψ̇ _F or a curvature actually driven κ _f with the relation κ _F = ψ̇ _F / v (with the driving speed v of the vehicle 1 ). The camera 2 or the control device 10 The lane curvature κ _S = 1 / R _{S of} the lane on the road 4 determine. In addition, the camera can 2 Information about a page deviation Dy of the vehicle 1 provide, for example, a lateral deviation between a center of a right and a left boundary line of the lane and a center of an axis of the vehicle 1 (eg a rear axle). If the vehicle runs exactly parallel to the lane, the following results: $$R_{FaHrzeuGSpur}=R_S-\Delta y$$

and thus: $${\kappa }_{FaHrzeuGSpur}=\frac{{\kappa }_S}{1-\Delta y\cdot {\kappa }_S}$$

The amount of curvature that the vehicle travels within the lane becomes overshoot Δψ of the vehicle 1 determined from the lane, ie an angular deviation from a tangent to a virtual center between a left and a right boundary line of the lane: $${\kappa }_{FaHrzeuGinSpur}=\frac{d\left(\Delta \Psi \left(t\right)\right)}{dt}\cdot \frac{1}{v}$$

This results in the actually driven curvature κ _F : $${\kappa }_F={\kappa }_{FaHrzeuGSpur}+{\kappa }_{FaHrzeuGinSpur}$$

The measured yaw rate ψ̇ can for example be previously filtered (eg via a PT1 filter), wherein the filter may have a time constant that depends on the driving speed v of the vehicle 1 depends.

The procedure M under M4 further comprises comparing a modeled eigenmovement 8b of the vehicle 1 based on the current steering wheel angle δ and a cross-dynamics vehicle model 3 of the vehicle 1 , The transverse dynamics vehicle model 3 is based on a single-track model, which is the current steering wheel angle δ as an input, the curvature driven κ _f of the vehicle 1 or the driven yaw rate ψ̇ _{F of} the vehicle 1 as output variable and three state variables, namely a float angle β _Z of the vehicle 1 , a yaw rate state quantity ψ̇ _{Z of} the vehicle 1 and a steering wheel angle offset state quantity Δδ _Z ,

$$B\text{'}=\left[\begin{array}{c}B\\ 0\end{array}\right]$$

$$C\text{'}=\left[\begin{array}{cc}C& 0\end{array}\right]$$

und $$B=\left[\begin{array}{c}\frac{cf}{m\cdot v\left(t\right)}\\ \frac{lf\cdot cf}{J}\end{array}\right]$$

$$A=\left[\begin{array}{cc}-\frac{cf+cr}{m\cdot v\left(t\right)}& -\frac{lf\cdot cf+lr\cdot cr}{m\cdot v{\left(t\right)}^2}-1\\ -\frac{lf\cdot cf-lr\cdot cr}{J}& -\frac{l{f}^2\cdot cf+l{r}^2\cdot cr}{J\cdot v\left(t\right)}\end{array}\right]$$

$$C=\left[\begin{array}{cc}0& 1\end{array}\right]$$

wobei lf einen Abstand zwischen einer Vorderachse und einem Schwerpunkt des Fahrzeugs 1, lr einen Abstand zwischen einer Hinterachse und einem Schwerpunkt des Fahrzeugs 1, cf eine Schräglaufsteifigkeit der Vorderachse, cr eine Schräglaufsteifigkeit der Hinterachse, J ein Gierträgheitsmoment des Fahrzeugs 1 und m eine Masse des Fahrzeugs 1 angeben, wobei weiterhin die gewählte Ausgangsmatrix C die Gierrate ψ̇_F des Fahrzeugs 1 ausgibt (die Fahrgeschwindigkeit v des Fahrzeugs wird hier als Funktion der Zeit t aufgefasst).The input vector B ' , the system matrix A ' and the output vector C ' of the lateral dynamics vehicle model 3 now result from the input vector as follows B , the system matrix A and the output vector C of the single-track model: $$A\text{'}=\left[\begin{array}{c}\begin{array}{cc}A& B\end{array}\\ \begin{array}{ccc}0& 0& 0\end{array}\end{array}\right]$$

