Method and apparatus for determining a filter output parameter of a filter for filtering a torsion bar torque of a steer-by-wire steering system for a vehicle

Abstract

Method (300) for determining a filter output variable (104) of a filter (100) for filtering a torsion bar torque of a steer-by-wire steering system of a vehicle depending on a filter input variable (102) to be filtered, wherein a predetermined filter property of the filter (100) is selected from a plurality of predetermined filter properties (108) depending on an input signal characteristic of the filter input variable (102) (304), wherein the input signal characteristic is determined depending on the filter input variable (102) at a current and at least one previous time point (302), and wherein the filter output variable (104) is determined by the filter (100) with the filter property (306).

Description

State of the art

[0001] The invention relates to a method and a device for determining a filter output parameter of a filter for filtering a torsion bar torque of a steer-by-wire steering system for a vehicle. The invention also relates to a vehicle with such a device.

[0002] In steer-by-wire steering systems, a steering handle receives the driver's directional input, and a hand torque actuator generates steering feedback for the driver. To create this feedback, the hand torque actuator includes a motor that delivers torque to a rotor shaft, which is then transmitted to the steering handle via a gearbox and steering column. The motor is positioned relatively close to the driver, resulting in noticeable audible and tactile feedback. This feedback becomes disruptive and unacceptable when the motor is driven by highly noisy signals. A common approach to handling noisy signals is to use low-pass filters to compensate for high-frequency noise, but this introduces an additional phase shift into the steer-by-wire steering system.Therefore, a compromise must be made regarding noise filtering and phase delay in order to avoid impairing the stability characteristics of the steer-by-wire steering system due to the additional phase delay.

[0003] Therefore, a filter is desirable that reduces noise without introducing a large phase delay and keeps filter error within a defined range.

[0004] From EP 2 576 321 B1, for example, a method and a device for determining a filter output variable of a filter of a steering system depending on a filter input variable to be filtered are known, wherein a predetermined filter property of the filter is selected from a plurality of predetermined filter properties by means of a filter characteristic determination unit depending on an input signal characteristic of the filter input variable, and wherein the filter output variable is determined by the filter with the filter property.

[0005] Similar methods and devices are also known from DE 10 2019 105 702 A1, DE 601 21 181 T2 and DE 102 09 610 A1. Disclosure of the invention

[0006] This is achieved by a method, a device and a vehicle according to the independent claims.

[0007] The method provides that, to determine a filter output variable of a filter for filtering a torsion bar torque of a vehicle's steer-by-wire steering system, depending on a filter input variable to be filtered, a predefined filter property of the filter is selected from a plurality of predefined filter properties based on an input signal characteristic of the filter input variable. The input signal characteristic is determined based on the filter input variable at a current and at least one previous time point. The filter output variable is determined by the filter with the selected filter property. The different filter properties of the filter influence the filter error. This makes it possible to limit the filter error to a maximum value.In addition, the filter operates in a favorable operating state for the input signal, thereby reducing phase delay and improving the properties of a control loop of the steer-by-wire steering system.

[0008] Preferably, the majority of the filter properties comprise a first filter property, a second filter property, and a third filter property, wherein: the first filter property sets the filter output signal to a midpoint, in particular a midpoint of a noise band, of the filter input signal; the second filter property operates the filter as a low-pass filter; and the third filter property operates the filter as a pass-through filter. These filter properties allow the filter to operate in different states. This makes it possible to compensate for noise in certain scenarios, such as a stationary steering wheel. This results in improved acoustic and haptic characteristics of the steer-by-wire steering system.

[0009] Preferably, the filter's predefined filter properties are parameterized by at least one application parameter. This application parameter makes it possible, for example, to limit the filter error to a specific value or to adjust individual filter properties. This allows the filter to be adapted to different operating conditions or different types of steer-by-wire steering systems.

[0010] Preferably, the at least one application parameter for the specified filter property characterizes a low-pass filter coefficient and / or a maximum permissible filter error.

[0011] Preferably, the input signal characteristic is determined depending on at least one application parameter. This allows the limits of the operational states and their operating ranges to be adjusted, thus improving the flexibility of the filter.

[0012] Preferably, the at least one application parameter for the input signal characteristic of the filter input variable characterizes a width of a noise band of the filter input variable, in particular in the form of quantization noise of the filter input variable and / or a rate of change of the filter input variable.

[0013] Preferably, it is provided that if the filter input variable for the input signal characteristic: remains within the width of the noise band, a first operating state of the filter is determined, whereby the first filter property for the filter is determined; shows a tendency to leave the width of the noise band, a second operating state of the filter is determined, whereby the second filter property for the filter is determined; shows a tendency to enter the width of the noise band, a third operating state of the filter is determined, whereby the second filter property for the filter is determined; is outside the width of the noise band and continuously moves away from it, a fourth operating state of the filter is determined, whereby the third filter property for the filter is determined;When the device is outside the width of the noise band and begins to approach it, a fifth operating state of the filter is determined, defining the second filter property. This gives the filter these operating states and their defined operating ranges. This allows the filter to be implemented efficiently and with minimal resource consumption in embedded software. This also has a positive impact on the filter's runtime.

