DE102012221006B4 - Method for adapting a vehicle dynamics control - Google Patents

Abstract

Method for adapting a vehicle dynamics control for a vehicle, which regulates the difference between a measured actual yaw rate and a target yaw rate calculated based on a vehicle model through driver-independent interventions in at least one wheel brake and / or the steering, characterized in that it is checked whether quasi-stationary cornering is present, and that in this case a parameter of the vehicle model that describes the driving behavior of the vehicle is redetermined for a current curve direction, in particular the vehicle model being the dynamic single-track model and preferably the parameter that describes the driving behavior of the vehicle is the tire lateral stiffness of the rear axle, and that the current curve direction , i.e. the presence of a left or right turn based on a difference in wheel speeds and/or the sign of a measured yaw rate and/or the sign of a measured lateral acceleration and/or the sign of a measured steering wheel angle and/or from a parameter of the vehicle model, in particular a Lateral force on the rear axle is detected, and / or that the newly determined parameter of the vehicle model that describes the driving behavior of the vehicle is rejected if the stationary cornering in a constant cornering direction is shorter than for a predetermined period of time.

Description

The invention relates to a method according to the preamble of claim 1 and a system for vehicle dynamics control according to the preamble of claim 7.

To increase driving safety, modern vehicles are usually equipped with vehicle dynamics control, also known as Electronic Stability Control (ESC). Here, a vehicle model is used to check whether the behavior of the vehicle follows the driver's specifications, in particular whether a deviation between the calculated reference yaw rate ψ̇ _Ref and the measured actual yaw rate ψ̇ _Ist exceeds a predetermined threshold value. If a deviation is detected, braking torque is built up on at least one vehicle wheel independently of the driver in order to reduce this deviation by specifically introducing a yaw moment. This means that the vehicle remains controllable for the driver (within physical limits) even if unstable driving conditions occur.

The so-called linear dynamic on-track model of the vehicle is often used in the yaw moment control of an ESC system, also known as DPSIP control, to calculate the reference yaw rate or generally the target value. The front and rear wheels are combined in pairs to form a wheel located on the vehicle's longitudinal axis. To describe driving behavior, the following equations of motion are used in the linear dynamic single-track model: $$\begin{array}{l}m\cdot \left(\dot{\beta }+\dot{\psi }\right)\nu +\left(C_r+C_{ƒ}\right)\cdot \beta -\left(C_r{l}_r+C_{ƒ}{l}_{ƒ}\right)\cdot \frac{\dot{\psi }}{\nu }=C_{ƒ}\cdot \delta \\ \mathrm{\Theta }\cdot \ddot{\psi }\mathrm{+}\left(C_r{l}_r-C_{ƒ}{l}_{ƒ}\right)\cdot \beta -\left(C_r{l}_r{}^2-C_{ƒ}{l}_{ƒ}{}^2\right)\cdot \frac{\dot{\psi }}{\nu }=C_{ƒ}{l}_{ƒ}\cdot \delta \end{array}$$

With equation (1) one can use given input variables steering angle or front wheel steering angle δ and driving speed v as well as the vehicle parameters m (vehicle mass), Θ (mass moment of inertia around the vertical axis), l _f (distance vehicle center of gravity to front axle), l _r (distance vehicle center of gravity to rear axle). ), C _f (tire lateral stiffness of the front axle), and C _r (tire lateral stiffness of the rear axle) calculate the important movement variables ψ̈ (yaw acceleration), ψ̇ (yaw rate) and β (swimming angular velocity). If the actual yaw rate Ψ̇ _Ist measured by a sensor is constantly compared with the calculated reference yaw rate ψ̇ _Ref , the difference DPSIP between Ψ̇ _Act and ψ̇ _Ref can be regulated or minimized through targeted braking and/or steering interventions.

