Method for determining the loading status of a vehicle

Abstract

Method for determining the loading status of a vehicle, wherein the wheel speeds of the vehicle's wheels are measured using wheel speed sensors, where from the wheel speed signals at least one tire pressure-insensitive parameter (λ) i ) is determined for the current load of the vehicle and is characterized by this, that the respective parameter (λ i , λ komp,i ) the mean of the amplitudes | S φ ' R φ ' R ( f ) | a selected pressure-invariant and load-sensitive frequency range between 36 and 49 Hz of a transfer function as a function of a frequency spectrum obtained from a wheel speed signal and is used directly or indirectly via derived quantities to determine the influence of the load change on quantities derived from the rolling circumference effect (Axle, Diag, Side).

Description

[0001] The invention relates to a method for determining the loading condition of a vehicle, wherein the wheel speeds of the vehicle's wheels are measured with the aid of wheel speed sensors.

[0002] Tire pressure monitoring of a vehicle, particularly a motor vehicle, can be performed directly using specially designed sensors or indirectly. Indirect tire pressure monitoring involves determining indicator values.

[0003] A change in a vehicle's load affects, among other things, the rolling circumference of a tire, thus influencing parameters used in indirect tire pressure monitoring. This can lead to load changes being mistakenly interpreted as changes in tire pressure and, for example, as tire pressure loss. If a load change and a tire pressure loss occur simultaneously, the pressure loss may not be reliably detected.

[0004] Indirect tire pressure monitoring therefore has the following disadvantages. In the worst case, a change in load is interpreted as a pressure loss, and a false pressure loss warning is issued. If the load changes and a simultaneous pressure loss occur, the system may not be able to issue a warning due to competing effects on the pressure loss criterion and may not detect the pressure loss.

[0005] DE 43 00 677 A1 describes how a sensor can measure the vibration spectrum of a vehicle's body and transform it into a frequency spectrum in real time. The vibration frequencies can be used to determine the weight of the body and the tire pressure.

[0006] JP 2005300254 A discloses a method in which a load condition is determined from a frequency analysis of the wheel speed sensors. For this purpose, the influence of tire pressure is calculated from the frequency spectra to eliminate this interference.

[0007] US 2012 0 245 787 A1 discloses determining vehicle mass from the frequency characteristics of the wheel speed sensors. This vehicle mass is then used to determine tire pressure loss.

[0008] DE 10 2008 056 664 A1 describes a method for indirect tire pressure monitoring, in which a pressure loss detection variable is determined based on the analysis of the vibration behavior of a wheel using the wheel speed signal.

[0009] The invention is based on the object of providing a method by means of which an estimate of the load change of a vehicle can be determined.

[0010] This object is achieved according to the invention in that at least one tire pressure-insensitive parameter for the current load of the vehicle is determined from the wheel speed signals.

[0011] Advantageous embodiments of the invention are the subject of the subclaims.

[0012] The invention is based on the idea that indirect tire pressure monitoring provides important information about the vehicle, as it allows existing or impending tire damage to be detected and increases driver safety if the tire pressure is too low or too high. However, indirect detection of tire pressure can lead to false detections if other influences affect the wheel speed signal. One of these influences is typically a change in the vehicle load. This should therefore be detected if possible to avoid false detections of the tire pressure. If the problem of tire load is considered independently of indirect tire pressure monitoring, an estimate of the tire load per wheel, i.e. for each individual wheel, has not yet been made in known systems.As has now been recognized, the change in load of a vehicle can be determined using parameters derived from the respective wheel speed signals. These parameters can be selected so that they provide reliable information about the change in load and are not influenced by changes in tire pressure. On the other hand, the determined change in load can be advantageously used to improve indirect tire pressure monitoring or to avoid false detections.

[0013] Advantageously, the frequency spectrum is determined from the respective wheel speed signal, with the characteristic value for the current vehicle load being determined from the frequency spectrum. The frequency spectrum is preferably determined by a Fourier transformation of the wheel speed sensor signal or the wheel speed as a function of time.

[0014] In a preferred embodiment, vehicle tests involving various pressure and load levels create a database on which the frequency spectra of the respective tire vibrations are calculated by Fourier transformation of the respective wheel speed signals. By comparing the values ​​for the different load and...

