DE102007034143B4 - A method for determining a model yaw rate and / or a model lateral acceleration and a model steering angle - Google Patents

Abstract

Method for determining a model yaw rate and / or a model lateral acceleration as so-called model size in an observer model for a vehicle dynamics control system of a motor vehicle having a steering system, wherein a measured steering quantity (yaw rate, aY) and the driving speed ( Vx) of the vehicle on which the determination is based, characterized in that the driver's manual torque (HM) applied to the steering system and used for determining the model size is taken from at least one three-dimensional characteristic map (Lookup Table (2-D)) at least one of these two variables (yaw rate or lateral acceleration aY) is retrieved via the relationship between the manual torque (HM) and the yaw rate or between the manual torque (HM) and the lateral acceleration (aY), which is stored in dependence on the vehicle speed (Vx).

Description

The invention relates to a method for determining a model yaw rate and / or a model lateral acceleration according to the preamble of claim 1 and to a method for determining a model steering angle according to the preamble of claim 4.

The prior art is on the DE 100 32 340 A1 referenced, which describes a vehicle steering system in which a variable torque assistance of the driver's steering request and triggered by a vehicle dynamics control intervention in the position of the steered wheels without mutual interference can be realized. For this purpose, among other things, the detected driver steering angle is modified depending on driving speed.

The term "vehicle dynamics control system" of a motor vehicle used herein can be taken to mean both longitudinal-dynamic as well as lateral-dynamic as well as vertical-dynamic control systems. The (current) yaw rate of a motor vehicle (when cornering) generally flows in addition to a longitudinally dynamic brake control system (so-called "ESP" or "DSC") in a lateral dynamic control system, which is usually a controllable steering system, for example steerable wheels of the front axle of a two-lane vehicle, if this example. With a steer-by-wire system or with a steering system with selectively variable steering ratio (eg., The so-called. Active steering of the applicant of the present patent application) is equipped, and / or for possibly Steerable rear wheels of a preferably two-lane vehicle, so this is equipped with a so-called. Rear-wheel steering. The (current) lateral acceleration of the vehicle (when cornering) generally flows into a vertical dynamic control system, which may be, for example, a suspension system with so-called. Federfußpunktverstellung or a system for changing the spring stiffness of a stabilizer, which, for example the applicant under the name "Dynamic Drive" in series.

Known both the yaw rate and the lateral acceleration of a motor vehicle by means of sensors mounted in the vehicle can be measured, but it can be on the one hand in the aforementioned vehicle dynamics control systems to high safety critical systems whose control intervention should never take place due to the signals of a single sensor; Furthermore, a vehicle dynamics control system is characterized by the fastest possible response. With regard to the latter requirement, a quasi-leading signal for the lateral acceleration or yaw rate is thus desired, which can be computationally determined by means of a so-called model in which a measurable steering variable, usually the steering angle set by the driver, and the current driving speed of the Vehicle are processed as input variables to determine from this an expected yaw rate or lateral acceleration. This leading yaw rate or lateral acceleration can then also be compared with the measured yaw rate or lateral acceleration, so that - without providing additional sensors - a safety-related redundancy is provided.

As the model mentioned here usually one or the so-called. Single track model is used, for example, via the so-called. Ackermann formula from the steering angle a model yaw rate and from this by multiplication with the vehicle speed a model lateral acceleration is determined. In most cases, however, such a model is valid only in the so-called linear range, but no longer in the limit of the tires and the driving dynamics, so that in many driving situations no valid model sizes can exist. Furthermore, the time lapse of the steering angle provided by the driver relative to the variables "yaw rate" and "lateral acceleration" measurable on / in the vehicle is relatively small, since a corresponding reaction must already have occurred for the detection or measurement of a steering angle on the driver's steering wheel.

An improved method according to the preamble of claim 1 is therefore an object of the present invention. The solution of this object is achieved by a method having the features of the independent claims. Advantageous developments are content of the dependent claims.