The vehicle parameters are expediently determined with the help of measurements on test vehicles. In the case of series vehicles in everyday use, the specific vehicle parameters may only describe the properties imprecisely; This can occur, for example, with heavy loads in the trunk or tires with special rubber compounds such as all-weather tires. It is therefore possible that the calculated reference yaw rate ψ̇ _Ref reflects the behavior desired by the driver so inaccurately that even in stable driving conditions, the magnitude of the difference between the reference yaw rate ψ̇ _Ref and the measured yaw rate Ψ̇ _Ist exceeds the threshold value for activating the vehicle dynamics control and thus a (due to unexpected change in the lateral dynamics and/or disturbing noises from a hydraulic pump) an uncomfortable incorrect control occurs for the driver.

In order to avoid incorrect regulations, the EP 1 089 901 B1 known to vary at least one of the parameters of the vehicle model depending on at least one separate measurement variable determined during stationary driving behavior.

From the DE 10 2008 030 667 A1 a method for determining parameters of a vehicle model is known, in which two estimating devices are provided for two driving situations. The estimation device provided for a driving situation is used to determine an estimated value.

From the DE 10 2007 029 958 A1 a method for preventing a vehicle from tipping over is known. A dynamic driving state variable such as the steering angle gradient is examined.

The DE 101 30 663 A1 describes a stability system for vehicles in which a special control mode is activated at high lateral accelerations, in which control interventions take place when the behavior is stable.

From the DE 10 2009 027 492 A1 It is known to determine a current driving situation by determining a large number of driving state variables. The steering torque provided is influenced based on the current driving situation.

The object of the present invention is to ensure an even more effective avoidance of incorrect control and/or a more precise vehicle dynamics control.

This object is achieved by a method according to claim 1 or a system for vehicle dynamics control according to claim 7.

A method is therefore provided for adapting a vehicle dynamics control for a vehicle, which regulates the difference between a measured actual yaw rate and a target yaw rate calculated based on a vehicle model through driver-independent interventions in at least one wheel brake and/or the steering. According to the invention, it is checked whether there is quasi-stationary cornering, and in this case a parameter of the vehicle model that describes the driving behavior of the vehicle is redetermined for an existing cornering direction. In particular, the dynamic single-track model is used as the vehicle model and the tire lateral stiffness of the rear axle is used as the parameter describing the driving behavior of the vehicle.

By using two different parameters C _r,new,left and C _r,new,right for the tire lateral stiffness of the rear axle in left and right curves, respectively, to calculate the reference yaw rate ψ̇ _Ref , an optimal match between the reference yaw rate ψ̇ _Ref and the measured yaw rate Ψ̇ is achieved _Is achieved for stable cornering even in vehicles with asymmetrical driving characteristics. This means that the control quality of the ESC control is significantly improved, which both increases driving safety and avoids incorrect controls that are uncomfortable for the driver.

It is useful if a quasi-stationary curve is recognized by the fact that the steering wheel angular velocity falls below a predetermined steering threshold value and/or the temporal change in the yaw rate falls below a predetermined yaw acceleration threshold value and/or the vehicle longitudinal speed exceeds a predetermined speed threshold value and/or the vehicle lateral acceleration lies within a predetermined tolerance band and/or the longitudinal direction of travel of the vehicle is recognized as forward and/or no driving stability control is active and/or the inclination of the road transverse to the direction of travel falls below a bank inclination threshold and/or the radius of curvature of the curve falls below a curvature threshold. In order to check the applicability of the vehicle model, current measured values from vehicle sensors (such as the steering angle sensor or the yaw rate sensor) or current values of variables of a vehicle model can be considered. A suitable driving situation can be recognized by the fact that one or more, preferably all, of the conditions mentioned are met.

According to the invention, the present curve direction, i.e. the presence of a left or a right curve, is determined based on a difference in wheel speeds and/or the sign of a measured yaw rate and/or the sign of a measured lateral acceleration and/or the sign of a measured steering wheel angle and/or from a parameter of the vehicle model, in particular a lateral force on the rear axle. It is particularly preferred to check whether the vehicle is moving forward by comparing the signs of the yaw rate and steering wheel angle and/or lateral acceleration. The newly determined parameter of the vehicle model that describes the driving behavior of the vehicle is expediently discarded if the stationary cornering in a constant cornering direction is shorter than for a predetermined period of time and / or the vehicle is reversing. The newly determined parameter of the vehicle model can then be clearly assigned to a left or right curve and was determined under defined conditions. It is particularly expedient if a determination of instantaneous values of the parameter of the vehicle model describing the driving behavior of the vehicle is carried out repeatedly as long as stationary cornering in a constant direction continues, and if an averaging of the newly determined instantaneous values is carried out. This means that statistical fluctuations in the parameter are suppressed.

