WO2003098098A1

WO2003098098A1 - Method and device for determining the curvature of a vehicle lane

Info

Publication number: WO2003098098A1
Application number: PCT/EP2003/004888
Authority: WO
Inventors: Gabriel Schwab
Original assignee: Volkswagen Ag
Priority date: 2002-05-16
Filing date: 2003-05-09
Publication date: 2003-11-27
Also published as: EP1506365A1; DE10221900A1

Abstract

The invention relates to a method and device for determining the curvature of a vehicle lane. The inventive method involves the determination of the vehicle speed, the yaw rate and the steering angle. The method is characterized in that, according to the measured vehicle speed and the measured yaw rate, the curvature of the lane is calculated by using a single-lane model based on the vehicle speed and the measured steering-wheel angle or by using a yaw rate model based on the measured yaw rate and the vehicle speed. The inventive device comprises a vehicle speed sensor (2), a yaw rate sensor (1) and a steering angle sensor (14). Said device is characterized by having a decision unit (7), which is coupled to both the vehicle speed sensor (2) and the yaw rate sensor (1) and which receives, as input signals, the vehicle speed (vSP) and measured yaw rate (dψ/dt), and these input signals enable a single-lane model or a yaw rate model to be selected as a calculation model on the basis of the speed (vSP) and the measured yaw rate (dψ/dt). The inventive device is also characterized by having a calculating device (8, 9, 11) with which the curvature of the lane can be calculated by using the selected model. When using the single-lane model, the curvature is determined based on the vehicle speed and the measured steering angle. When using the yaw rate model, the curvature is calculated based on the measured yaw rate and the vehicle speed.

Description

Method and device for determining the curvature of a lane

vehicle

The present invention relates to a method and a device for determining the curvature of a lane of a vehicle.

Information about the curvature of the lane that a vehicle is traveling on can be used in particular to assist the driver when cornering. For example, in so-called adaptive motor vehicle headlight systems, the curvature values can be used to control a cornering light function. The direction of light emission from the headlamp is at least partially pivoted in the direction of the curve being driven. The cornering light function is particularly important in the case of adaptive motor vehicle headlight systems, since the illumination by the low beam of a conventional headlight in the curve is insufficient in many situations. Especially in the case of a left-hand bend (right-hand traffic is assumed in the following explanations; in the case of left-hand traffic, the page information should be reversed accordingly), on the one hand, the vehicle's own lane is poorly illuminated and, on the other hand, oncoming traffic is dazzling more than in the right-hand bend. It is therefore important for the cornering light function to know what the course of the road is in order to achieve the best possible illumination of the course of the road, while at the same time avoiding dazzling other road users.

In addition, driver assistance systems are known, in which various systems for environment detection of a motor vehicle are used. With automatic distance control, e.g. automatically maintains a sufficient safety distance from the vehicle in front. With this system, it is important to know whether the vehicle is cornering or not. In the known automatic distance control, the current curvature value for the driving lane is determined with the help of driving dynamics sensors. However, the models that are used to calculate the lane curvature have the disadvantage that they only insufficiently estimate the curvature of the lane in front of the vehicle.

This disadvantage can be remedied by calculation methods that use navigation data or video sensor data. Such video sensor data is obtained from images that are recorded by video cameras and that show the driving optically record the tool environment. The images obtained are evaluated using digital image processing so that the course of the road ahead of the vehicle can be determined. Disadvantages of the video sensor systems, however, are the very high hardware costs and the still insufficient image processing for determining the road.

With navigation systems, the current location of the vehicle, e.g. determined with a GPS (Global Positionings System) receiver. The location determined in this way is compared with the data on a digital map and the position of the vehicle is determined in this way. The data from the digital map can then be used to determine the exact course of the road on the road traveled by the vehicle. A disadvantage of the curvature determination using navigation data is that it is too imprecise. This is due on the one hand to the error in determining the current position using the GPS receiver and on the other hand to inaccuracies in digital maps. Furthermore, short-term changes in the course of the road cannot be taken into account. In addition, the hardware costs of such a system are relatively high.