[0015] The pressure-invariant, load-sensitive spectral range of the frequency spectrum can be determined from the spectra characteristic of pressure scenarios. Following this "offline" determination, this frequency range is used during ongoing operation (i.e., "online") to calculate a tire-pressure-insensitive parameter for the current vehicle load. To theoretically motivate these steps, a rigid-belt tire model coupled to a quarter-vehicle static position serves as the basis. Its transfer function, formed as the quotient of the model's vibration behavior and the road excitation, approximately corresponds to a realistic frequency spectrum of the tire vibration.

[0016] Preferably, at least one environmental influence and / or one vehicle influence is compensated for in the respective characteristic or loading characteristic, from which a compensated characteristic is determined. An example of an influence of the vehicle is the vehicle speed or the engine torque, which can change the vibration spectrum indirectly, i.e., by changing the tire properties (e.g., sidewall stiffness), or directly, i.e., by exciting other vibration modes (e.g., less torsional and more vertical excitation).

[0017] The respective parameter is preferably used directly or indirectly via derived variables to determine the influence of the load change on variables derived from the rolling circumference effect.

[0018] According to the invention, the respective parameter is used directly or indirectly via derived variables to compensate for the influence of the change in loading on variables derived from the rolling circumference effect.

[0019] The influence of the load change on the rolling circumference effect derived in this way is advantageously used to verify the plausibility of pressure loss situations and / or to verify the plausibility of the position detection of a pressure loss.

[0020] The derived influence of the load change on the rolling circumference effect is preferably used to verify the plausibility of the estimated values ​​of other load estimation tools, in particular acceleration sensors.

[0021] From the at least one tire pressure insensitive parameter (λ j) In a preferred embodiment, the displacement of the center of gravity in the vehicle plane is determined for the current load. The vehicle plane refers to a plane that runs parallel to the vehicle surface when the vehicle's surface is level.

[0022] The determined shift of the center of gravity is advantageously provided to a vehicle dynamics control system, in particular ESP or TCS.

[0023] The advantages of the invention lie in particular in the fact that by directly estimating the effects of load changes on rolling circumference dimensions and the resulting variables, knowledge of the load condition can be obtained without additional sensors or measurements. Furthermore, the position of the load can be determined. The post-processing of other physical effects for vehicle mass estimation (e.g., from high-resolution acceleration signals) can be verified. The indirect tire pressure monitoring system is thus independent of high-resolution longitudinal acceleration sensors or the torque signal of the engine control unit, thus enabling indirect tire pressure monitoring functionality in projects where the latter are not available or only available to a limited extent.

[0024] The method can also be used outside of tire pressure monitoring, specifically to determine a shift in the vehicle's center of gravity in the vehicle's xy plane, caused, for example, by a change in load. This information is useful for adjusting the control of vehicle dynamics control systems such as ESP.

[0025] An embodiment of the invention is explained in more detail with reference to a drawing, which shows, in a highly schematic representation: Fig. 1 two exemplary representations of a tire with different tire pressure; Fig. 2 an exemplary spectrum of a wheel speed sensor signal; Fig. 3 a motor vehicle with four wheels; Fig. 4 is a representation of the rigid belt model of a quarter-vehicle pneumatic tire with a body, a belt and a rim; Fig. 5 a diagram of relevant quantities of the rigid belt model according to Fig. 4; Fig. 6 shows an exemplary diagram of a transfer function as a function of a frequency spectrum obtained from a wheel speed signal; Fig. 7 shows an exemplary diagram of two transfer functions as a function of a frequency spectrum obtained from a wheel speed signal for two different tire pressures and an enlarged section thereof; Fig. 8 shows an exemplary diagram of a transfer function as a function of a frequency spectrum obtained from a wheel speed signal during a load change and an enlarged section thereof; Fig. 9 an exemplary classification value for different states of a vehicle; Fig. 10 an exemplary calibration phase of two process variables; Fig. 11 three exemplary process variables for different vehicle conditions; Fig. 12 a diagram of an exemplary determination of a process variable; Fig. 13 a vehicle in an xy-plane; and Fig. 14 shows an example of the relationship between two process variables.

