Method and system for controlling a vehicle

Abstract

Method for compensating for disturbances caused by lateral forces in an electromechanical steering system (2) of a vehicle, wherein the vehicle comprises a steering angle controller (3) which has a compensation device (3.1) for compensating for disturbances caused by lateral forces, wherein an estimation device (4) is provided for determining an estimated steering disturbance variable caused by the lateral forces, wherein the estimation device (4) has a Kalman filter (5) which determines the estimated steering disturbance variable based on input information which comprises a model-based calculated steering disturbance variable and at least one measured variable provided by a sensor or derived therefrom, wherein the estimated steering disturbance variable or a variable derived therefrom is made available to the compensation device (3.1) of the steering angle controller (3) as an input signal.

Description

[0001] The invention relates to a method and a system for compensating disturbances in an electromechanical steering system of a vehicle resulting from lateral forces acting on the vehicle. In particular, a method and a system for compensating disturbances are disclosed that estimate a steering disturbance variable using a Kalman filter, which is used in a disturbance compensator of the steering angle controller to reduce the disturbances.

[0002] In the lateral vehicle control functions (lane guidance assistant, lane departure warning, etc.), a steering angle controller is used to influence the lateral vehicle guidance, in particular to keep the vehicle in the center of the lane or to change lanes.

[0003] Steering angle controllers are already known that enable a model-based estimation of a steering disturbance in the area of ​​the steering rod or the steering actuator of the electromechanical steering system in order to be able to compensate for the steering disturbance based on this.

[0004] A disadvantage of the state of the art is that known systems assume that the mathematical model used to determine the steering disturbance completely and correctly represents the vehicle dynamics and therefore does not take into account any inaccuracies in the system.

[0005] The document DE 10 2011 119 462 A1 discloses a method for operating a crosswind assistant in a vehicle, wherein a crosswind compensation intervention can be carried out as a steering intervention.

[0006] The document DE 10 2004 057 928 A1 discloses a method for crosswind stabilization of a vehicle, which uses a Kalman filter to estimate the crosswind force magnitude, the lateral velocity magnitude and the road surface cross slope magnitude.

[0007] The publication DE 10 2004 017 638 A1 describes a method for a vehicle for determining a crosswind force value of a crosswind influence generated by a crosswind acting on the vehicle, in particular for implementing a steering assistance system that compensates for crosswind influences. The vehicle comprises estimation means for estimating the crosswind force value using an observer or Kalman filter based on a linearized lateral dynamics single-track model of the vehicle, in particular based on a lateral acceleration value detected by a lateral acceleration sensor and a yaw rate value detected by a yaw rate sensor.

[0008] Based on this, it is the object of the invention to provide a method for compensating for disturbances in an electromechanical steering system of the vehicle, which method enables an improved estimation of the steering disturbance variable.

[0009] This object is achieved by a method having the features of independent patent claim 1. Preferred embodiments are the subject of the subclaims. A system for compensating for interference in an electromechanical steering system is the subject of independent patent claim 11.

[0010] According to a first aspect, a method for compensating for disturbances in an electromechanical steering system of a vehicle is disclosed. The vehicle comprises a steering angle controller having a compensation device for compensating for disturbances caused by lateral forces. An estimation device is provided for determining an estimated steering disturbance variable caused by the lateral forces. The estimation device has a Kalman filter that determines the estimated steering disturbance variable based on input information comprising a model-based calculated steering disturbance variable and at least one measured variable provided by a sensor or derived therefrom. The estimated steering disturbance variable or a variable derived therefrom is provided as an input signal to the compensation device of the steering angle controller.

[0011] The disclosed method has the technical advantage that, using the Kalman filter, variances or covariances of one or more mathematical models, in particular a mathematical model used to calculate the steering disturbance, and preferably also a mathematical model used to calculate the lateral forces acting on the vehicle's wheels, can be taken into account. This allows model inaccuracies, which may also be dependent on the driving situation, to be taken into account when determining the estimated steering disturbance. This allows for a more accurate estimation of the steering disturbance.

[0012] According to one embodiment, at least two measured variables provided by sensors or derived therefrom are used to determine the estimated steering disturbance, with a first measured variable being the steering angle rate and a second measured variable being the longitudinal acceleration of the vehicle. If these measured variables exceed predetermined or vehicle-specific value ranges, this is an indication that the underlying mathematical model no longer reliably models the actual situation. Thus, the higher inaccuracy of the mathematical model can be taken into account in the Kalman filter by dynamically adapting covariance matrices.