It is therefore the object of the present invention to provide a method and a device for determining the curvature of a lane of a vehicle, with which the curvature of the lane is determined as accurately as possible, the method and the system should at the same time be able to be implemented cost-effectively.

According to the invention, this object is achieved by a method according to claim 1 and a device according to claim 12, with advantageous developments and further developments resulting from the subclaims.

According to the invention, the curvature of the vehicle's lane is calculated on the basis of vehicle dynamic data. The movement of a vehicle can be described by different approximation models. It was examined which of these models and under which conditions is particularly suitable for determining the curvature of a lane.

The so-called single-track model at M. Mitschke is: dynamics of motor vehicles - driving behavior; 2nd edition, volume C, Springer Verlag, Berlin, 1990. To simplify this model, two assumptions are made: 1. The vehicle's center of gravity is at road level. From this it follows that the centrifugal force acting in the center of gravity has no effect on the wheel loads. The additional load on the outside of the curve and the relief on the inside of the curve are neglected. A rotational movement about the transverse axis of the vehicle, the so-called pitching, and a rotational movement about the longitudinal axis of the vehicle, the so-called rolling, are neglected.

2. There is a linear system. This means that there is a linear relationship, for example, between the lateral force to be applied and the slip angle. This linearization is particularly permissible for lateral accelerations of less than 4 m / s ² . Different values apply to the tire stiffness (slip coefficient) for the front and rear axles. The deformation of the steering is also assumed to be linear. Their rigidity is combined with the front tire rigidity to the total rigidity of the front axle.

With the single-track model, the curvature of the lane can be calculated with a known vehicle speed and a known steering wheel angle, taking into account characteristic stiffness values for the vehicle. In particular, the tire stiffness and the deformation of the steering are taken into account.

When driving very slowly, the dynamics of the vehicle can be described by the so-called Ackermann model. It is approximately assumed that there are no lateral forces and therefore no slip angle on the wheels. In this case, the wheels roll without lateral slippage.

Finally, the driving dynamics can also be determined in the so-called yaw rate model from the change in time of the so-called yaw angle and the center of gravity speed of the vehicle. The yaw angle describes the rotation of the vehicle around a vertical axis.

Furthermore, the curvature of the lane can be calculated directly from the steering wheel angle, with corrections due to the deformability of vehicle components being incorporated in this model, not as in the single-track model. Finally, the curvature can be calculated from the direct measurement of the lateral acceleration and the center of gravity speed of the vehicle. It has now been found that the curvature of the lane can be calculated particularly well as a function of the current speed and yaw rate of the vehicle, ie the change in the yaw angle over time, using either the single-track model or the so-called yaw rate model. In the case of the single-track model, the curvature is determined by means of the vehicle speed and the measured steering wheel angle, in the case of the yaw rate model by means of the measured yaw rate and the vehicle speed.

In the method according to the invention for determining the curvature of a lane of a vehicle, the vehicle speed, the yaw rate and the steering wheel angle are measured. From this, depending on the speed and the measured yaw rate, the curvature of the lane is calculated either using a single-track model or using the yaw rate model.

According to an advantageous embodiment of the method according to the invention, if the vehicle speed is below a limit speed and the measured yaw rate is below a limit yaw rate, the curvature of the lane is calculated using the single-track model, otherwise using the yaw rate model. The limit speed is advantageously between 20 km / h and 40 km / h, preferably between 25 km / h and 35 km / h and particularly preferably 30 km / h. The limit yaw rate is advantageously between 1.5 degrees / s and 2.5 degrees / s, preferably between 1.8 degrees / s and 2.2 degrees / s and particularly preferably 2 degrees / s.