Preferably, a present longitudinal dynamics situation is differentiated according to acceleration, overrun and braking, with each longitudinal dynamics situation being given a tolerance band for the current longitudinal acceleration, and a parameter of the vehicle model that describes the driving behavior of the vehicle, in particular a tire lateral stiffness of the rear axle, is redetermined for the present longitudinal dynamics situation. This makes it possible to take the shift in axle load distribution into account and thus achieve further improved accuracy of the vehicle model. The current longitudinal acceleration is particularly preferably determined from the signals from wheel speed sensors and/or a longitudinal acceleration sensor.

It is advantageous if a standard value of the parameter of the vehicle model that describes the driving behavior of the vehicle is or is saved permanently between two ignition runs, since the driving characteristics usually change slowly and continuously. By rejecting a newly determined value of the vehicle model parameter describing the driving behavior of the vehicle if the deviation between the newly determined value and the standard value exceeds a predetermined difference threshold value, incorrect adjustments to the parameter are avoided. On the other hand, if the newly determined parameter is plausible, the standard value can be modified based on the newly determined value, in particular a weighted averaging takes place, in which the newly determined value is weighted more heavily than the previous standard value. This means that the driving behavior is always described in the best possible way.

According to a preferred embodiment of the invention, when cornering is stationary, both the value of the parameter of the vehicle model that describes the driving behavior of the vehicle that is assigned to a fixed direction of the curve and a value for this parameter that is independent of the direction of the curve are determined, and the value assigned to a fixed direction of the curve is discarded if this deviates from the value independent of the direction of the curve by more than a predetermined difference threshold value. The value that is independent of the direction of the curve is determined particularly frequently and therefore has only small fluctuations, while a value that is dependent on the direction of the curve better describes asymmetries in the vehicle. This means that suitable driving situations are used particularly efficiently to adapt the parameter and at the same time implausible or incorrect adaptations are avoided.

The invention further relates to a system for driving dynamics control of a vehicle, comprising at least a yaw rate sensor, a lateral acceleration sensor, four wheel speed sensors and wheel brakes that can be operated independently of the driver and/or a steering system that can be operated independently of the driver. According to the invention, the system comprises means for determining one or more parameters describing the driving behavior of the vehicle, in particular a microprocessor which carries out a method according to the invention.

Further preferred embodiments result from the subclaims and the following description of an exemplary embodiment.

An embodiment of the side-by-side adaptive single-track vehicle model can be described in detail as follows:

1. Determination of the vehicle parameters in their original condition using test vehicles

To apply the ESC controller, ie adapting the ESC controller to the vehicle type, vehicle parameters such as the vehicle mass, the distance from the vehicle's center of gravity to the front axle and the distance from the vehicle's center of gravity to the rear axle are first measured directly on test vehicles. Further vehicle parameters, such as the mass moment of inertia about the vertical axis Θ, the tire lateral stiffness of the front axle C _f and the tire lateral stiffness of the rear axle C _r are identified by measurements on the test vehicle or vehicles during suitable driving maneuvers. In this parameter identification, the left and right curves are expediently taken into account equally and average values for the parameters are calculated. These original vehicle parameters are saved and used for vehicle dynamics control in new vehicles.