[0013] According to one embodiment, the estimation device receives lateral force information containing information regarding the lateral forces acting on the vehicle as input information. The lateral force information is information calculated using a mathematical model. The lateral force information indicates, in particular, which lateral forces are acting on the vehicle's wheels.

[0014] According to one exemplary embodiment, the lateral force information is calculated using a single-track model and converted into the model-based calculated steering disturbance variable. The model-based calculated steering disturbance variable indicates, in particular, the magnitude of the forces acting on the steering actuator of the electromechanical steering system or the rack of the electromechanical steering system, which are caused by the lateral forces. The lateral force information is converted into the model-based calculated steering disturbance variable, for example, using a factor that specifies the ratio between the lateral force occurring at the vehicle wheel and the force acting on the steering actuator of the electromechanical steering system or the rack.

[0015] According to one embodiment, the Kalman filter comprises one or more covariance matrices that can be adapted depending on the driving situation. This allows for the consideration of driving-situation-dependent uncertainty in the mathematical model, which leads overall to better estimates of the steering disturbance by the Kalman filter.

[0016] According to one embodiment, the Kalman filter comprises at least one steering system covariance matrix that can be adjusted depending on the driving situation. This matrix takes into account the variance of the mathematical model of the steering system and whose values ​​can be adjusted depending on the measured variable provided by the at least one sensor. Preferably, the at least one sensor provides information that characterizes the current driving situation. Thus, it is possible to adjust the entries of the steering system covariance matrix based on the sensor information.

[0017] According to one embodiment, the steering system covariance matrix is ​​adaptable based on the steering angle rate and / or the longitudinal acceleration of the vehicle. In particular, when a threshold value of the steering angle rate and / or a threshold value of the longitudinal acceleration (positive or negative) is exceeded, the errors of the mathematical model of the steering system can increase significantly, so that the reliability of the mathematical model decreases. This can be taken into account accordingly in the steering system covariance matrix based on the information on the steering angle rate and / or the longitudinal acceleration of the vehicle.

[0018] According to one embodiment, the steering system covariance matrix has a variance parameter, wherein the variance parameter is formed by the square of the estimated steering disturbance resulting from the change in the steering angle or by the square of the estimated steering disturbance resulting from the longitudinal acceleration of the vehicle, depending on whether the change in the steering angle or the longitudinal acceleration of the vehicle leads to a higher value of the square of the estimated steering disturbance.

[0019] According to one embodiment, the steering system covariance matrix has entries that represent the inaccuracy of a first model state with respect to the steering angle rate and / or the inaccuracy of a second model state with respect to the steering disturbance.

[0020] According to one embodiment, the Kalman filter comprises at least one measured value covariance matrix that can be adjusted depending on the driving situation and takes into account the variance of at least one sensor providing the measured variables and / or the variance of the model-based calculated lateral forces acting on the vehicle. This allows the inaccuracies of the sensor system or the mathematical model used to calculate the lateral forces acting on the wheels to be taken into account when estimating the steering disturbance.

[0021] According to one embodiment, the measured value covariance matrix, which can be adjusted depending on the driving situation, has entries that indicate a variance of the measured steering angle rate and / or a variance of the estimated lateral forces acting on the vehicle. This allows for the consideration of variances in the measured values ​​or estimated values ​​used to correct the Kalman filter estimates.

[0022] According to a further aspect, the invention relates to a system for compensating for disturbances caused by lateral forces in an electromechanical steering system of a vehicle, wherein the vehicle comprises a steering angle controller having a compensation device for compensating for disturbances caused by lateral forces, wherein an estimation device is provided for determining an estimated steering disturbance variable caused by the lateral forces, wherein the estimation device has a Kalman filter which determines the estimated steering disturbance variable based on input information comprising a model-based calculated steering disturbance variable and at least one measured variable provided by a sensor or derived therefrom, wherein the estimated steering disturbance variable or a variable derived therefrom forms the input signal of the compensation device of the steering angle controller.

[0023] The terms “approximately”, “essentially” or “about” mean, in the sense of the invention, deviations from the exact value by + / - 10%, preferably by + / - 5% and / or deviations in the form of changes that are insignificant for the function.

[0024] Further developments, advantages, and possible applications of the invention will become apparent from the following description of exemplary embodiments and from the figures. All described and / or illustrated features, individually or in any combination, are fundamentally part of the invention, regardless of their summary in the claims or their reference back to them. The content of the claims is also incorporated into the description.