It has been found that the single-track model delivers particularly precise curvature values at speeds below these limit speeds and at low yaw rates below the specified limit yaw rates. If the limit speed or the limit yaw rate is exceeded, the yaw rate model delivers more precise curvature values, so that it is used according to the invention for calculating the curvature.

When calculating the curvature according to the single-track model, variables that are characteristic of the rigidity of the vehicle are preferably taken into account.

In the single-track model, the curvature can preferably be calculated using the following formula:

where v _{SP is} the center of gravity speed of the vehicle, δi is the steering wheel angle, the steering ratio,

/ is the center distance, l _v or l is the distance between the center of gravity and the front or rear axle, m is the mass of the vehicle, c _αV or c _{aH is} the stiffness of the front or rear tires, and c'av is the stiffness the front axle is.

In the yaw rate model, the curvature can preferably be calculated using the vehicle speed and the measured yaw rate using the following formula:

1 dψ κ =

'SP dt

where K is the curvature, dψ / dt the measured yaw rate and \ _ψ the center of gravity of the vehicle.

The vehicle speed is preferably the center of gravity speed of the vehicle. It is determined on the basis of the measured wheel speeds of the vehicle and the yaw rate.

According to an advantageous development of the method according to the invention, the curvature of a lane of a vehicle which is traveling on a roadway of known curvature is determined by means of the single-track model and / or the yaw rate model, and a speed-dependent correction factor is calculated from the determined curvature and the known roadway curvature and the calculated curvature is corrected in Dependence on the speed by the correction factor. The correction factor results from the quotient of the calculated curvature and known curvature. It has been found that the correction factor is in a range between 0.98 and 1.20. The device according to the invention for determining the curvature of a lane of a vehicle comprises a vehicle speed sensor, a yaw rate sensor and a steering wheel angle sensor. The device is characterized by a decision unit, which is coupled to the vehicle speed sensor and the yaw rate sensor and receives the vehicle speed and the measured yaw rate as input signals, and by which, depending on the speed and the measured yaw rate, either a single-track model or a yaw rate model can be selected. Furthermore, the device is characterized by a calculation device by means of which the curvature of the lane can be calculated on the basis of the selected model, the curvature being determined on the basis of the vehicle speed and the measured steering angle in the single-track model and on the basis of the measured yaw rate and the vehicle speed in the yaw rate model is calculated.

A limit speed and a limit yaw rate are preferably stored in the decision unit, the decision unit selecting the single-track model when the speed is below the limit speed and the yaw rate below the limit yaw rate, and otherwise selecting the yaw rate model as the calculation model.

When selecting the single-track model, the calculation device preferably calculates the curvature using the following formula:

κ = is and

where v _{SP is} the center of gravity speed of the vehicle, δ _/ . is the steering wheel angle, k is the steering ratio,

/ is the center distance, l _v or l is the distance between the center of gravity and the front axle or rear axle, m is the mass of the vehicle, c _αV or c _{a is} the stiffness of the front tires or the rear tires, and c ' _{aV is} the stiffness the front axle is. According to an advantageous development of the device according to the invention, it further comprises a memory unit in which correction factors are stored, the correction factors resulting from known and calculated curvatures as a function of vehicle speeds. The storage unit is coupled to the calculation device, so that curvatures calculated by the calculation device can be corrected by the correction factors stored in the storage unit.

The curvatures of the lane of a vehicle determined with the method and the device according to the invention can be used particularly well when controlling the cornering light function of an adaptive headlight system. The specific curvature can also be used in other driver assistance systems, e.g. an automatic distance control.

The invention will now be explained using exemplary embodiments with reference to the drawings.

1 schematically shows an exemplary embodiment of the device according to the invention for determining the curvature of a lane of a vehicle, FIG. 2 shows the results of measurements for different curvature models for driving at different speeds, FIG. 3 shows the values for a measurement run calculated according to the single-track model, the actual curve radii as well as the amount error of the single track model and

FIG. 4 shows the values calculated for the curve radii for the same measurement run as that on which FIG. 3 is based, according to the yaw rate model, and the amount of error for the yaw rate model.