$$B\text{'}=\left[\begin{array}{c}B\\ 0\end{array}\right]$$

$$C\text{'}=\left[\begin{array}{cc}C& 0\end{array}\right]$$

and $$B=\left[\begin{array}{c}\frac{cf}{m\cdot v\left(t\right)}\\ \frac{lf\cdot cf}{J}\end{array}\right]$$

$$A=\left[\begin{array}{cc}-\frac{cf+cr}{m\cdot v\left(t\right)}& -\frac{lf\cdot cf+lr\cdot cr}{m\cdot v{\left(t\right)}^2}-1\\ -\frac{lf\cdot cf-lr\cdot cr}{J}& -\frac{\mathrm{<figure-callout\; id="2"\; label="l\; f"\; filenames="DE102018209038A1\_0013.png"\; state="\{\{state\}\}">l\; </figure-callout>}{f}^{}{}^2\cdot cf+\mathrm{<figure-callout\; id="2"\; label="l\; r"\; filenames="DE102018209038A1\_0013.png"\; state="\{\{state\}\}">l\; </figure-callout>}{r}^{}{}^2\cdot cr}{J\cdot v\left(t\right)}\end{array}\right]$$

$$C=\left[\begin{array}{cc}0& 1\end{array}\right]$$

where lf is a distance between a front axle and a center of gravity of the vehicle 1 , lr a distance between a rear axle and a center of gravity of the vehicle 1 , cf one Front axle skew stiffness, cr Rear skew stiffness, J Inertia moment of inertia of the vehicle 1 and m is a mass of the vehicle 1 indicate, wherein furthermore the selected output matrix C the yaw rate ψ̇ _{F of} the vehicle 1 outputs (the driving speed v the vehicle is understood here as a function of time t).

wobei $$ei{g}_{Fahrzeug}=eig\left(A\right)$$

The procedure M also includes under M5 comparing the actual proper motion 8a of the vehicle 1 with the modeled proper motion 8b of the vehicle 1 , where a self-motion difference 9 between the actual self-movement 8a of the vehicle 1 and the modeled proper motion 8b of the vehicle 1 is determined. Under M6 then becomes the determined eigenmotion difference 9 from a state observer 6 to the vehicle model 3 returned in the form of a control loop. The eigenvalues of the system matrix A 'can now be calculated in the usual way, where they are supplemented by an additional eigenvalue for the definition of a learning dynamics, which can possibly be speed-dependent, for example: $$ei{G}_{BeObacHter}=\left[ei{G}_{FaHrzeuG}+ei{G}_{Lerndynamik}\right]$$

in which $$ei{G}_{FaHrzeuG}=eiG\left(A\right)$$

By placing the poles, a corresponding feedback matrix can be determined. The steering wheel angle offset state quantity Δδ _Z of the lateral dynamics vehicle model 3 Due to the feedback, the steering wheel angle offset approaches more or less quickly Δδ on, ie converges to this. The speed of the approach, ie the learning process, can be adjusted by means of polar specification.

The learning may additionally be limited to stationary situations (steady cornering), where a change in the yaw rate ψ̇ and / or the steering wheel angle δ correspondingly small fails. In addition, can be limited to situations with low longitudinal and / or transverse dynamics, eg with a longitudinal acceleration a _x of the vehicle 1 and / or a lateral acceleration a _y of the vehicle 1 less than 3 m / s ² . It can thus be ensured that the vehicle model used also has validity.

In the foregoing detailed description, various features have been summarized to improve the stringency of the illustration in one or more examples. It should be understood, however, that the above description is merely illustrative and not restrictive in nature. It serves to cover all alternatives, modifications and equivalents of the various features and embodiments. Many other examples will be immediately and immediately apparent to one of ordinary skill in the art, given the skill of the art in light of the above description.

