DE102008047750A1 - Driver assistance system for vehicle, has module for analysis of torsional vibration of wheel of vehicle and for generating estimated value of coefficient of friction on basis of analysis - Google Patents

Abstract

The driver assistance system (100) has modules (105,106,109) for analysis of a torsional vibration of a wheel of a vehicle and for generating an estimated value of a coefficient of friction (111) on basis of the analysis. A data fusion module (107) is provided for combination of the estimated value with another estimated value of the coefficient of friction. A classification of the coefficient of friction is carried out. Independent claims are included for the following: (1) a method for assisting a driver of a vehicle; (2) a program element for implementing on a processor; and (3) a computer-readable medium for storing program element.

Description

Field of the invention

The The invention relates to the assistance and control technology for vehicles. In particular, the invention relates to a driver assistance system for a vehicle, a vehicle with a driver assistance system, a method for Assistance of a driver of a vehicle, a program element and a computer readable medium.

Technological background

at usual Anti-lock braking systems (ABS systems) and also with usual Electronic Stability Program (ESP) systems it requires a coefficient of adhesion or coefficient of friction between Estimate tires and road surface and this then one of the further control of the brake pressure lying algorithm first pretend. Such a relatively accurate estimate of the coefficient of friction requires its own sensors and is complex.

Summary of the invention

It An object of the invention is to provide a coefficient of friction estimation. which works in normal driving and with few sensors gets along.

It are a driver assistance system, a vehicle, a procedure, a Program element and a computer-readable medium according to the features the independent one claims specified. Further developments of the invention will become apparent from the dependent claims.

The described embodiments concern equally the driver assistance system, the vehicle, the procedure, the computer-readable Medium and the program element. In other words, the referred to below with regard to the driver assistance system Features also in the process, the program element and in the computer readable Implement medium or the vehicle, and vice versa.

According to one embodiment The invention provides a driver assistance system for a vehicle, which a first module for analyzing a torsional vibration of a wheel of the Vehicle and for generating a first estimate of a coefficient of friction based on the analysis. Furthermore, a data fusion module provided, which for combining the first estimate with a second estimate of the friction coefficient is, whereby a classification of the coefficient of friction takes place. The Driver assistance system is used for adaptation based on the classification the coefficient of friction.

In In other words, it is not necessary for a precise coefficient of friction estimated becomes. Rather, the coefficient of friction in, for example, three or more classes divided. So it will be the road conditions classified. This information can then be used to adapt driver assistance systems, such as ACC (Adaptive Cruise Control) or APIA (Active-Passive Integration Approach), or also for adapting a recuperation operation of the vehicle serve.

Especially the coefficient of friction estimation works or classification during constant ride of the vehicle under evaluation of the wheel vibration behavior. A constant ride represents a typical driving situation. To this coefficient of friction estimation to support and to enlarge the workspace will be more Procedures combined, differing in their validity complete.

According to one another embodiment The invention provides a second module, which is the generation the second estimate serves.

According to one another embodiment The invention provides a third module, which is the generation a third estimate the coefficient of friction, which is then fed to the data fusion module, wherein the data fusion module for combining the first estimate with the second estimate and with the third estimate of the friction coefficient is, whereby the classification of the coefficient of friction takes place.

According to a further embodiment of the invention, a fourth module is provided, which is designed to generate a fourth estimated value of the friction coefficient, which is then the Datenfusionsmo dul is supplied. In this case, the data fusion module is designed to combine the fourth estimated value with, for example, the first estimated value or the first and the second estimated value, or the first, the second and the third estimated value, or with a combined result from the three estimated values, whereby the classification of the friction coefficient he follows.

In In other words, the fusion module, for example, also in several parts accomplished and in a first step, the first and the second estimate with each other merge (ie combine). After that will be in a subsequent step the third estimate combined with the fusion result. Also, in a first step the first three estimates then the fourth estimate with this result combined.

The The result is a classification of the current friction coefficient in three or more classes.