The device for determining the curvature of a lane of a vehicle comprises a yaw rate sensor 1, a vehicle speed sensor 2 and a steering wheel angle sensor 14. These sensors 1, 2 and 14 determine the vehicle speed, the yaw rate and the steering wheel angle while driving.

As yaw rate sensor 1, speed sensor 2 and steering wheel angle sensor 14, sensors can be used, for example, which deliver the data for an electronic stability program (ESP) for driving dynamics control. Traditionally, the electronic stability program comprises a total of four as vehicle dynamics sensors Wheel speed sensors, a steering wheel angle sensor, a yaw rate sensor and a lateral acceleration sensor. The data from these arrangements are made about ^"a bus available and can also be used in this way for the calculation of the lane curvature.

Here, vehicle speed is understood to mean the speed v _S p of the vehicle's center of gravity. It is not measured directly, but determined from the signals of the individual wheel speeds. Since the trajectory of the vehicle's center of gravity is a superposition of a purely translational movement with the speed | v _s p | and the rotational movement with the angular velocity dψ / dt, which is the same for each point of the vehicle, the center of gravity velocity can be calculated as follows:

v _SP = Vi -? _; x άψ / German

The yaw rate vector dψ / dt has only one component in the vertical direction. The distances η are the distances from the respective wheel contact point to the center of gravity of the vehicle.

The wheel speeds of the front axle still have to be corrected by the steering angle. Since the wheel speed sensors can only measure the proportion of the speed in the wheel plane, an error results for the axis direction of the wheel, which is, however, advantageously corrected if the float angle is known.

In addition to the measured values of the wheel speeds and the yaw rate, the speed sensor 2 also detects whether the brake pedal is actuated. A distinction is then made between two cases when calculating the center of gravity:

1. The brake is applied. In this case, the maximum value of the calculated center of gravity speeds for the individual wheels is selected:

VSP = max (V _S PI, v _S p ₂ , v _SP3 , v _S P4)

2. The brake is not applied. In this case the average of the calculated center of gravity speeds of the non-driven axis is used: 'SP3 + V SP4

'SP

The above-described calculation of the vehicle speed is carried out in the unit 5. The speeds of the individual wheels, the yaw rate and the information as to whether the brake pedal is actuated serve as input variables. Unit 5 transmits the vehicle speed to units 6, 11 and 14 as an output variable.

The yaw rate determined by the yaw rate sensor 1 is also transmitted to the unit 3, in which the absolute amount of the yaw rate is formed. This absolute amount is finally transferred to unit 4. In unit 4, the measured yaw rate is compared with a limit yaw rate stored in unit 4. If the measured yaw rate exceeds the stored limit yaw rate, unit 4 transmits a corresponding signal to unit 7. Furthermore, the measured yaw rate is transmitted to units 9 and 11, as will be explained later. In the exemplary embodiments of the present invention, the limit yaw rate is in a range between 1.5 degrees / s and 2.5 degrees / s, preferably in a range between 1.8 degrees / s and 2.2 degrees / s and in a particularly preferred range The exemplary embodiment is the limit yaw rate 2 degrees / s.

In unit 6, the vehicle speed calculated in unit 5 is compared with a limit speed and it is determined whether this limit speed is exceeded. If the limit speed is exceeded, the unit 6 transmits a corresponding signal to the unit 7. In the exemplary embodiments of the present invention, the limit speed is in a range between 20 km / h and 40 km / h, preferably in a range between 25 km / h and 35 km / h and in a particularly preferred embodiment is 30 km / h.