The exemplary embodiments have been selected and described in order to represent the principles underlying the invention and their possible applications in practice in the best possible way. As a result, those skilled in the art can optimally modify and utilize the invention and its various embodiments with respect to the intended use. In the claims as well as the description, the terms "including" and "having" are used as neutral language terms for the corresponding terms "comprising". Furthermore, a use of the terms "a", "an" and "an" a plurality of features and components described in such a way should not be excluded in principle.

LIST OF REFERENCE NUMBERS

1

vehicle

2

camera

3

Lateral dynamic vehicle model

4

roadway

5

Steering wheel angle sensor

6

state observer

7

camera data

8a

actual proper movement

8b

modeled proper motion

9

Proper motion difference

10

control device

δ

current steering wheel angle

Δδ

Steering wheel angle offset

Δδ _Z

Steering wheel angle offset state variable

ψ̇ _Z

Yaw rate state variable

β _Z

Slip angle state variable

ψ̇ _F

driven yaw rate

κ _F

driven curvature

κ _S

lane curvature

Δψ

orientation deviation

Dy

lateral deviation

a _x

longitudinal acceleration

a _y

lateral acceleration

M

method

M1

step

M2

step

M3

step

M4

step

M5

step

M6

step

Claims (8)

A method (M) of automatically learning a steering wheel angle offset (Δδ) during operation of a vehicle (1), comprising: measuring (M1) a current steering wheel angle (δ) of the vehicle (1) with a steering wheel angle sensor (5); Detecting (M2) a current driving situation of the vehicle (1) on a lane (4) for the current steering wheel angle (δ) by a camera (2) of the vehicle (1), wherein the current driving situation in the form of camera data (7) is detected ; Calculating (M3) an actual proper movement (8a) of the vehicle (1) from the camera data (7) for the current driving situation of the vehicle (1); Determining (M4) a modeled proper motion (8b) of the vehicle (1) based on the current steering wheel angle (δ) and a lateral dynamics vehicle model (3) of the vehicle (1), the lateral dynamics vehicle model (3) determining a steering wheel angle offset state quantity (8) Δδ _Z ) as degree of freedom; Comparing (M5) the actual proper motion (8a) of the vehicle (1) with the modeled proper motion (8b) of the vehicle (1), wherein a proper motion difference (9) between the actual proper motion (8a) of the vehicle (1) and the modeled proper motion (8b) of the vehicle (1) is determined; and returning (M6) the determined self-motion difference (9) to the vehicle model (3) in a closed loop such that the steering wheel angle offset state quantity (Δδ _Z ) converges to the steering wheel angle offset (Δδ). Method (M) after Claim 1 , wherein the eigenmotion difference (9) is fed back via a state observer (6). Method (M) after Claim 1 or 2 wherein the transverse dynamics vehicle model (3) is based on a one-track model comprising as state variables a slip angle (β _z ) of the vehicle (1), a yaw rate (ψ̇ _Z ) of the vehicle (1) and the steering wheel angle offset state variable (Δδ _Z ) , Method (M) after Claim 3 wherein the lateral dynamics vehicle model (3) comprises the current steering wheel angle (δ) as an input. Method (M) after Claim 3 or 4 wherein the transverse dynamics vehicle model (3) comprises a driven curvature (κ _F ) of the vehicle (1) or a driven yaw rate (ψ̇ _F ) of the vehicle (1) as an output variable. Method (M) according to one of Claims 1 to 5 wherein the actual intrinsic motion of the vehicle (1) is based on a lane curvature (κ _S ) of a lane of the vehicle (1) on the lane (4), a lateral deviation (Δy) of the vehicle (1) on the lane and an orientation deviation ( Δψ) of the vehicle (1) is calculated from the lane. Method (M) according to one of Claims 1 to 6 wherein the method (M) is applied to a longitudinal acceleration (a _x ) of the vehicle (1) and / or a lateral acceleration (a _y ) of the vehicle (1) less than 3 m / s ² . Regulating device (10), which for carrying out a method (M) according to one of Claims 1 to 7 is trained.

Images