2. Online identification of vehicle parameters for series vehicles during everyday driving

Since the behavior of series vehicles in everyday driving can differ from that of the test vehicles depending on the vehicle condition (mechanical tolerances, load) and tire condition, the vehicle behavior in everyday driving is constantly monitored and re-identified. In order to easily record and quickly identify changes in driving behavior, i.e. to be able to describe the current behavior more precisely, a special method is used in which only a single parameter is varied, namely the tire lateral stiffness of the rear axle C _r . The other vehicle parameters are kept constant, while the tire lateral stiffness of the rear axle C _r is selected such that the behavior of the vehicle model according to equation (1) best reflects the behavior of the series vehicle during stable cornering. For this purpose, a least squares criterion can be considered, ie the squared deviation between target and actual behavior can be minimized for one or more variables. It makes sense to rewrite equation (1) as follows: $$\begin{array}{l}m\cdot a_y+\left(C_r+C_{ƒ}\right)\cdot \beta -\left(C_r{l}_r+C_{ƒ}{l}_{ƒ}\right)\cdot \frac{\dot{\psi }}{\nu }=C_{ƒ}\cdot \delta \\ \mathrm{\Theta }\cdot \ddot{\psi }\mathrm{+}\left(C_r{l}_r-C_{ƒ}{l}_{ƒ}\right)\cdot \beta -\left(C_r{l}_r{}^2+C_{ƒ}{l}_{ƒ}{}^2\right)\cdot \frac{\dot{\psi }}{\nu }=C_{ƒ}{l}_{ƒ}\cdot \delta \end{array}$$

In equation (2), the slip angle β is the only unknown quantity for a vehicle with ESC sensors, while the remaining movement variables such as lateral acceleration a _y , yaw rate ψ _̇ , steering angle δ, driving speed v and yaw acceleration ψ̈ are measured directly or indirectly with sensors or from the measured quantities can be derived. If you insert these movement variables into equation 2 and eliminate them

the slip angle β, you can calculate a new value for the tire lateral stiffness of the rear axle C _r for each loop run or loop of a vehicle dynamics control: $$C_{r,new}=\frac{m{l}_{ƒ}\frac{a_y}{\nu \dot{\psi }}-\frac{\mathrm{\Theta }\ddot{\psi }}{\nu \dot{\psi }}}{\frac{l^2}{{\nu }^2}-\frac{l\delta }{\nu \dot{\psi }}+\frac{m{l}_r\frac{a_y}{\nu \dot{\psi }}}{C_{ƒ}}+\frac{\mathrm{\Theta }\ddot{\psi }}{C_{ƒ}\nu \dot{\psi }}}$$

Equation (3) is used as the basis for the online identification of the vehicle parameter C _r, ie for determining the tire lateral stiffness of the rear axle while driving.

3. The calculation of the parameter change compared to the original parameter

In order to be able to better observe the parameter change and to avoid or detect possible calculation errors or sensor errors, it makes sense to directly assign the new value for the tire lateral stiffness of the rear axle C _r,new calculated using equation (3) to the original parameter C _r,Ori compare, preferably during each loop iteration. If the deviation exceeds a predetermined threshold value, this indicates an error. About the comparison

To simplify, the new value C _r,new is divided by the original parameter C _r,Ori : $${ƒ}_{C,new}=\frac{C_{r,new}}{C_{r,Ori}}=\frac{m{l}_f\frac{a_y}{v\dot{\psi }}-\frac{\mathrm{\Theta }\ddot{\psi }}{v\dot{\psi }}}{\frac{C_{r,Ori}{l}^2}{v^2}-\frac{C_{r,Ori}l\delta }{v\dot{\psi }}+\frac{C_{r,Ori}m{l}_r\frac{a_y}{v\dot{\psi }}}{C_{ƒ}}+\frac{C_{r,Ori}\mathrm{\Theta }\ddot{\psi }}{C_{ƒ}v\dot{\psi }}}$$

If you combine all parameters and movement variables in equation (4), then this equation can be rewritten as follows: $${ƒ}_{C,new}=\frac{C_{r,new}}{C_{r,Ori}}=\frac{\left(m{l}_f\right)\frac{a_y}{\dot{\psi }}-\frac{\mathrm{\Theta }\ddot{\psi }}{\dot{\psi }}}{\left(C_{r,Ori}{l}^2\right)\frac{1}{v}-\left(C_{r,Ori}l\right)\frac{\delta }{\dot{\psi }}+\left(\frac{C_{r,Ori}m{l}_r}{C_{ƒ}}\right)\frac{a_y}{\dot{\psi }}+\left(\frac{C_{r,Ori}\mathrm{\Theta }}{C_{ƒ}}\right)\frac{\ddot{\psi }}{\dot{\psi }}}$$