According to one another embodiment According to the invention, the first estimate is based on monitoring a rotation of the wheel by a first detection device in the vehicle, with the second estimate on a monitor the slip of the wheel by a second detection device in based on the vehicle.

For each individual wheel of the vehicle can in this case provided its own sensor be.

According to one another embodiment According to the invention, the third estimate is based on monitoring a driving dynamics of the vehicle by a third detection device.

According to one another embodiment According to the invention, the fourth estimate is based on a statistical estimate Evaluation of different input variables.

According to one another embodiment belonging to the invention the three detection devices to an ESP sensor of the vehicle. New sensors need not installed.

According to one another embodiment The invention combines the estimated values in the data fusion module based on averaging of the estimates.

Also can, according to one another embodiment invention, the combination of estimates in the data fusion module based on a fuzzy logic.

According to one another embodiment The invention relates to the first detection device as a wheel speed sensor executed the analysis of the torsional vibration of the wheel based on a model the torsional vibration takes place.

According to one another embodiment The invention relates to a vehicle with a driver assistance system described above specified.

According to one another embodiment The invention relates to a method for assisting a driver of a vehicle Vehicle, in which an analysis of the torsional vibration of a Wheel of the vehicle takes place. Then the generation of a first estimate a coefficient of friction based on the analysis. Then the first estimated value with a second estimate of the friction coefficient combined, whereby a classification of the friction coefficient he follows. With the result of the classification, an adaptation of the Driver assistance system are made.

Also Is it possible, the classification result for to use the control of a Rekuperationsbetriebs.

According to one another embodiment The invention is a program element specified that, when it on running a processor will instruct the processor, the process steps given above perform.

In this case, the program element z. B. be part of a software that is stored on a processor of the vehicle management. The processor can also be the subject of the invention. Furthermore, this embodiment of the invention comprises a program element, which already uses the invention from the beginning, as well as a program element, which by updating an be standing program for the use of the invention causes.

According to one another embodiment The invention provides a computer readable medium on which a program element is stored that when on a processor accomplished will instruct the processor, the process steps given above perform.

in the Below, with reference to the figures, embodiments of the invention.

Brief description of the figures

1 shows a schematic representation of a system for Reibbeiwertschätzung according to an embodiment of the invention.

2 shows a wheel of a vehicle with sensors for evaluating the torsional vibration of the wheels according to an embodiment of the invention.

3 shows the dependence of the frequency spectrum of the coefficient of friction α.

4 shows a system for Reibbeiwertschätzung according to another embodiment of the invention.

5 shows an LM estimate.

6 shows a vehicle according to an embodiment of the invention.

7 shows a flowchart of a method according to an embodiment of the invention.

8th shows a system for Reibbeiwertschätzung according to another embodiment of the invention.

Detailed description of exemplary embodiments

The Representations in the figures are schematic and not to scale.

In The following figure description will be for the same or similar Elements use the same reference numbers.

1 shows a system for Reibbeiwertschätzung according to an embodiment of the invention. The system 100 has four detection devices (sensors) 101 . 102 . 103 and 104 on. Each of the detection devices can in this case have a plurality of individual sensors.

In the detection device 101 they are sensors for detecting a torsional vibration of a tire. These are, for example, wheel speed sensors.

The detection device 102 includes sensors to monitor the slip. The detection device 103 includes, for example, sensors for monitoring the driving dynamics of the vehicle, such as the yaw rate, the speed, the steering wheel angle and / or the slip angle.

All these sensors can already be part of the ESP sensor system of the vehicle.

The recorded measured values of the individual sensors are then sent to corresponding modules 104 . 105 . 106 and 109 to hand over. The first module 104 serves the analysis of the torsional vibration of the respective wheels and generates from the analysis result an estimated value of the coefficient of friction, which is then sent to the fusion module 107 is handed over.

At It should be noted that the vehicle is at a motor vehicle or can also act on a motorcycle.