In decision unit 7, depending on the measured vehicle speed and the measured yaw rate, either the single-track model or the yaw rate model is selected as the calculation model for the curvature. If the measured or determined vehicle speed is below the limit speed and the measured yaw rate is below the limit yaw rate, the single-track model is selected, otherwise the yaw rate model. This selection is transmitted to the calculation device 8.

In the calculation device 8, the curvature of the is calculated based on the selected calculation model, ie either by means of the single-track model or the yaw rate model Lane calculated. If the calculation is based on the single-track model, the curvature in the computing unit 9 is calculated using the following formula:

ιr w WoUUbeBIi vv _ch == - 7 _t ^c ' ^{aV Caiil is} 1st and m [c _aH l _H -c ^, _aV l _{V i}

where v _{SP is} the center of gravity speed of the vehicle, δi. is the steering wheel angle, i _{L is} the steering ratio,

/ the center distance is, l _v or ϊ _{H is} the distance between the center of gravity and the front or rear axle, m is the mass of the vehicle, c _aV or c _{aH is} the stiffness of the front or rear tires, and c ' _{aV is} the Stiffness of the front axle is.

For this purpose, the computing unit 9 receives the vehicle speed from the unit 5 and the steering wheel angle from the sensor 14. The stiffness data for the respective vehicle are stored in a storage unit.

If the yaw rate model was selected as the calculation model, the curvature in connection with the unit 11 is calculated as follows:

1 dψ κ = - v ςp dt

where dψ / dt is the measured yaw rate, ie the temporal change in the yaw angle, and v _S p is the center of gravity speed of the vehicle. With this formula, the change in the slip angle ß was neglected. It has been found that, for normal situations on country roads, the swimming angle ß in the range of] ß | <0.5 to 0.8 degrees. The change in the slip angle dß / dt is small compared to the yaw rate dψ / dt and can therefore be neglected.

Finally, the calculated curvatures can be corrected using correction factors. The correction factors depend on the speed. They are each for the track model and the yaw rate model stored in the storage unit 14. They were determined as follows:

For curvatures of roads with geometry known from road maps, the value ranges of the measured driving dynamics signals, i.e. in particular the yaw rate and the wheel speeds, and the most frequently occurring values are determined. These values are used to calculate the curvature and are related to the actual curvature. The quotient of the calculated curvature and the curvature taken from the road maps forms the correction factor. It has been found that the correction factor is in a range between 0.98 and 1.20. For curvature and speed ranges in which no measurement data are available, the correction factor is chosen to be one.

The correction factors and the selection of the model are transmitted from the storage unit 14 and the unit 7 to the correction factor calculation unit 16 via the unit 15. The calculated curvature is transmitted from the calculation device 8 to the multiplication unit 12, and the corresponding correction factor from the correction factor calculation unit 16. These two values are multiplied together in the multiplication unit 12 and output to the unit 13.

In order to be able to compare the calculation models for the curvature used in the exemplary embodiments of this invention, ie the single-track model and the yaw rate model, with other models for calculating the curvature, comparative measurements were carried out. A test vehicle was used for the measurements, in which the sensor signals wheel speeds, steering angle, yaw rate and lateral acceleration of the electronic stability program ESP, the information "driver brakes" and the time signal of a watch were recorded. The measurements were carried out on a dry road and with a constant load Before the start of the test, new "Continental ContiSportContact" and 205/55 R16 tires were fitted and run in for approx. 1000 km. Before each measurement, the tire pressure was checked and the tires warmed up. The vehicle-specific parameters required for the single-track model were measured. The measuring section was a circle with a radius of R = 205 m. For better understanding, the analyzed curvatures K are shown as radius values R = 1 / κ. The circular path was traversed at speeds between 10 km / h and 110 km / h. The results the curvature or radius calculations for the different curvature models are shown in FIG. 2. The mean values for the radius and the ranges of the standard deviation are shown. The area delimited by the standard deviation is indicated by the narrow bars.