sowohl für den Zähler als auch für den Nenner ergibt: $${ƒ}_{C,new}=\frac{C_{r,new}}{C_{r,ori}}=\frac{\left(\frac{C_{ƒ}{l}_{ƒ}}{C_{r,ori}{l}_r}\right)\frac{a_y}{\dot{\psi }}-\left(\frac{\mathrm{\Theta }{C}_{ƒ}}{C_{r,ori}m{l}_r}\right)\frac{\ddot{\psi }}{\dot{\psi }}}{\left(\frac{C_{ƒ}l}{m{l}_r}\right)\frac{1}{v}-\left(\frac{C_{ƒ}l}{m{l}_r}\right)\frac{\delta }{\dot{\psi }}+\frac{a_y}{\dot{\psi }}+\left(\frac{\mathrm{\Theta }}{m{l}_r}\right)\frac{\ddot{\psi }}{\dot{\psi }}}$$

Another paraphrase with division by $\frac{C_{r,Ori}m{l}_r}{C_{ƒ}}$

for both the numerator and the denominator results in: $${ƒ}_{C,new}=\frac{C_{r,new}}{C_{r,Ori}}=\frac{\left(\frac{C_{ƒ}{l}_{ƒ}}{C_{r,Ori}{l}_r}\right)\frac{a_y}{\dot{\psi }}-\left(\frac{\mathrm{\Theta }{C}_{ƒ}}{C_{r,Ori}m{l}_r}\right)\frac{\ddot{\psi }}{\dot{\psi }}}{\left(\frac{C_{ƒ}l}{m{l}_r}\right)\frac{1}{v}-\left(\frac{C_{ƒ}l}{m{l}_r}\right)\frac{\delta }{\dot{\psi }}+\frac{a_y}{\dot{\psi }}+\left(\frac{\mathrm{\Theta }}{m{l}_r}\right)\frac{\ddot{\psi }}{\dot{\psi }}}$$

The terms in the brackets depend exclusively on the original vehicle parameters and can be replaced with 5 summary new parameters Para1 to Para5. Equation (6) can then be simplified as follows: $${ƒ}_{C,new}=\frac{C_{r,new}}{C_{r,Ori}}=\frac{Para1\cdot \frac{a_y}{\dot{\psi }}-Para2\frac{\ddot{\psi }}{\dot{\psi }}}{Para3\cdot \frac{1}{\nu }-Para4\frac{\delta }{\dot{\psi }}+\frac{a_y}{\dot{\psi }}+Para5\cdot \frac{\ddot{\psi }}{\dot{\psi }}}$$

By comparing equations (6) and (7), the 5 new parameters can be clearly defined and calculated. With equation (7), the ratio f _c,new of the new tire lateral stiffness of the rear axle C _r,new to the original parameter C _rOri can be calculated at any time (e.g. every loop or loop pass).

As long as the vehicle behavior remains stable, the ratio f _c,new should remain constant, and if the everyday vehicle were identical to the test vehicles, the ratio should be

f _c,new must be equal to 1. This is not the case in practice for various reasons, including tolerances of mechanical components, various disruptive influences on the vehicle and faulty sensors for the movement variables. In order to identify the driving behavior of an everyday vehicle, additional measures must be taken to keep possible external interference and the influence of inaccurate measurement signals as small as possible.

4. General conditions for the online identification of the tire lateral stiffness of the rear axle C _r,new and the relative change f _c,new

• • Es liegt keine Aktivität einer Fahrstabilitätsregelung vor, d.h. weder Antiblockiersystem (ABS) noch Traktionskontrolle (TCS) noch ESC sind aktiv.
• • Das Fahrzeug durchfährt eine Kurve in Vorwärtsrichtung, was z.B. anhand der Vorzeichen von Gierrate und Querbeschleunigung überprüft werden kann.
• • Die Fahrbahn bzw. das Fahrzeug weist keine große Querneigung auf. Dies kann überprüft werden, indem das Produkt aus Gierrate und Fahrzeuggeschwindigkeit mit der Querbeschleunigung verglichen wird.
• • Die Bahnlinie überschreitet eine bestimmte Krümmung bzw. der Kurvenradius unterschreitet einen vorgegebenen Schwellenwert.
• • Die Querbeschleunigung liegt innerhalb eines vorgegebenen Toleranzbands, vorzugsweise beträgt die Querbeschleunigung zwischen 0,2 g und 0,4 g. Somit ist gewährleistet, dass die Fahrzeugeigenschaften in dem stabilen linearen Bereich liegen.
• • Die Gierbeschleunigung und Lenkwinkelgeschwindigkeit sollten jeweils einen vorgegebenen Schwellenwert unterschreiten, damit eine quasistationäre Fahrt vorliegt, die durch eine geringe Dynamik gekennzeichnet ist.