[0026] Identical parts are provided with the same reference numerals in all figures.

[0027] A pneumatic car tire 20 is shown schematically in Fig. 1 shown on the left side at target pressure p soll and on the right side at lower inflation pressure, with unchanged load. The tire 20 with outer radius r (unloaded, pressure p soll ) springs in the rigid belt model under load with a force F Z = M B g by the height ρ = z T - z D one, where z is T the current displacement of the tire belt (see Fig. 5) and z Dis the effective height of the road profile. One full rotation of the tire after compression corresponds to the path of a tire with rolling circumference radius re=r−η(zr−zD) with a tire-dependent factor η and the current vertical displacement of the rim z R (see Fig. 5). Equations are designated by a number in parentheses, which is then used to refer to this equation in the text. For a constant load, the rolling circumference is determined by the spring deflection (z R - z D ) pressure-dependent. By comparing the rotation times T (available from the wheel speeds of the brake control unit) i = 2πr e,i / v Kfz the wheels of a vehicle (v Kfzis the vehicle speed), this results in three pressure-dependent measured variables. In general, in indirect tire pressure monitoring, after performing the reset and setting the target pressure on all wheels, a reference value is averaged over time at different speeds within a predefined time using the equations below and stored (calibration). Once calibration is complete, a single-wheel pressure loss can be detected, e.g., via the change in a variable Axle after a detection threshold is exceeded (detection). From the four rotation times T i For example, the three dimensions Axle, Diag, and Side can be determined as follows: Axle=T1+T2T3+T4−1Diag=T1+T4T2+T3−1Side=T1+T3T2+T4−1

[0028] In Fig. 3 schematically shows a vehicle 6 with four tires numbered 1, 2, 3 and 4. In the above formulas, for example, T1 = 2π r e,1 / vKfz d the orbital time of wheel 1 (in the Fig. 3 left front). The same applies to wheels 2, 3, and 4.

[0029] Independent of the rolling circumference radius r e A second pressure-dependent criterion is a frequency effect. The tire belt and the rim, when excited by the road surface, oscillate both against each other and in phase around the rotation axis. The oscillations of the rim lead to the modulation of the yaw rate signals and can be represented as a spectrum using a Fourier transformation. Fig. 2, a curve 8 shows the spectrum at a first tire pressure value p1 and a curve 10 shows the spectrum at a second, lower tire pressure value p2, ie p1 > p2. The spectrum is dominated, among other things, by the natural frequencies f p1 and f p2, whose absolute value depends on the material (tire), pressure and external factors (e.g. speed). Analogous to the rolling circumference, f p1 and f p2 during the calibration phase after setting the target pressure and stored. In case of pressure loss, the two natural frequencies shift to lower values ​​(indicated by the two arrows in Fig. 2 indicated).

[0030] After compensation of the external influencing factors, a single-wheel or multi-wheel pressure loss can be detected individually for each wheel by a deviation of the currently determined frequencies from the learned values.

[0031] A disadvantage of conventional indirect tire pressure monitoring is the susceptibility to false warnings due to load changes. A change in load F z (loading) of a tire leads to a tire-dependent change in the rolling circumference radius r via the equations given above. eand thus, for example, via equation (2) for Axle to change a criterion ΔAxle = Axle aktuell - Axle Kalibrierung compared to a learned or calibrated value. Due to availability and systematic errors in compensating for environmental influences, a plausibility check of the rolling circumference effect using the frequency effect is not possible without error. In the worst case, a load change is interpreted as a pressure loss, and a false pressure loss warning is issued.

[0032] In the event of a change in load and simultaneous pressure loss, an indirect tire pressure monitoring system may not be able to warn due to competing effects, e.g., via the ΔAxle criterion, and may not detect the pressure loss.

[0033] The method according to the invention aims at compensating the loading effect on the rolling circumference, thereby preventing false warnings and restoring the reliable warning capability of an indirect tire pressure monitoring system even under load changes.