It was recognized that a sufficiently precisely reproducible, nonlinear relationship exists between the so-called hand moment and the yaw rate of the vehicle and between the so-called hand moment and the lateral acceleration of the vehicle and finally also between the so-called hand moment and the currently set steering angle although in each case depending on the driving speed of the vehicle. These three relationships can thus be represented (over the driving speed of the vehicle) in three-dimensional maps, after these relationships have preferably been determined experimentally for a particular vehicle type. With a known manual torque and a known driving speed, a stationary value for each of these three variables (yaw rate, lateral acceleration, steering angle) can thus easily be retrieved from the respective three-dimensional characteristic map.

In principle, a value determined in this way can already be determined with the corresponding value determined with a sensor (for yaw rate or Lateral acceleration or steering angle) in order to obtain information about the reliability of the sensor signal (in the sense of redundancy).

It is of particular advantage that at least in recent power assisted steering systems of motor vehicles, namely in the so-called. Electric steering (with electromotive steering assistance) said hand moment already exists as a measure, since this is needed for a suitable control of the electric motor. Advantageously, this already existing measured variable (of the driver to the steering system, in particular steering wheel, applied manual torque) can be used to from a suitable observer model with the help of at least one of the three maps mentioned at least one of the three model sizes, namely a model Yaw rate or a model lateral acceleration or a model steering angle. For a further use in a vehicle dynamics control system, the so-called stationary value of each of these three variables determined so far is not ideally suited, for which reason it is proposed in the sense of an advantageous development that those from the mentioned (respective) characteristic map The quantity called up is passed through a high-pass filter or a low-pass filter and is then compared with a corresponding quantity determined by measurement and / or in another (conventional) model, in particular via a weighted averaging in order to obtain the desired model size. The latter can then be incorporated directly into a vehicle dynamics control system.

In the mentioned high-pass filter or low-pass filter, the signal-specific transient or stationary properties can be crystallized out. In this way it is possible to control stationarily the respective sensor value and nevertheless advantageously use the so-called phase advance of the signal calculated from the hand moment, which is understood to mean the so-called "advance" of the respective signal described above. In other words, by using the measured manual torque applied by the driver to his steering wheel and not waiting for a steering angle to be set thereby, the desired model sizes can be obtained earlier in time.

As already mentioned above, the output signal of the observer model proposed here, namely the so-called model size, is finally calculated by a weighted averaging, to which at least one corresponding size determined by other means is used. For example. Thus, a model yaw rate can be determined by suitable averaging from the yaw rate value determined via the manual torque and the associated three-dimensional characteristic map and a value determined by means of a yaw rate sensor and a yaw rate value determined from the steering angle via the aforementioned Ackermann formula Mean of these three values mentioned).

The weighting factors on which this weighting is based can depend on various boundary conditions; in particular, however, they can be dependent on the validity of the respective underlying model under the current boundary conditions, ie, if the model validity is insufficient (for example because the vehicle is moving near its dynamic driving limit range), the corresponding one in this model hidden model value and the other values are weighted accordingly higher. For example, the known single-track model is invalid at high lateral acceleration, low vehicle speed or high wheel slip. On the other hand, a model lateral acceleration or model yaw rate determined from the moment of hand, as proposed here, can be relatively inaccurate if the vehicle is heavily oversteered. In principle, however, the availability of the observer models proposed here on the basis of the manual torque is better than that of the known (and mentioned) single track model, since the specified relationships contained in three-dimensional maps (between the manual torque and the yaw rate or between the manual torque and the lateral acceleration or between the manual torque and the steering angle) in each case are valid until into the dynamic driving limit area.

As an alternative to such weighted averaging, it is also possible to use a Lünberger observer basically known to the person skilled in the art for calculating the model yaw rate or model lateral acceleration and applying the manual torque applied by the driver to the steering system of the vehicle using said relationships (in three-dimensional maps) ,

In the attached 1 - 3 various observer models are shown as possible embodiments of the present invention.