The second module 105 generates a second estimate, the third module 106 generates a third estimate and the fourth module 109 generates a fourth estimate, in this case based on a statisti evaluation of the input variables of the sensors 110 ,

All four estimates are handed over to the fusion module and then converted into a classification of the coefficient of friction.

The classified coefficient of friction 111 then becomes a module 108 for further use.

In the module 108 For example, it is an adaptive cruise control system for an ACC system, a control unit for the ESP or, more generally, a control unit for a driver assistance system. The classified coefficient of friction can also be used for other systems, for example for setting the recuperation operation.

The following are the four modules 104 . 105 . 106 and 109 and their operation described in more detail.

In the first module 104 the evaluation of the torsional vibration of the tires takes place. Vehicle tires have a typical vibration whose frequency depends on the coefficient of friction and the air pressure. By a parameter estimation or identification of the associated vibration model, the friction coefficient can thus be estimated.

For this will be For example, the signals of the wheel speed sensors (as in the so-called. DDS +) corrected by the pole pitch error and then equidistant interpolated. Typical wheel speed sensors consist of a pole wheel and a sensor that detects the edges of the pole wheel. As a result receives one the time stamp of the individual flanks. By manufacturing tolerances The division of the pole wheel is not equidistant. These deviations can identified and compensated by identification or averaging become. From the timestamps and the distances between the individual flanks The speed can then be calculated on the pole wheel. The timeline however, these values are dependent from the occurrence of the flanks and thus variable. That's why the interpolation is the values on a constant timeline needed.

at DDS + is a so-called "Deflation Detection System." It recognizes one Tire pressure loss indirectly from the data of the wheel speed sensors the electronic brake system. Because the rolling circumference of a tire decreases with pressure loss. This is not just the changes the dynamic rolling circumference, but with an additional Software module also a changed Vibration behavior of the tire evaluated. This allows one precise Calculation of tire inflation pressure and an early one Information of the driver.

Based on these signals, a model of the rotational vibration of the tire is identified. This can be done in the time domain (eg in the case of RLS or when using a Kalman filter) or in the frequency domain. For this model, the movement of the rim (ω ₁ ) depending on the wheel speed over ground is formulated with the coefficient of friction (represented by α).

The force model for road contact is:

T describes the running-in behavior of the tire and depends on the speed and the contact surface with the road.

A schematic structure of such a wheel with corresponding sensors is also in 2 shown. Also shows 2 a curve representing the dependence of F _x (ie the force in the direction of travel) and the slip s.

From the following equation, the angular velocity can be calculated ω ₁ (s). ν is the vehicle speed, which is superimposed on the velocity component ν _s . ν _s describes the influence of the dynamic rolling radius, which depends, among other things, on the vertical excitations through the road and changes in the road:

T

wie oben die Zeitkonstante des Einlaufverhaltens des Reifens;

K

die Tosionssteifigkeit des Reifens;

ω₂

die Drehgeschwindigkeit des Gürtels;

J₁

das Trägheitsmoment der Felge;

J₂

das Trägheitsmoment des Gürtels;

R

der Radradius;

s

die sogenannte LaPlace Größe.

Here are:

T

as above, the time constant of the running-in behavior of the tire;

K

the torsional stiffness of the tire;

ω ₂

the speed of rotation of the belt;

J ₁

the moment of inertia of the rim;

J ₂

the moment of inertia of the belt;

R

the wheel radius;

s

the so-called LaPlace size.

201 refers to the wheel hub, 202 the tire circumference, 203 ω ₂ , 204 J ₂ , 205 the radius of the tire R, 206 the force in the X-direction, 207 a first spring, 209 a second spring and 208 the force in the Z direction. 209 denotes J ₁ , 210 ω ₁ . The model of power transmission between the wheel and the roadway is exemplified by the arrangement 211 (Serial spring-damper element) described. The symbol 211 symbolizes a contact to a fixed point.

3 shows the dependence of the frequency spectrum of α. The horizontal axis 301 denotes the frequency between 0 and 150 Hz. The vertical axis 302 denotes the oscillation amplitude | G |.