[0025] The invention is explained in more detail below with reference to exemplary embodiments and the figures. They show: Fig. 1 shows, by way of example, a schematic representation of a system for compensating for disturbances caused by lateral forces in an electromechanical steering system; and Fig. 2 shows an example of a schematic representation of the processes in the Kalman filter for determining the estimated steering disturbance.

[0026] Fig. Figure 1 shows an exemplary and schematic illustration of a system 1 for compensating for lateral forces acting on the vehicle. The disturbances that occur due to the lateral forces in the area of ​​the vehicle's steering rod, i.e., the rack to which the steering actuator acts, are estimated, and this estimated disturbance information is fed to a compensation device, allowing direct compensation for the vehicle's steering angle control.

[0027] System 1 uses a Kalman filter 5 to estimate the disturbances occurring at the steering rod due to lateral forces based on a mathematical model that simulates the steering dynamics.

[0028] The Kalman filter 5 also receives measurement information provided by one or more sensors in order to correct the estimates made using the mathematical model. For example, steering angle information provided by a steering angle sensor is used as measurement information. The steering angle information can, in particular, be the steering angle rate (change in the steering angle over time). Furthermore, the Kalman filter 5 uses a model-based calculated steering disturbance variable that depends on the lateral forces acting on the vehicle. The model-based calculated steering disturbance variable can, for example, be a calculated handlebar disturbance variable that arises due to the lateral forces acting on the vehicle. The model-based calculated steering disturbance variable can be calculated from lateral force information provided by a lateral force estimator.The lateral force information can, in particular, be lateral force information acting on the vehicle's wheel. The lateral force estimator can calculate the lateral force information, for example, based on a single-track model and the current steering angle.

[0029] The Kalman filter 5 is designed to weight the predicted and measured states relative to one another using one or more covariance matrices. This means that the contribution to the estimation result made by the information calculated by the mathematical model or the information provided by the at least one sensor can be adjusted depending on the situation. In other words, the contribution made by the measurement results to correcting the estimation result can be varied.

[0030] By means of the Kalman filter 5, it is possible to take into account the dynamics of the steering system by means of a steering system model of the vehicle and the vehicle mass by means of the lateral force information estimated from a single-track model when estimating the steering disturbance (output information of the Kalman filter 5).

[0031] The structure and functionality of the steering disturbance estimation system 1 is described below based on the Fig. 1 is explained in more detail.

[0032] The estimator 4, also referred to as a steering rod disturbance estimator, is coupled to a lateral force estimator 6, from which the estimator 4 receives lateral force information. The lateral force information indicates, in particular, the magnitude of the lateral forces acting on the vehicle's wheels. The lateral force estimator 6 can be formed by a higher-level unit that calculates the lateral force information based on a model (for example, based on a single-track model of the vehicle and based on one or more sensor information items).

[0033] This lateral force information, which indicates the forces in the area of ​​the vehicle's wheels, can then be converted into a steering rod disturbance variable, i.e., the forces occurring in the area of ​​the steering rod due to the lateral forces acting on the vehicle. The conversion is performed using a factor that takes the vehicle's wheel size into account. This calculated steering rod disturbance variable forms an input variable that is used in the Kalman filter 5 for the model-based prediction of the steering disturbance variable.

[0034] To calculate the steering disturbance, in particular the steering rod disturbance, the following mathematical model is used, which models the dynamics of the vehicle's electromechanical steering (EPS: electronic power steering): {θ¨mT˙l}=[−bmJm−1Jm00]{θ˙mTl}+[1Jm 0]{Tm}

[0035] This includes: J m : Equivalent moment of inertia of the EPS motor (kg.m 2 ) b m: Equivalent damping of the EPS motor (N*m / (grad / s)) T l : Steering disturbance at the steering rod (Nm) T m : EPS motor torque (Nm) θ̇ m : Steering angle rate (degrees / s)

[0036] The model is based on the assumption that the first derivative of the steering disturbance at the handlebar Ṫ l is zero.

[0037] The estimation device 4 also receives steering angle information from a steering angle sensor 7 and information about the vehicle speed.

[0038] The system 1 is designed to take into account errors in the model-based prediction of the steering disturbance and errors due to sensor variances or inaccuracies of the model used to calculate the lateral forces occurring on the vehicle by adapting the covariance matrices used in the Kalman filter 5.

[0039] As previously explained, the mathematical model simulating the dynamics of the electromechanical steering assumes that the first derivative of the steering disturbance at the steering rod Ṫ l is zero. However, this assumption is not correct in all driving situations.