HM_Sensor

ist das mittels eines Sensors ermittelte, vom Fahrer des Fahrzeugs an dessen Lenksystem angelegte Handmoment;

Vx

ist die Fahrgeschwindigkeit des Fahrzeugs;

Gierrate

ist die vorzugsweise mittels eines Gierratensensors ermittelte Gierrate des Fahrzeugs;

HM

ist (abermals) das Handmoment, jedoch muss dieses nicht zwangsweise von einem Sensor gemessen sein;

aY

ist die vorzugsweise mittels eines Querbeschleunigungssensors ermittelte Querbeschleunigung des Fahrzeugs;

Radschlupf

ist der (vorzugsweise geschätzte) Schlupf zwischen den Fzg.-Rädern und der Fahrbahn;

Schwimmwinkelgeschwindigkeit

ist die (vorzugsweise geschätzte) Schwimmwinkelgeschwindigkeit des Fahrzeugs;

Lookup Table (2-D)

ist das genannte dreidimensionale Kennfeld, welches fahrgeschwindigkeitsabhängig (deshalb hier als 2-D bezeichnet) den Zusammenhang zwischen Handmoment und einer der drei Größen „Gierrate”, „Querbeschleunigung”, Lenkwinkel” enthält;

Gewichtungsfaktoren

in diesem Block werden aus den angegebenen Eingangsgrößen verschiedene Gewichtungsfaktoren gebildet, nämlich:

Faktor_HM

als Gewichtungsfaktor für die über das Handmoment aus dem „Lookup Table (2-D)” bestimmte Größe,

Faktor_ESM

als Gewichtungsfaktor für die über das Einspurmodell bestimmte Größe;

Faktor_Sensor

als Gewichtungsfaktor für die aus einem (gemessenen) Sensorsignal bestimmte Größe;

Hochpass

bezeichnet ein Hochpassfilter für dass jeweilige Signal;

Tiefpass

bezeichnet ein Tiefpassfilter für dass jeweilige Signal;

Einspurmodell

in diesem Block wird aus den angegebenen Größen, nämlich der Fahrgeschwindigkeit Vx und dem vom Fahrer eingestellten Lenkwinkel ein Modellwert für die Querbeschleunigung (= aY_ESM) in bekannter Weise bestimmt;

LW_Sensor

stellt das Sensorsignal für den Lenkwinkel (LW) dar;

AY_Sensor

ist ein gemessenes Sensorsignal für die Querbeschleunigung des Fahrzeugs;

x

bezeichnet eine mathematische Multiplikation;

+

bezeichnet eine mathematische Addition;

The following are initially only in the 1 . 2 briefly explained:

HM_Sensor

is determined by a sensor, the driver of the vehicle at the Steering system applied manual torque;

Vx

is the driving speed of the vehicle;

yaw rate

is the yaw rate of the vehicle, preferably determined by means of a yaw rate sensor;

HM

is (again) the hand moment, but this does not necessarily have to be measured by a sensor;

aY

is the transverse acceleration of the vehicle, preferably determined by means of a lateral acceleration sensor;

wheel slip

is the (preferably estimated) slip between the vehicle wheels and the roadway;

Swimming angular velocity

is the (preferably estimated) float velocity of the vehicle;

Lookup Table (2-D)

is said three-dimensional map, which depending on driving speed (therefore referred to here as 2-D) contains the relationship between manual torque and one of the three variables "yaw rate", "lateral acceleration", steering angle ";

weighting factors

In this block, different weighting factors are formed from the specified input variables, namely:

Faktor_HM

as a weighting factor for the size determined by the hand moment from the "Lookup Table (2-D)",

Faktor_ESM

as a weighting factor for the size determined via the single-track model;

Faktor_Sensor

as a weighting factor for the variable determined from a (measured) sensor signal;

highpass

denotes a high pass filter for that respective signal;

lowpass

denotes a low pass filter for that respective signal;

single-track

In this block, a model value for the lateral acceleration (= aY_ESM) is determined in a known manner from the given variables, namely the driving speed Vx and the steering angle set by the driver;