For a meaningful online identification of the tire lateral stiffness of the rear axle C _r,new, an important prerequisite is that there is quasi-stationary and stable cornering, since only then are the simplified equations of motion valid for the single-track model. Therefore, the fulfillment of one or more, preferably all, of the present general conditions is checked:

• • There is no driving stability control activity, ie neither the anti-lock braking system (ABS), traction control (TCS) nor ESC are active.
• • The vehicle curves in the forward direction, which can be checked, for example, using the signs of the yaw rate and lateral acceleration.
• • The road or the vehicle does not have a large bank. This can be checked by comparing the product of yaw rate and vehicle speed with the lateral acceleration.
• • The path line exceeds a certain curvature or the curve radius falls below a specified threshold value.
• • The lateral acceleration is within a specified tolerance band, preferably the lateral acceleration is between 0.2 g and 0.4 g. This ensures that the vehicle properties are in the stable linear range.
• • The yaw acceleration and steering angular velocity should each fall below a predetermined threshold value so that there is quasi-stationary travel, which is characterized by low dynamics.

It is expedient to check whether the general conditions mentioned above are met at the same time, whereupon an online identification of the tire lateral stiffness of the rear axle C _r,new is carried out. This is repeated continuously while driving, with each loop or loop pass generating a new value for the tire lateral stiffness of the rear axle C̈̈̂ _r,new or f̈ _c,new according to Eq. (7) is calculated. The calculated values are preferably temporarily stored. If the calculations have been carried out for a certain duration (e.g. n loops), a learning cycle is fired and the average value for C _{r ,new} or f _c,new determines: $${\overline{ƒ}}_{c,new}=\frac{\mathrm{\Sigma }{\widehat{ƒ}}_{c,new}}{An\mathrm{e.g}aHlderLOOps}$$

If the above general conditions are still met at the same time, then a new learning cycle is expediently started.

5. Differentiated identification of the relative change f _c,new for left and right curves and/or division for different driving situations

• Gasgeben,
• Bremsen und
• Schubbetrieb unterteilt. Für die aktuell vorliegende Situation wird die entsprechende Reifenseitensteifigkeit der Hinterachse nach einem erfolgten Lernzyklus zweckmäßigerweise wie folgt mit einem Tiefpassfilter neu berechnet:

$${ƒ}_{c,new,S}=\left(1-\lambda \right)\cdot {ƒ}_{c,old,S}+\lambda \cdot {\overline{ƒ}}_{c,new}$$

For example, due to different wheel loads, a vehicle in different longitudinal dynamic situations also has different driving behavior for lateral dynamics. It is therefore advantageous to assign the new mean value for the tire lateral stiffness of the rear axle learned according to equation (8) to a specific driving situation and a clear cornering direction. According to a preferred classification, the longitudinal dynamic situations in

• Step on the gas,
• brakes and
• Overrun operation divided. For the current situation, the corresponding tire lateral stiffness of the rear axle is expediently recalculated using a low-pass filter after a learning cycle has been completed:

$${ƒ}_{c,new,S}=\left(1-\lambda \right)\cdot {ƒ}_{c,Old,S}+\lambda \cdot {\overline{ƒ}}_{c,new}$$

Here λ is the weighting factor for the newly learned mean f _c,new and the index S denotes the corresponding longitudinal dynamic situation when accelerating, braking or coasting.