[0034] The frequency effect of a pneumatic tire is determined using a Fig. 4, coupled to a quarter-vehicle static test, is discussed. A pneumatic tire 20 is modeled by a rigid belt 26 with mass m. T and moment of inertia J T , which is connected via a spring-damper system in torsional and radial direction to a rim 32 with mass m R and moment of inertia J R is coupled, where c t and cr Spring constants and d t and d r Damping. The interface of the tire 20 with a road surface 36 with height z D (t) forms the tire contact patch, modulated by a longitudinal stiffness c p and relaxation length σ. The deflection ρ = z T (t) - z D (t) of the tire 20 under load F N is approximately described by a vertical stiffness over FN=cvertρ .

[0035] The connection of the wheel or tire 20 to a body 42 of mass m B is carried out via a spring-damper system in the x- and z-direction. The point A on the tire belt with radius r is determined by translation and rotation of the rim around x R , e.g. R or φ R and subsequent translation and rotation of the belt by x T , e.g. T or φ Tto point C. Ω is the angular velocity around the co-moving origin.

[0036] Based on Fig. 5, the equations of motion of tire 20 are discussed. Tire 20 lies in the zx-plane of a coordinate system. The equations of motion are derived from the consideration of the motion of a belt point A, which is to be transferred to point C by rotation and translation of rim 32 or belt 26 (via B) (see Nojek, Rafal: Insufficient Pressure and Damage Detection for Tires with Supporting Inserts, Dissertation, Technical University of Koszalin, 2010). mBx¨B=FBX mBz¨B=FBZ mRz¨R=FRX−FBx mRz¨R=FRZ−FBZ IRϕ¨R=Mt mTx¨T=−FRX+Fƒcos a−Fnsin a mTx¨T=−FRX+Fƒsin a−Fncos a ITϕ¨T=−Mt−Fƒre−Fnreƒrr σF˙ƒ=−vbFƒ−cκvs FBX=cX(xR−xB)+dX(x˙R−x˙B) FBZ=cZ(zR−eg)+dZ(z˙R−z˙B) FRX=πr(ct+cr)(xT−xR)+πr(dt+dr)(x˙T−x˙R) −πr(dt+dr)(Ω+ϕR)(zT−zR) FRZ=πr(ct+cr)(zT−zR)+πr(dt+dr)(z˙T−z˙R) +πr(dt+dr)(Ω+ϕ˙)(xT−xR) Mt=2πdtr3(ϕ¨T−ϕ¨R)+2πctr3(ϕT−ϕR) Fn=cvert(zD−zT) re=r−η(zR−zD)+Δr σ=12l cκ=2cpl2 vS=x˙T−vb=x˙T−re(Ω+ϕ˙T) frr=fR0

[0037] The vibration behavior φ' R (t) of the rim 32 can be determined independently of the road excitation z D (t) for small displacements, for this purpose the above system of equations is linearized. After Laplace transformation, the transfer function can be Θφ'RzD(s)=φ'^R(s) / z^D(s) represent graphically, with s = i 2πf, see Fig. 6, in which a curve 48 shows the absolute value of the transfer function. The frequency is shown on the x-axis 52, while the absolute value of the transfer function is plotted on the y-axis 56. A natural frequency analysis of the above system of equations (4) allows the visible resonances to be identified with vibration modes. In addition to a translational, in Fig. 6, the in-phase natural vibration of rim 32 and belt 26, designated E1, shows the torsional excitations with (E4) and without (E2) phase shift of rim 32 and belt 26, as well as a vertical vibration E3.

[0038] In Fig. Figure 7 shows the absolute value of the transfer function of a tire (225 / 45 R17) under pressure change. Curve 60 represents the absolute value of the transfer function at a tire pressure p = 2.5 Pa, and curve 66 represents the absolute value of the transfer function at a tire pressure of 1.5 Pa, where Pa represents the pressure unit Pascal.

[0039] The natural frequencies show a trend towards smaller absolute values ​​when pressure drops. Within the discussed model, a strong dependence on the pressure-dependent stiffnesses c t and c r and the vertical stiffness c vert , the longitudinal stiffness c p and the relaxation length σ (tire contact patch), which were used to model the pressure loss. The shift of the natural frequencies, see also Fig. 2, can serve as a detection criterion for a pressure loss.

[0040] In a box 70, the two curves 60, 66 are shown enlarged in a selected frequency range of the frequency f from 36 to 49 Hz. The two curves 60, 66 are almost identical in this frequency range, thus the spectrum shows an extremely low sensitivity to tire inflation pressure in this frequency range.