Further measurements not shown here were carried out for radii in a range between 15 m and 305 m. For the single-track model, it follows that the curvature or the radius with little scatter is calculated regardless of the speed. The valid limit of the linear single-track model was exceeded only at high lateral accelerations> 0.65 g.

The Ackermann model only delivered accurate curvature values for the slowest speed at 13 km / h. The calculated curvatures became very imprecise for higher speeds.

In the yaw rate model, deviations of the calculated curvature from the actual curvature occur especially at low speeds. At speeds that correspond to a yaw rate of> 4 degrees / s, the curvature values calculated with the yaw rate model are subject to only slight variations and provide very good curvature values.

The curvature calculations with the lateral acceleration model are clearly falsified in most cases. The radius is calculated too small, especially for higher speeds.

The wheel speed difference model of the rear axle calculates the radius at low speed relatively accurately, although the standard deviation is very large. At higher speeds, the radius is calculated far too small, especially with large curve radii. The wheel speed difference model of the front axle delivers similar results.

As a result of these measurements, it was found that the single-track model and the yaw rate model are particularly well suited for the curvature calculation, the single-track model being preferred at lower speeds and lower yaw rates, and the yaw rate at higher speeds and higher yaw rates. In order to determine the preferred range of the single-track model and the yaw rate model more precisely, further measurements were carried out on a test section in which radii in the range from 400 to 500 m occurred. The results of these measurements and calculations are shown in FIG. 3 for the single-track model and in FIG. 4 for the yaw rate model. Curves 1 show the reference values for the radii on the test route, curve 2 of FIG. 3 shows the calculated radii on the basis of the Single track model, curve 2 of FIG. 4 shows the calculated radii on the basis of the yaw rate model, and curves 3 each show the amount error of the single track model or the yaw rate model, represented as the difference between the actual curvature and the calculated curvature.

These measurements reveal greater differences between the two models than with the circular drives.

The results of the measurements can be summarized as follows:

The single track model and the yaw rate model provide the most accurate curvature values. If one also takes into account the bank's banked slopes, the yaw rate model is the only one of the models examined that is insensitive to banked slopes. The yaw rate model has greater accuracy at higher yaw rates. For low speeds below a certain limit speed, the single-track model is therefore selected to calculate the curvature if the yaw rate is below a limit yaw rate. For yaw rates above the limit yaw rate and for speeds above the limit speed, the yaw rate model is used to calculate the curvature. The limit speed is advantageously in a range between 20 km / h and 40 km / h, preferably in a range between 25 km / h and 35 km / h and particularly preferably 30 km / h. The limit yaw rate is advantageously in a range between 1.5 degrees / s and 2.5 degrees / s, preferably in a range between 1.8 degrees / s and 2.2 degrees / s, and is particularly preferably 2 degrees / s.

Claims

claims

1. A method for determining the curvature of a lane of a vehicle, in which one measures the vehicle speed, the yaw rate and the steering wheel angle, characterized in that, depending on the measured vehicle speed and the measured yaw rate, the curvature of the lane using a single-track model using the vehicle speed and of the measured steering wheel angle or on the basis of a yaw rate model using the measured yaw rate and the vehicle speed.

2. The method according to claim 1, characterized in that, if the vehicle speed is below a limit speed and the measured yaw rate is below a limit yaw rate, the curvature of the lane is calculated using the single-track model, otherwise using the yaw rate model.

3. The method according to claim 2, characterized in that the speed limit between 20 km / h and 40 km / h and the limit yaw rate is between 1.5 degrees / s and 2.5 degrees / s.

4. The method according to claim 1, characterized in that the speed limit between 25 km / h and 35 km / h and the limit yaw rate is between 1.8 degrees / s and 2.2 degrees / s.

5. The method according to claim 2, characterized in that the limit speed is 30 km / h and the limit yaw rate is 2 degrees / s.

6. The method according to any one of the preceding claims, characterized in that in the single-track model, in the calculation of the curvature, data which for the

Rigidity of the vehicle are characteristic.