This means that a distinction is preferably made between the three different longitudinal dynamic situations as well as between left and right turns. It is advantageous to determine the direction of the curve using the sign of the lateral force on the rear axle from the single-track model. If the lateral force is positive, then there is a left turn, otherwise a right turn is detected. Depending on the direction of the curve, the corresponding tire lateral stiffness of the rear axle is recalculated as follows using a low-pass filter: $${ƒ}_{c,new,S,R}=\left(1-\lambda \right)\cdot {ƒ}_{c,Old,S,R}+\lambda \cdot {\overline{ƒ}}_{c,new}$$

Here the index R denotes the direction of the curve.

If the three situation-dependent tire lateral stiffnesses of the rear axle are learned from equation (9) and two additional tire lateral stiffnesses of the rear axle that depend on the direction of the curve are learned from equation (10) for each driving situation, then a total of six side-by-side adaptive values and three side-independent adaptive values can accurately describe the driving behavior of the vehicle.

6. Use of the identified new parameters in the single-track model

The learned tire lateral stiffness of the rear axle is used in the single-track model depending on the driving situation, which means that the best possible reference yaw rate is available for each side and depending on the situation.

• • Die Anzahl der Lernzyklen für die vorgegebene Fahrsituation und Kurvenrichtung hat einen bestimmte Schwelle (z.B. 5) überschritten.
• • Der seitenunabhängige adaptive Wert liegt zwischen den beiden seitenweise adaptiven Werten.
• • Der Unterschied zwischen jedem seitenweise adaptiven Wert und dem seitenunabhängigen adaptive Wert liegt unter einem Schwellenwert, insbesondere sollte der relative Unterschied 5% unterschreiten.

The reliability of the newly learned tire lateral stiffness usually depends on the number of learning cycles. The more frequently the situation and the direction of the curve occur while driving, the more reliable the corresponding learned value is. Since the page-by-page adaptive values are determined from fewer learning cycles than the page-independent adaptive value of the same driving situation, a plausibility check of the page-by-page adaptive values is preferably carried out before they are used for the single-lane model. Conveniently, the page-by-page adaptive values are considered reliable and used if one or more, in particular all, of the following conditions are met:

• • The number of learning cycles for the specified driving situation and curve direction has exceeded a certain threshold (eg 5).
• • The page-independent adaptive value is between the two page-wise adaptive values.
• • The difference between each page-wise adaptive value and the page-independent adaptive value is below a threshold value, in particular the relative difference should be less than 5%.

Otherwise, the side-independent adaptive value of the parameter is used in the single-track model.

Because a braking situation occurs much less frequently than the other two situations during normal driving, it may be necessary

be, the adaptive values for the braking situation learned for coasting or accelerating are used in the single-track model. For example, the number of learning cycles can be used as a measure of the quality of a new learned value for the lateral stiffness of the rear axle.

7. Saving and reading out the identified new parameters in a suitable storage medium

The new parameters for tire lateral stiffness of the rear axle identified with the help of equation (9) and equation (10) are expediently stored in a suitable storage medium with a suitable format when the ignition is switched off so that they can also be used for future journeys. Saving the identified new parameters can be made dependent on conditions. For example, the reliability of the identified new parameters can be assessed based on a statistical evaluation of the individual loops or determined values. Furthermore, it can be required that the difference between new parameters and parameters present at the start of the journey and/or the original parameters lies within a predetermined tolerance band.

It is advantageous if the stored parameters or the tire lateral stiffness of the rear axle are read from the storage medium the next time the ignition is switched on and checked for plausibility. If, for example, the values lie within a specified tolerance band and are therefore suitable, they are preferably used directly in the single-track model.

By learning or identifying vehicle parameters that describe driving behavior depending on the left or right curve as well as the existing longitudinal dynamics situation and using them in the single-track model when calculating the reference yaw rate in particular, vehicles with asymmetrical properties can be better described. This improves the quality of the control, which on the one hand increases driving safety and on the other hand leads to a reduced occurrence of incorrect controls.