[0041] In Fig. Figure 8 again shows the magnitude of the transfer function as a function of frequency. Curve 76 shows the magnitude of the transfer function at a tire pressure of 2.5 Pa. Curve 80 shows the magnitude of the transfer function at a load change of approximately 100 kg. The load change was modeled using a mass change Δm B , a change in the vertical stiffness c vert , the longitudinal stiffness c p and the relaxation length σ (tire footprint) as well as a slight increase in the stiffness c t and c r .

[0042] Box 86 shows the two curves 76 and 80 for a frequency range between 36 and 49 Hz. The two curves differ significantly from each other in this frequency range.

[0043] The method according to the invention makes use of this behavior of the transfer function. A comparison of the Fig. 7 and Fig. Figure 8 shows approximately pressure-invariant regions in the transfer function, here, for example, at 35 Hz < f < 50 Hz (see Box 70), ie, the change in tire pressure, e.g., due to pressure loss, leaves the spectrum unchanged at this point. On the other hand, according to Fig. 8 amplitudes of this range under increasing load, so the spectrum is load-sensitive at this point and can be used as a loading criterion.

[0044] The spectrum of the tire vibrations generated by Fourier transformation of the realistic wheel speed signals on the vehicle still contains the effect of the road excitation (surface z D ), which is explicitly calculated in the model analysis by using the transfer function Θφ'RzD(s) The average road influence can be calculated by estimating the car power spectrum of the road SZDZD(f)=σr2aπvKFZa2v2KFZ+(2πf)2 with speed-dependent parameters a, σ r The realistic car performance spectrum of the wheel speed signals (rim) on the vehicle is then Sφ'Rφ'R(f)=|Θφ'RzD(f)|SzDzD(f).

[0045] In the Fig. 9 is the mean of the amplitudes |Sφ'Rφ'R(f)|¯ of the frequency range selected in boxes 70 and 86 for the various loading cases and a pressure loss with the same surface excitation (“smooth”). A first point 90 corresponds to the case of an unloaded vehicle and a target tire pressure p Soll . A second point 94 corresponds to the case of a loaded vehicle and the target tire pressure p Soll . A third point 98 corresponds to the case of an unloaded vehicle and a pressure loss in the tire, ie the tire pressure is below the target tire pressure. A fourth point 102 corresponds to the case of an unloaded vehicle with the target tire pressure, but a rough road surface. The different loading levels are clearly distinguishable. It is also evident in the case of different road roughness that the criterion λ=|Sφ'Rφ'R(f)|¯ still has a road dependency.

[0046] If we now consider the individual criteria of the 4 wheel positions, λj=|Sφ'Rφ'R(f)|¯j the tire pressure-insensitive and load-sensitive parameters of the procedure, it can be approximately assumed |Sφ'Rφ'R(f)|¯j≈|Θφ'RzD(f)|JSzDzD(f¯) for an average frequency f and the road dependency can be determined, for example, by forming the axle-wise quotient Λ=λ1+λ2λ3+λ4−1 be eliminated. The difference ΔΛ(m)=Λcurrent−Λcalib, so that a change in loading compared to the calibration leads to a non-zero difference ΔΛ. The method in a preferred embodiment can now determine the change in the loading state as follows. For this purpose, the detection criterion Axle is considered as an example, and the corresponding loading criterion from equation (8). For the diagonal and side-by-side pressure loss criteria, Diag and Side, respectively, a diagonal or side-by-side loading criterion can be used analogously to equation (8).

[0047] The detection criterion ΔAxle = Axle aktuell - Axle Kalibrierung is implicitly dependent on the load via Eq. (1) ΔAxle(m) = ΔAxle(Λ(m)). The exact functional relationship ΔAxle model(Λ) can be determined offline by determining the characteristic curve ΔAxle(m) vs. A(m) depending on the axle load m. Thus, the loading influence can be explicitly compensated when the axle load in the vehicle changes by (see also Fig. 11) ΔAxlecomp=ΔAxle(m)−ΔAxlemodel(Λ).

[0048] In Fig. 11, in the upper diagram, the value ΔAxle is represented by a curve 130 at a load condition of 700 kg, by a curve 134 in the unloaded condition and by a curve 138 in the unloaded condition with pressure loss at the rear axle.