LW_Sensor

represents the sensor signal for the steering angle (LW);

AY_Sensor

is a measured sensor signal for the lateral acceleration of the vehicle;

x

denotes a mathematical multiplication;

+

denotes a mathematical addition;

The observer model after 1 used to determine a model lateral acceleration as shown both a lateral acceleration measurement, namely "AY_Sensor", as well as a model value from the Einspurmodell, namely "aY_ESM", as well as from the manual torque on the said three-dimensional map ("Lookup Table (2 -D) ") and weight these three values with different weighting factors to obtain a representative model value for vehicle lateral acceleration by weighted averaging.

The observer model after 2 is similar to the one after 1 constructed and also determines a model lateral acceleration, but without explicit measurement of the lateral acceleration by means of a sensor. The here proposed determination of a model size (here the model lateral acceleration) using the measured manual torque of the driver, who this applies to the steering system, in particular to his steering wheel, can thus replace a commonly provided lateral acceleration sensor.

The still exemplified observer model according to 3 serves to determine a model steering angle. In addition to the designations already described, there are two integrator links and also a Luenberger observer model. The weighting factors used are a static factor (Faktor_stat) for the static steering angle determination and a dynamic factor (Faktor_dyn) for the dynamic steering angle determination. As can be seen, the hand momentum is integrated twice to obtain a model dynamic steering angle. From the aforementioned three-dimensional characteristic map ("Lookup Table (2-D)") results in a stationary relationship between the manual torque and the steering angle. As in 3 shown, the dynamic steering angle is corrected via a Luenberger observer model. Subsequently, the stationary and the dynamic steering angle are combined via a weighted averaging. This model can thus replace a steering angle sensor. It goes without saying that a model yaw rate and a model lateral acceleration can again be calculated from this model steering angle via the known single track models, wherein it should also be pointed out that quite a variety of details can be deviated from the above explanations without departing from the content of the patent claims.

Claims (4)

Method for determining a model yaw rate and / or a model lateral acceleration as so-called model size in an observer model for a vehicle dynamics control system of a motor vehicle having a steering system, wherein a measured steering quantity (yaw rate, aY) and the driving speed ( Vx) of the vehicle on which the determination is based, characterized in that the driver's manual torque (HM) applied to the steering system and used for determining the model size is taken from at least one three-dimensional characteristic map (Lookup Table (2-D)) at least one of these two variables (yaw rate or lateral acceleration aY) is retrieved via the relationship between the manual torque (HM) and the yaw rate or between the manual torque (HM) and the lateral acceleration (aY), which is stored in dependence on the vehicle speed (Vx). A method according to claim 1, characterized in that the retrieved from the map (Lookup Table (2-D)) size is passed through a high pass filter (high pass) or a low pass filter (low pass) and then with a corresponding, by measurement and / or in calculated from another (conventional) model, in particular via a weighted averaging in order to obtain the desired model size. A method according to claim 2, characterized in that the or the weighting factor (s) for the averaging of the validity of the respective model depending on the current boundary conditions is / are. A method for determining a model steering angle in an observer model for a vehicle dynamics control system of a motor vehicle with a steering system, wherein a measurement-based steering variable (yaw rate, aY) and vehicle speed (Vx) of the vehicle, the determination can be based on, characterized in in that the driver's manual torque (HM) applied to the steering system is used to determine the model steering angle by selecting from a three-dimensional characteristic map (Lookup Table (2-D)) on the relationship stored therewith in dependence on the vehicle speed (Vx) between the manual torque (HM) and the steering angle (LW) a stationary steering angle is retrieved, from which the model steering angle is determined by the stationary steering angle by two times integration and correction via another observer model, a dynamic steering angle is determined, which steering angle, namely the stationary and the dynamic steering angle, over e weighted averaging can be merged.

Images