The curves 303 . 304 and 305 show three frequency spectra at three different coefficients of friction α. The frequency spectra are two resonances 306 . 307 at approx. 40 Hz and approx. 60 Hz. These resonances can be taken into account by identifying these two frequency ranges separately with a second order model or the entire frequency range with a third or fourth order model (as in the above equation). Furthermore, the energies or the energy ratios in all resonances can be taken into account. This is followed by a unit that fuses the various pieces of information (fusion module).

4 shows a schematic representation of the method according to an embodiment of the invention.

Unterer Frequenzbereich (um 40 Hz) Block 408

Oberer Frequenzbereich (um 70 Hz) Block 409

The evaluation of the wheel speeds is based on the equation XXX (w1 = ...). Since the individual orders are excited differently depending on the excitation, equation XXX is divided into three submodels depending on the valid frequency range and simplified: Entire frequency range block 407

Lower frequency range (around 40 Hz) Block 408

Upper frequency range (around 70 Hz) Block 409

The blocks 401 . 402 and 403 represent bandpass filters (BP). In 4 the exemplary evaluation of the models with the parameter estimation method Recursive-Least-Squares (RLS) is described. For this purpose, first the required derivatives in the blocks 404 . 405 and 406 determines, the actual parameter estimation takes place in the blocks 407 . 408 and 409 , The estimation errors of the models are in the block 410 determined and consulted for Reibwertschätzung. Furthermore, the ratios of the resonances in the upper and lower energy range in block 411 determined and also used for friction value monitoring.

From the information of the blocks 407 - 411 will be in the block 107 a coefficient of friction determined. At the block 107 it can be a multi-dimensional map that relates the different inputs to a coefficient of friction. An exemplary implementation is a neural network to which the dependencies are trained based on measurements.

This Method works on a wheel-individual basis, so that wheel-individual statements and page-wise statements on μ-split detection are possible.

In the second detection unit 102 (please refer 1 ) the slip monitoring takes place. In normal driving conditions, different levels of slippage occur in different road conditions (dry, wet, snow, icy). The slip is determined from the page-by-side comparison between front and rear wheel speed.

For example runs parallel for the torsional vibration evaluation, a model using the ABS / ESP sensor technology as Input and the slip on high friction has as output. This Model is learned specific to the tire (so-called reset button) or vehicle-specific established. From the deviations between the measured and estimated slip the road condition can be determined.

By the page-wise processing is also possible here μ-split detection.

If the deviation between the measured and the estimated slip is very small, a Hochreibwert can be recognized, if others Method and especially the statistical described below Evaluation points to a reduced coefficient of friction.

In the third detection unit 103 the monitoring of the driving dynamics takes place. The measured values are in the module 106 to hand over. The driving dynamics variables such as the yaw rate, speed, steering wheel angle or slip angle are physically related, with the friction coefficient or the friction-dependent skew stiffnesses playing a central role. By monitoring these quantities (by the detection unit 103 ) can be closed, for example via a Kalman filter on the coefficient of friction. This is done in the module 106 (please refer 1 ).

• – Außentemperatur unterhalb einer voreingestellten Schwelle. Daraus ergibt sich, dass der Reibbeiwert kleiner ist.
• – ABS/ESP/TCS-Eingriffe. Je häufiger diese Eingriffe auftreten, umso kleiner ist der mögliche Reibbeiwert.
• – Verwendung der Reibbeiwertschätzung von ABS/ESP/TCS.
• – Häufiges Anschlagen der anderen Verfahren. Daraus ergibt sich ebenfalls ein kleinerer Reibbeiwert.
• – Im Fusionsmodul 107 (siehe 1, 4, 8) werden die berechneten Reibbeiwerte der verschiedenen Verfahren miteinander kombiniert. Die Kombination erfolgt beispielsweise durch Mittelwertbildung, gewichtete Mittelwertbildung unter Verwendung von Gütewerten der einzelnen Verfahren, Fuzzy-Logik oder indem jedes Verfahren ein Vertrauensintervall liefert, wobei die verschiedenen Vertrauensintervalle überlagert werden und der Bereich mit der größten Überschneidung die fusionierte Aussage ergibt.