[0040] The system is designed to adapt at least one covariance matrix, in particular a steering system covariance matrix, which models the steering dynamics of the electromechanical steering of the vehicle, based on sensor information which indicates that the aforementioned assumption is not fulfilled.

[0041] In particular, in the following driving situations, it may happen that the assumption that the first derivative of the steering disturbance at the steering rod Ṫ l is zero, does not apply: - During steering movements, ie the steering angle rate is not equal to zero; - When accelerating or braking the vehicle while moving in a transverse direction, as this causes the transverse forces to increase or decrease.

[0042] During the above situations, the influence of the changing steering disturbance on the handlebar is determined by the following disturbance covariance matrix: QAsmpt,k=q[Δt33Jm2−Δt22Jm−Δt22JmΔt]

[0043] This includes: q: variance value; Δt: sampling time (sec); J m : Equivalent moment of inertia of the EPS motor (kg.m 2 ).

[0044] The variance value q is preferably determined online, i.e. depending on the situation, by selecting the maximum value from the square value of the change in the steering disturbance due to steering movements or from the square value of the change in the steering disturbance due to acceleration or deceleration. In other words, it is determined whether the change in the steering disturbance is greater due to steering movements or due to acceleration or deceleration of the vehicle, and the greater value is used as the variance value q. The change in the steering disturbance as a function of the steering angle rate can either be calculated using a mathematical function or determined using a look-up table. The same applies to the change in the steering disturbance as a function of the acceleration or deceleration of the vehicle, i.e. the change in the steering disturbance as a function of the acceleration or deceleration.The braking of the vehicle can either be calculated using a mathematical function or determined using a look-up table.

[0045] Mathematically expressed, the variance value q is determined as follows: q=max((f(abs(LWR),P_01))2,(f(abs(a_long),P_02))2)

[0046] This includes: LWR: Steering angle rate a_long: Longitudinal acceleration of the vehicle (positive or negative) P_01: Parameter for the change of the steering disturbance [Nm / s] depending on the absolute value of the steering angle rate; P_02: Parameter for the change of the steering disturbance [Nm / s] depending on the absolute value of the vehicle longitudinal acceleration; abs(x): Absolute value of x.

[0047] The parameters P_01 and P_02 are determined in advance by means of simulations or test drives.

[0048] The final steering system covariance matrix Q f, which is preferably a 2x2 matrix, is determined by the following relationship: Qf=Qtunned+QAsmpt,k

[0049] Where Q Asmpt,k the above disturbance covariance matrix and Q tunned is a matrix by which a system inaccuracy is introduced by setting diagonal values.

[0050] The matrix Q tunned be trained as follows: Qtunned=[Qtunned,1100Qtunned,22]

[0051] The element Q tunned,11 the inaccuracy of the state of the steering angle rate in the model, whereas the element Q tunned,22 indicates the uncertainty of the state of the steering disturbance in the model.

[0052] The values ​​of the elements Q tunned,11 and Q tunned,22 can be determined in a simulation or through test drives.

[0053] In addition, the system 1 can provide a measured value covariance matrix R that can be adapted to the driving situation.f The measured value covariance matrix R f can be determined as follows: Rf=RSensor+RHf,k

[0054] This includes: R Sensor : Matrix indicating the uncertainty of the estimated lateral force information (provided by lateral force estimator 6) and the sensor measurements; R Hf,k : Matrix indicating the variance due to high-frequency jumps in the steering angle sensor signal and high-frequency jumps in the estimated lateral force information;

[0055] The matrix R Sensorreflects the uncertainties caused by the sensors or the lateral force estimator 6. In particular, the lateral force information provided by the lateral force estimator 6 may exhibit driving-situation-dependent uncertainty, since the mathematical model used differs in model quality depending on the driving situation, i.e., in certain driving situations, the model exhibits a greater model error than in others. The following situations, in particular, can lead to greater model inaccuracy: - No strong lateral movement (the linear tire model included in the single-track model is only valid up to lateral accelerations of less than 0.4g, especially 0.3g.); - Stable lateral travel, ie the steering angle rate or the yaw rate is smaller than a specified threshold value.

[0056] The matrix R Sensor can be formed by the following matrix: Rsensor=[Rsensor00Rsensor,22]

[0057] Where R sensor,11 the inaccuracy of the measured steering angle rate, ie R Sensor,11 indicates the variance of the information provided by the steering angle sensor (unit degrees 2 / s 2 ). This information is either provided by the sensor manufacturer or can be determined through simulations or test drives.