Claims (9)

Method for adapting a vehicle dynamics control for a vehicle, which regulates the difference between a measured actual yaw rate and a target yaw rate calculated based on a vehicle model through driver-independent interventions in at least one wheel brake and/or the steering, characterized in that it is checked whether quasi-stationary cornering is present, and that in this case a parameter of the vehicle model that describes the driving behavior of the vehicle is redetermined for a current curve direction, in particular the vehicle model being the dynamic single-track model and preferably the parameter that describes the driving behavior of the vehicle is the tire lateral stiffness of the rear axle, and that the current curve direction , i.e. the presence of a left or right turn based on a difference in wheel speeds and/or the sign of a measured yaw rate and/or the sign of a measured lateral acceleration and/or the sign of a measured steering wheel angle and/or from a parameter of the vehicle model, in particular a Lateral force on the rear axle is detected, and / or that the newly determined parameter of the vehicle model that describes the driving behavior of the vehicle is rejected if the stationary cornering in a constant cornering direction is shorter than for a predetermined period of time. Procedure according to Claim 1 , characterized in that a quasi-stationary curve is recognized by the fact that the steering wheel angular velocity falls below a predetermined steering threshold value and / or the temporal change in the yaw rate falls below a predetermined yaw acceleration threshold value and / or the vehicle longitudinal speed exceeds a predetermined speed threshold value and / or the vehicle lateral acceleration within a predetermined tolerance band and/or the longitudinal direction of travel of the vehicle is recognized as forward and/or no driving stability control is active and/or the inclination of the road transverse to the direction of travel falls below a bank inclination threshold and/or the radius of curvature of the curve falls below a curvature threshold. Method according to at least one of the Claims 1 until 2 , characterized in that an existing longitudinal dynamic situation is differentiated according to acceleration, overrun and braking, with each longitudinal dynamic situation being given a tolerance band for the current longitudinal acceleration, which is preferably determined from the signals from wheel speed sensors and / or a longitudinal acceleration sensor, and that the driving behavior of the vehicle descriptive parameters of the vehicle model, in particular tire lateral stiffness of the rear axle, are redetermined for the present longitudinal dynamic situation. Method according to at least one of the Claims 1 until 3 , characterized in that a determination of instantaneous values of the parameter of the vehicle model describing the driving behavior of the vehicle is carried out repeatedly as long as stationary cornering continues in a constant direction, and that an averaging of the newly determined instantaneous values is carried out. Method according to at least one of the Claims 1 until 4 , characterized in that a standard value of the parameter of the vehicle model describing the driving behavior of the vehicle is stored or is that a newly determined value of the parameter of the vehicle model describing the driving behavior of the vehicle is discarded if the deviation between the newly determined value and the standard value exceeds a predetermined difference threshold exceeds, and that the standard value is modified based on the newly determined value, in particular a weighted averaging takes place, in which the newly determined value is weighted more heavily than the previous standard value. Method according to at least one of the Claims 1 until 5 , characterized in that in the presence of stationary cornering, both the value of the parameter of the vehicle model that describes the driving behavior of the vehicle, which is assigned to a fixed curve direction, and a value for this parameter that is independent of the direction of the curve is determined, and that the value assigned to a fixed curve direction is discarded if this deviates from the value independent of the direction of the curve by more than a predetermined difference threshold value. System for driving dynamics control of a vehicle, at least comprising a yaw rate sensor, a lateral acceleration sensor, four wheel speed sensors and, in addition, wheel brakes that can be operated independently of the driver and/or a steering system that can be operated independently of the driver, characterized in that this means for determining parameters describing the driving behavior of the vehicle, in particular a microprocessor who carries out a method according to one of the preceding claims. Vehicle dynamics control system Claim 7 , characterized in that the means or means for determining parameters describing the driving behavior of the vehicle further comprise means for temporarily storing and/or filtering certain or measured variables or parameters, in particular a value independent of the direction of the curve and two values each assigned to a direction of the curve being determined. Vehicle dynamics control system Claim 7 , characterized in that the means or means for determining parameters describing the driving behavior of the vehicle further comprise means for temporarily storing and/or filtering specific or measured variables or parameters, whereby a value independent of the direction of the curve and two values each assigned to a direction of the curve are determined, whereby Each value is assigned a measure of the quality of this value, in particular a statistical measure, and that the values assigned to a curve direction are only used for vehicle dynamics control if they have sufficient quality, otherwise the value independent of the curve direction is used for vehicle dynamics control.