[0049] In the middle diagram, the value ΔΛ is shown by curves 142, 146, 150 for these three vehicle conditions (“load of 700 kg”, “unloaded”, “unloaded with pressure loss on the rear axle”), in the lower diagram the value ΔAxle komp through the three curves 154, 158, 162.

[0050] After compensating for the loading effect, ΔAxle komp instead of ΔAxle and the indirect tire pressure control via rolling circumference is now fully capable of warning even when the load changes.

[0051] Without the exact knowledge of the relationship ΔAxle model (Λ), ΔAxle can also be checked for plausibility using the difference ΔΛ in the detection phase. If a limit value ΔΛ thresh a change in load can be assumed. A possible pressure loss warning should then be delayed until the pressure loss criteria of the frequency effect, f p1 and f p2 , provide a clear picture. This prevents load-induced false warnings.

[0052] The method in a preferred embodiment is discussed below. In a first step, measurements are performed on the vehicle with load changes. A functional relationship between ΔAxle(m) and ΔΛ(m) is established, which is in particular essentially linear, with both variables depending on the load m. From this, the model variable ΔAxle is calculated. model (Λ) is determined.

[0053] During a calibration phase, a synchronous determination of the averaged values ​​of Λ and Axle is carried out, this is exemplified in Fig. 10 shown.

[0054] The average value of A is Λ Kalib , the average value of Axle is Axle kalib .

[0055] The following process steps are carried out during the life cycle or during operation of the vehicle.

[0056] The quantity ΔΛ = Λ aktuell - Λ kalibThe quantity ΔAxel kompl = (Axle aktuell - Axle kalib ) - ΔAxle modell (ΔΛ). The value of ΔAxel kompl is now used in an indirect tire pressure control so that the load change can be taken into account.

[0057] The described preferred embodiment of the method is schematically illustrated again with the aid of the Fig. 11 and Fig. 12, as an example for the detection criterion Axle, for the pressure loss criteria, Diag or Side, the procedure can be analogous: During calibration runs, the calibration values ​​for Axle and Λ are determined from the wheel speed signals and then the respective differences ΔAxle and ΔΛ are calculated in tests with load changes.

[0058] The ΔAxle model model (Λ) can then be determined offline, see Fig. Figure 12, which presents the cumulative data from several tests with load changes, here using the example of a load change from 200 kg to 700 kg and variable tires. In this example, a linear relationship is suggested.

[0059] After implementation, ΔAxle model (Λ) can be used for the load compensation of ΔAxle. Fig. Figure 11 shows this for a calibration (700 kg payload) and two subsequent vehicle tests at variable speed and road profiles. After calibration, the payload was removed, and the differences in ΔAxle and ΛΔ that occur in the subsequent test (“test without load”) due to the change in load are clearly visible. After compensation to ΔAxle komp The latter can be used in DDS+ post-processing, i.e. in indirect tire pressure control with post-processing, as a pressure loss criterion, the minimal deviation of which indicates normal pressure.

[0060] The case of load changes and modified pressures on the rear axle is in the second test case (“test unloaded, HA pressure loss”, Fig. 11). After compensation, ΔAxle komp a clear deviation from the (loaded) calibration case, which leads to a pressure loss warning after passing through the corresponding DDS+ warning path.

[0061] The method includes non-parametric spectral method implementations: By Fourier transforming the wheel speed signal, the pressure-invariant, load-sensitive part of the frequency spectrum of the pneumatic tire vibration can be determined. From this, an indicator variable, λ, for load changes on the vehicle can be derived.

[0062] The method also includes parametric spectral method implementations: An approximation for the transfer function Θφ'RzD(s) of equation (5) (or for the corresponding system for φ R) allows the formulation of a differential equation for φ R or φ' R' whose coefficients can be estimated recursively from the wheel speed signals. This allows the estimation of the transfer function Θφ'RzD(s) (or an approximation thereof). From the latter, the pressure-invariant, load-sensitive spectral component can be determined and the procedure can be analogous to the case of non-parametric spectral methods.