[0058] R sensor,22 is the inaccuracy of the estimated lateral force information.

[0059] R sensor,22 can be determined, for example, as follows: Rsensor,22=max(Temp_01,max(Temp_02,Temp_03))

[0060] This is Temp_01=f(abs(vehicle lateral acceleration),P_03) Temp_02=f(abs(vehicle yaw rate),P_04) Temp_03=f(abs(steering angle rate),P_05) P_03: Parameters regarding the variance of the estimated lateral force [Nm 2] depends on the absolute lateral acceleration of the vehicle; P_04: Parameters regarding the variance of the estimated lateral force [Nm 2 ] depends on the absolute yaw rate of the vehicle; P_05: Parameters regarding the variance of the estimated lateral force [Nm 2 ] depends on the absolute steering angle rate of the vehicle;

[0061] The parameters P_03, P_04 and P_05 are determined by simulation or tests, for example during test drives.

[0062] The function f() for determining the values ​​of Temp_01, Temp_02 or Temp_03 can, for example, be a linear interpolation function. For example, Temp_01 is interpolated from parameter P_03 for the absolute lateral acceleration prevailing at a certain point in time. P_03 can, for example, be determined using a lookup table depending on the lateral acceleration. Furthermore, Temp_02 is interpolated from parameter P_04 for the absolute vehicle yaw rate prevailing at a certain point in time. P_04 can, for example, be determined using a lookup table depending on the vehicle yaw rate. Furthermore, Temp_03 is interpolated from parameter P_05 for the absolute steering angle rate prevailing at a certain point in time. P_05 can, for example, be determined using a lookup table depending on the steering angle rate.

[0063] The matrix R Hf,k preferably has the following form: RHf,k=[Var(noise steering angle rate)00Var(noise steering disturbance)]

[0064] The calculation of the matrix entries of the matrix R Hf,k can be done as follows: Raw standard deviation of steering angle rate noise = abs(measured steering angle rate−low-pass filtered steering angle rate) Standard deviation of the steering angle rate noise (Var(Noise Steering Angle Rate)) = Low-pass filtered raw standard deviation of the steering angle rate noise

[0065] The low-pass filter is designed for discrete-time applications as follows: Filter output k = filter coefficient ∗ (filter input k − filter output k − 1) + filter output k − 1

[0066] When filtering the standard deviation of the steering angle rate noise, a coefficient of 0.9 is used for rising edges and 0.3 for falling edges. The resulting signal can thus quickly follow an increase in noise intensity and then slowly decrease again. This slightly overestimates the noise, but this is generally better than underestimating it. Raw standard deviation of the steering disturbance noise = abs(measured steering disturbance − low-pass filtered steering disturbance) Standard deviation of the steering disturbance noise (Var(Steering disturbance noise)) = Low-pass filtered raw standard deviation of the steering disturbance noise;

[0067] The low-pass filter is designed for discrete-time applications as follows: Filter output k = filter coefficient ∗ (filter input k − filter output k − 1) + filter output k − 1

[0068] When filtering the standard deviation of the steering noise, a coefficient of 0.9 is used for rising edges and 0.3 for falling edges. The resulting signal can thus quickly follow an increase in noise intensity and then slowly decrease again. This slightly overestimates the noise, but this is generally better than underestimating it.

[0069] Based on the covariance matrices adapted to the driving situation, the Kalman filter determines the estimated steering disturbance.

[0070] Fig. Figure 2 shows a flow chart of how, based on predictions and measurements used for correction, an iterative estimation of the steering disturbance is performed and error covariance matrices are updated.

[0071] The previously described mathematical steering dynamics model is converted into a discrete-time model. The discrete-time model can be described as follows: {θ˙m,k+1 Tl,k+1}=[(1−bmΔtJm)−ΔtJm01]{θ˙m,kTl,k}+[ΔtJm 0]{Tm,k}

[0072] Here, k specifies the iteration step counter, so that k+1 is the iteration step following k.

[0073] The time-discrete steering dynamics model can be compared with the Fig. 2 shown nomenclature: A=[(1−bmΔtJm)−ΔtJm01]; B=[ΔtJm 0]; xk+1={θ˙m,k+1Tl,k+1}; xk={θ˙m,k Tl,k};

[0074] The discrete-time output matrix is Y=

[1001] {θ˙m,k Tl,k}; H=

[1001] ; Q=Qf; R=Rf;

[0075] The error covariance matrix P kis a symmetric matrix to ensure filter stability. Due to numerical problems, the error covariance matrix P k becomes asymmetrical.