In the module 109 a statistical evaluation is carried out. Based on various inputs from the fourth detection unit 110 a base friction coefficient is determined. Here, different input variables deliver a coefficient of friction, which is then linked to other friction coefficients (eg via fuzzy logic). Possible input variables are:

• - Outside temperature below a preset threshold. It follows that the coefficient of friction is smaller.
• - ABS / ESP / TCS interventions. The more frequently these interventions occur, the smaller the possible coefficient of friction.
• - Use of the coefficient of friction estimation of ABS / ESP / TCS.
• - Frequent striking of other procedures. This also results in a smaller coefficient of friction.
• - In the fusion module 107 (please refer 1 . 4 . 8th ), the calculated coefficients of friction of the different methods are combined. The combination is done, for example, by averaging, weighted averaging using quality values of the individual methods, fuzzy logic, or by providing each method with a confidence interval, superimposing the different confidence intervals, and giving the merged statement the area with the largest intersection.

The Result of the merger is a classification of the coefficient of friction, for example, in high, medium, low.

For example decides the statistical evaluation about whether a high or a mean friction coefficient is present.

If one of the other three methods responds, the friction coefficient class is reduced by one step. If two other methods are involved, the friction coefficient class is reduced by two levels. If three methods respond or if one of the methods responds particularly strongly, the low friction coefficient is output.

5 shows a result of Reibbeiwertklassifikation according to an embodiment of the invention. On the transverse axis 501 the time is shown from 70 to 95 seconds. On the vertical axis 502 is the so-called LM indicator shown. A high LM value (LM = Low μ), such as the value at t = 90 s, means a small coefficient of friction and a low LM value, as at t = 72 s means a high coefficient of friction.

In the first period, the vehicle is on a main road 507 , In the second section 508 turns the vehicle to the right. In the third section 509 the vehicle is on a narrow street and in the fourth section 510 the vehicle approaches an intersection.

As in the curve 503 can be seen, the LM indicator depends on the current situation. In the measurement made, the detection and recovery times are less than 300 milliseconds. This is especially true for the area 505 , In the area 504 the Vehicle Dynamics Observer / ESP switches on. In the area 506 is braked, so that the corresponding slip model is used.

6 shows a vehicle according to an embodiment of the invention. The vehicle 601 has a system 100 for friction coefficient estimation. In particular, the system uses sensor data from ESP sensors and / or brake system sensors.

7 shows a flowchart of a method in which in step 701 an analysis of the torsional vibration of a wheel of the vehicle is performed. In step 702 an estimate is generated based on the analysis and in step 703 a combination of this estimated value with a further estimated value with regard to the friction coefficient, whereby a classification of the coefficient of friction takes place. In step 704 then the adaptation of a driver assistance system takes place on the basis of the classification of the friction coefficient.

8th shows another system for Reibbeiwertschätzung, in which the measured values of the detection devices 101 . 102 . 103 to the appropriate further processing modules 104 . 105 . 106 be handed over.

Furthermore, a model 801 used, the measured values a _x 802 and _ay 803 be supplied. This lobe model relates to the longitudinal dynamics and the lateral dynamics of the vehicle. Alternatively to this model can also another model 806 from the values a _x and a _y the values R _TOA 807 , R _WSM 808 and RIO 809 calculated. These values are valid values (for example, between 0 and 1) of the three methods 104 - 106 , (Tire Oscillation-Analsis, Wheel-Slip-Model and Vehicle-Dynamics-Observer).

Based on one of these models as well as the measured values of the detection devices 101 . 102 and 103 takes place in the modules 104 . 105 and 106 in each case an estimate of the friction coefficient. This is followed by the first part of the fusion module 107 a first weighting, in which also the model 801 respectively. 806 is taken into account. The result is a (fused) friction coefficient 810 , the second part of the fusion module 107 is handed over. This second part of the fusion module 107 will also be the result of the statistical evaluation 811 to hand over. In this second part, a weighting takes place, taking into account the model 801 or 806 , The final result is then the classified coefficient of friction 111 for example, a driver assistance unit 108 is handed over.