[0063] The method can be used to determine environmental influences on the above-mentioned loading indicator λ i (tire position i). This includes the quotient formation to A according to equation (7), analogously for diagonal and sidewise, or pairwise quotients, compare equation (2), and the subsequent compensation Λ komp = f Λ (λ,v kfz, e,...) with a model function f that can be determined offline by vehicle tests Λ, where e is the road excitation estimated from the non-parametric or parametric spectral method and "..." allows for further influencing variables, such as the engine torque. Likewise, the λ i directly compensate to λ komp = f λ (λ,v kfz, e, ...) .

[0064] In a preferred embodiment of the method, a shift in the center of gravity of the vehicle in the xy-plane is detected, wherein the xy-plane runs essentially parallel to the roadway or the ground, as is the case in Fig. 13. The following procedural steps are carried out. In vehicle tests with load changes (shifting, payload), the wheel loads m1, m2, m 3, m4 of the four wheel positions, see Fig. 13, measured under load change. The following conditions are defined MAxle=m1+m2m3+m4−1=:Mm−1 and MSide=m1+m3m2+m4−1.

[0065] Using equation (7), the loading indicator Λ Axle (and analogously Λ side ) is calculated and recorded online in the vehicle for the various loading cases.

[0066] A compensation model Λ is set up offline Axle,komp = f(Λ, v Kfz, e) (analogous for Λ Side,komp ) to eliminate environmental influences. The characteristic curve Λ Axle,Side,komp as a function of M Axle,Side be determined, see Fig. 14, ΛAxle,model(MAxle).

[0067] The model equation (11) is inverted for later estimation of the displacement of the vehicle's center of gravity in the x-direction M Axle = Λ -1 Axle,model (Λ) and implemented (analogously for M side ). In vehicle tests, the measurement of Λ Axle,komp so on the ratio M Axel (see equation (10)).

[0068] The values ​​M determined in this way Axle , MSide can be made available to other functional models (AYC, TCS). Alternatively, if the (symmetric) center distance to the coordinate origin, Δx, is known, see Fig. 13, directly the approximate displacement of the vehicle’s center of gravity in the x-direction via RCoG,x=|Δx|(Mm−1Mm+1)=|Δx|(MAxleMAxle+2) be determined (analog R CoG,x about M Side ).

Claims

[1] Method for determining the loading condition of a vehicle, whereby the wheel speeds of the vehicle’s wheels are measured using wheel speed sensors, where from the wheel speed signals at least one tire pressure insensitive parameter (λ i ) is determined for the current load of the vehicle characterized by , that the respective parameter (λ i , λ komp,i ) the mean value of the amplitudes |Sφ'Rφ'R(f)| a selected pressure-invariant and load-sensitive frequency range between 36 and 49 Hz of a transfer function as a function of a frequency spectrum obtained from a wheel speed signal and is used directly or indirectly via derived variables to determine the influence of the load change on variables derived from the rolling circumference effect (Axle, Diag, Side). [2] Method according to claim 1, wherein the frequency spectrum is determined from the respective wheel speed signal, and wherein the characteristic value (λ i ) is determined for the current load of the vehicle. [3] Method according to claim 1 or 2, wherein at least one environmental influence and / or an influence of the vehicle in the respective parameter (λ i ), resulting in a compensated parameter (λ komp,i ) is determined. [4] Method according to one of claims 1 to 3, wherein the respective parameter (λ i , λ komp,i ) is used directly or indirectly via derived variables to compensate for the influence of the load change on variables derived from the rolling circumference effect (Axle, Diag, Side). [5] Method according to one of claims 1 to 4, wherein the derived influence of the change in load on the rolling circumference effect is used to check the plausibility of pressure loss situations and / or to check the plausibility of the position detection of a pressure loss. [6] Method according to one of claims 1 to 5, wherein the derived influence of the load change on the rolling circumference effect is used to check the plausibility of the estimated values ​​of other load estimation means, in particular acceleration sensors. [7] Method according to one of claims 1 to 6, wherein the at least one tire pressure-insensitive parameter (ë i ) the displacement of the center of gravity in the vehicle plane is determined for the current load. [8] Method according to claim 7, wherein the determined displacement of the center of gravity is provided to a vehicle dynamics control system, in particular ESP or TCS.

Images