[0076] To ensure the symmetry of the error covariance matrix P k To obtain this, the following measure can be taken: Pk=Pk+PkT2, ie the error covariance matrix P k and their transposes are added and the result is divided by two. This results in a symmetric error covariance matrix P k receive.

[0077] To further prevent destabilization of the filter, a limit on the magnitude of the elements of the error covariance matrix P k Preferably, the elements are limited in each iteration step to avoid resulting calculation errors.

[0078] The invention has been described above using exemplary embodiments. It is understood that numerous changes and modifications are possible without departing from the scope of protection defined by the patent claims. List of reference symbols 1 system 2 electromechanical steering 3 steering angle controllers 3.1 Compensation device 4 Estimating device 5 Kalman filters 6 lateral force estimators 7 Steering angle sensor

Claims

[1] Method for compensating for disturbances caused by lateral forces in an electromechanical steering system (2) of a vehicle, wherein the vehicle comprises a steering angle controller (3) which has a compensation device (3.1) for compensating for disturbances caused by lateral forces, wherein an estimation device (4) is provided for determining an estimated steering disturbance variable caused by the lateral forces, wherein the estimation device (4) has a Kalman filter (5) which determines the estimated steering disturbance variable based on input information which comprises a model-based calculated steering disturbance variable and at least one measured variable provided by a sensor or derived therefrom, wherein the estimated steering disturbance variable or a variable derived therefrom is made available to the compensation device (3.1) of the steering angle controller (3) as an input signal. [2] Method according to claim 1, characterized bythat at least two measured variables provided by sensors or derived therefrom are used to determine the estimated steering disturbance variable, wherein a first measured variable is the steering angle rate and a second measured variable is the longitudinal acceleration of the vehicle. [3] Method according to claim 1 or 2, characterized by that the estimation device (4) receives lateral force information containing information regarding the lateral forces acting on the vehicle as input information, wherein the lateral force information is information calculated by a mathematical model. [4] Method according to claim 3, characterized by that the lateral force information is calculated using a single-track model and converted into the model-based calculated steering disturbance. [5] Method according to one of the preceding claims, characterized bythat the Kalman filter (5) comprises at least one steering system covariance matrix which is adaptable depending on the driving situation and which takes into account the variance of the mathematical model of the steering system and whose values ​​are adaptable depending on the measured variable which is provided by the at least one sensor. [6] Method according to claim 5, characterized by that the steering system covariance matrix is ​​adjustable based on the steering angle rate and / or the longitudinal acceleration of the vehicle. [7] Method according to claim 6, characterized bythat the steering system covariance matrix has a variance parameter, wherein the variance parameter is formed by the square of the estimated steering disturbance resulting from the change in the steering angle or by the square of the estimated steering disturbance resulting from the longitudinal acceleration of the vehicle, depending on whether the change in the steering angle or the longitudinal acceleration of the vehicle leads to a higher value of the square of the estimated steering disturbance. [8] Method according to one of claims 5 to 7, characterized by that the steering system covariance matrix has entries that represent the inaccuracy of a first model state of the steering angle rate and / or the inaccuracy of the second model state of the steering disturbance. [9] Method according to one of the preceding claims, characterized bythat the Kalman filter (5) comprises at least one measured value covariance matrix which can be adapted depending on the driving situation and which takes into account the variance of at least one sensor providing the measured variables and / or the variance of the model-based calculated lateral forces acting on the vehicle. [10] Method according to claim 9, characterized by that the measured value covariance matrix, which can be adapted depending on the driving situation, has entries that indicate a variance of the measured steering angle rate and / or a variance of the estimated lateral forces acting on the vehicle. [11] System for compensating for disturbances caused by lateral forces in an electromechanical steering system (2) of a vehicle, wherein the vehicle comprises a steering angle controller (3) which has a compensation device (3.1) for compensating for disturbances caused by lateral forces, wherein an estimation device (4) is provided for determining an estimated steering disturbance variable caused by the lateral forces, wherein the estimation device (4) has a Kalman filter (5) which determines the estimated steering disturbance variable based on input information which comprises a model-based calculated steering disturbance variable and at least one measured variable provided by a sensor or derived therefrom, wherein the estimated steering disturbance variable or a variable derived therefrom forms the input signal of the compensation device (3.1) of the steering angle controller (3).

Images