7. The method according to any one of the preceding claims, characterized in that the curvature of the single-track model is calculated using the following formula:

κ = where is and

where v _{SP is} the center of gravity speed of the vehicle, δi. is the steering wheel angle, k is the steering ratio, / is the center distance, l _v or IM is the distance between the center of gravity and the front or rear axle, m is the mass of the vehicle, c _aV or c _{aH is} the stiffness of the front tires or the rear tires, and c ' _a v is the stiffness of the front axle.

8. The method according to any one of the preceding claims, characterized in that the curvature is calculated using the yaw rate model using the following formula:

κ = - ^ v _S p dt

where K is the curvature, dψ / dt the measured yaw rate and v _{SP is} the center of gravity speed of the vehicle.

9. The method according to any one of the preceding claims, characterized in that the vehicle speed is the center of gravity speed of the vehicle and that the rotational speeds of the wheels of the vehicle are measured and the center of gravity speed of the vehicle is calculated from the speeds of the wheels and the yaw rate.

10. The method according to any one of the preceding claims, characterized in that the curvature of a lane of a vehicle which is traveling on a roadway of known curvature is determined in advance by means of the single-track model and / or the yaw rate model, and a speed-dependent correction factor is calculated from the determined curvature and the known roadway curvature, and that the calculated curvature is corrected by the correction factor depending on the speed.

11. The method according to claim 10, characterized in that the correction factor results from the quotient of the calculated curvature and known curvature and that the correction factor is in a range between 0.98 and 1.20.

12. Device for determining the curvature of a lane of a vehicle, comprising a vehicle speed sensor (2), a yaw rate sensor (1) and one

Steering angle sensor (14), characterized by

- A decision unit (7), which is coupled to the vehicle speed sensor (2) and the yaw rate sensor (1) and receives the vehicle speed (v _SP ) and measured yaw rate (dψ / dt) as input signals and by which in Depending on the speed (v _SP ) and the measured yaw rate (dψ / dt), a single-track model or a yaw rate model can be selected as the calculation model, and

- A calculation device (8, 9, 11) by means of which the curvature of the lane can be calculated on the basis of the selected model, the curvature being determined on the basis of the vehicle speed and the measured steering angle in the single-track model and on the basis of the measured yaw rate and the vehicle speed in the case of the yaw rate model is calculated.

13. The apparatus according to claim 12, characterized in that a limit speed and a limit yaw rate is stored in the decision unit (7) and that the decision unit (7) selects the single-track model if the speed is below the limit speed and the yaw rate is below the limit yaw rate and otherwise select the yaw rate model as the calculation model.

14. The apparatus according to claim 12 or 13, characterized in that

that the calculation device (8, 9) calculates the curvature using the following formula when selecting the single-track model:

κ is and

where v _{SP is} the center of gravity speed of the vehicle, δ. is the steering wheel angle, k is the steering ratio,

/ is the center distance, l _v or l is the distance between the center of gravity and the front axle or rear axle, m is the mass of the vehicle, c _{aM is} the stiffness of the front tires or the rear tires, and c ' _a v is the stiffness of the front axle ,

15. Device according to one of claims 12 to 14, characterized in that the calculation unit calculates the curvature using the following formula when selecting the vehicle speed and yaw rate as the calculation model:

1 dψ

K - -

¹ SP German

where K is the curvature, dψ / dt the measured yaw rate and v _S p is the center of gravity speed of the vehicle.

16. The device according to one of claims 12 to 15, characterized in that a memory unit (14) is also contained, in which correction factors are stored, the correction factors being known and calculated curvatures in Dependency on vehicle speeds, and that the storage unit (14) is coupled to the calculation unit (8, 9, 11, 16), so that curvatures calculated by the calculation unit (8, 9, 11, 16) by the in the storage unit (14 ) stored correction factors can be corrected.