In addition to that noted that "comprising" and "having" no other elements or excludes steps and "one" or "one" does not exclude a multitude. Further It should be noted that features or steps with reference to one of the above embodiments have been described, also in combination with other features or steps of other embodiments described above can. Reference signs in the claims are not limitations to watch.

Claims (15)

Driver assistance system for a vehicle, the driver assistance system comprising: a first module ( 104 ) for analyzing a torsional vibration of a wheel of the vehicle and for generating a first estimate of a friction coefficient based on the analysis; a data fusion module ( 107 ) for combining the first estimate with a second estimate of the coefficient of friction, thereby classifying the coefficient of friction; wherein the driver assistance system is adapted for adaptation based on the classification. The driver assistance system of claim 1, further comprising: a second module ( 105 ) for generating the second estimate. Driver assistance system according to one of the preceding claims, further comprising: a third module ( 106 ) for generating a third estimate of the friction coefficient associated with the data fusion module ( 107 ) is supplied; wherein the data fusion module ( 107 ) is carried out for combining the first estimated value with the second estimated value and with the third estimated value of the friction coefficient, whereby the classification of the friction coefficient takes place. The driver assistance system of claim 1 or 2, further comprising: a fourth module ( 109 ) for generating a fourth estimate of the friction coefficient associated with the data fusion module ( 107 ) is supplied; wherein the data fusion module ( 107 ) is performed to combine the first estimate with the second estimate, the third estimate, and the fourth estimate of the coefficient of friction, thereby classifying the coefficient of friction. Driver assistance system according to one of claims 2 to 4, wherein the first estimated value on a monitoring of rotation of the wheel by a first detection means ( 101 ) in the vehicle. the second estimated value being based on a monitoring of the slip by a second detection device ( 102 ) in the vehicle. Driver assistance system according to one of claims 3 to 5, wherein the third estimated value on a monitoring of driving dynamics of the vehicle by a third detection device ( 103 ). Driver assistance system according to one of claims 4 to 6, where the fourth estimate based on a statistical evaluation of different input variables. Driver assistance system according to one of the preceding claims, wherein the three detection devices ( 101 . 102 . 103 ) belong to an ESP sensor of the vehicle. Driver assistance system according to one of the preceding claims, wherein the combination in the data fusion module ( 107 ) based on an averaging of the estimates. Driver assistance system according to one of the preceding claims, wherein the combination in the data fusion module ( 107 ) based on a fuzzy logic. Driver assistance system according to one of the preceding claims, wherein the first detection device ( 101 ) is designed as a wheel speed sensor; wherein the analysis of the torsional vibration of the wheel is based on a model of torsional vibration. Vehicle with a driver assistance system after a the claims 1 to 11. Method for assisting a driver of a vehicle, the method comprising the steps: Analysis of a torsional vibration a wheel of the vehicle; Generation of a first estimate a coefficient of friction based on the analysis; Combination of the first estimated value with a second estimate of the friction coefficient, whereby a classification of the friction coefficient he follows; Adaptation of a driver assistance system based on the Classification. Program element that when on a processor accomplished will guide the processor to perform the following steps: analysis a torsional vibration of a wheel of the vehicle; Generation of a first estimate a coefficient of friction based on the analysis; Combination of the first estimated value with a second estimate of the friction coefficient, whereby a classification of the friction coefficient he follows; Adaptation of a driver assistance system based on the Classification. Computer-readable medium on which a program element which, when executed on a processor, is stored instruct the processor to perform the following steps: analysis a torsional vibration of a wheel of the vehicle; Generation of a first estimate a coefficient of friction based on the analysis; Combination of the first estimated value with a second estimate of the friction coefficient, whereby a classification of the friction coefficient he follows; Adaptation of a driver assistance system based on the Classification.