[0014] In a vehicle, the system can be designed to measure the torsion bar torque, determine the filter input variable to be filtered based on the measured torsion bar torque, and determine a motor torque for the steer-by-wire steering system based on the filter output variable and a target torque for the torsion bar torque. This allows the torsion bar torque control of the steer-by-wire steering system to be adapted to the situation and improved overall.

[0015] The device for determining the filter output of a filter for filtering a torsion bar torque of a vehicle's steer-by-wire steering system, depending on a filter input, is configured to select a predefined filter property from a plurality of predefined filter properties based on the input signal characteristic of the filter input. It determines the input signal characteristic at a current and at least one previous time point, and then determines the filter output through the filter with the selected filter property. The different filter properties allow for influencing the filter error, thereby limiting it to a maximum value.Furthermore, it is possible for the filter to operate in a state favorable to the input signal, thereby reducing phase delay and improving the properties of a control loop of the steer-by-wire steering system.

[0016] Preferably, the filter features comprise a first filter feature, a second filter feature, and a third filter feature. The device is configured to use the first filter feature to shift the filter output signal to a midpoint, particularly a midpoint of a noise band, of the filter input signal; to operate the filter as a low-pass filter using the second filter feature; and to operate the filter as a pass-through filter using the third filter feature. These filter features allow the filter to operate in different states. This makes it possible to compensate for noise in certain scenarios, such as a stationary steering wheel. This results in improved acoustic and haptic characteristics of the steer-by-wire steering system.

[0017] Preferably, the device is designed to parameterize the filter's predetermined filter properties based on at least one application parameter. This application parameter makes it possible, for example, to limit the filter error to a specific value or to adjust individual filter properties. This allows the filter to be adapted to different operating conditions or different types of steer-by-wire steering systems.

[0018] Preferably, at least one application parameter for the specified filter property characterizes a low-pass filter coefficient and / or a maximum filter error.

[0019] Preferably, the device is designed to determine the input signal characteristics depending on at least one application parameter. This makes it possible to adjust the limits of the operational states and their operating ranges, thus improving the flexibility of the filter.

[0020] Preferably, the at least one application parameter for the input signal characteristic of the filter input variable characterizes a width of a noise band of the filter input variable, in particular in the form of quantization noise of the filter input variable and / or a rate of change of the filter input variable.

[0021] Preferably, the device is designed to determine, when the filter input parameter for the input signal characteristic remains within the width of the noise band, a first operating state of the filter, as well as the first filter property for the filter, shows a tendency to leave the width of the noise band; a second operating state of the filter, as well as the second filter property for the filter, shows a tendency to enter the width of the noise band; a third operating state of the filter, as well as the second filter property for the filter, is located outside the width of the noise band and is continuously moving away from it; a fourth operating state of the filter, as well as the third filter property for the filter, is located outside the width of the noise band and is beginning to approach the noise band.To determine a fifth operational state of the filter, as well as the second filter property, the filter must be defined. This gives the filter these operational states and their defined operating ranges. This allows the filter to be implemented efficiently and with minimal resource consumption in embedded software. This also has a positive impact on the filter's runtime.

[0022] The device can improve the steer-by-wire function in a vehicle. In this case, the vehicle comprises the device, which is configured to measure the torsion bar torque, determine the filter input variable to be filtered depending on the measured torsion bar torque, and determine a motor torque for the steer-by-wire steering system depending on the filter output variable and a target torque for the torsion bar torque.

[0023] Further advantageous embodiments will become apparent from the following description and the drawing. The drawing shows: Fig. 1 a schematic representation of a filter, Fig. 2 a schematic representation of a state graph of the filter, Fig. 3. A flowchart of a procedure for determining a filter output variable of the filter, Fig. Figure 4 shows a schematic representation of a vehicle. Fig. 5 represents a control loop.

[0024] Fig. Figure 1 shows a filter 100 in a schematic representation. A filter input variable 102 is provided to the filter 100. Depending on the filter input variable 102, the filter 100 determines a filter output variable 104. The filter input variable 102 is also fed to an evaluation unit 106. In this example, the filter 100 includes the evaluation unit 106. The evaluation unit 106 may be arranged as a separate unit from the filter 100. The filter 100 also has a plurality of filter properties 108. The plurality of filter properties 108 includes a first filter property 110, a second filter property 112, and a third filter property 114. The plurality of filter properties 108 may comprise a lower or higher number of filter properties.The evaluation unit 106 determines, based on the input signal characteristic of the filter input variable 102, a filter property from the plurality of filter properties 108, which is used to process the filter input variable 102 by the filter 100. The filter output variable 104 is determined by the filter 100 based on the filter input variable 102 and the determined filter property of the filter 100. The input signal characteristic is determined by the evaluation unit 106 based on the filter input variable 102 at a current time and at least one previous time.

[0025] Filter 100 operates with the first filter property 110 when the filter input 102 alternates around a midpoint and between two quantization values. The upper and lower quantization values ​​are limits for a noise band and define its width. The first filter property 110, when the filter input 102 lies within the noise band, causes the filter output 104 to be set to the midpoint of the noise band. Therefore, with the first filter property 110, filter 100 ensures that the filter output 104 remains at a piecewise constant value and does not exhibit the alternating behavior of the filter input 102.

[0026] Filter 100 is operated with the second filter property 112 when the filter input 102 exhibits dynamic behavior. The filter input 102 remains close to the midpoint of the noise band but shows a tendency to either leave or enter the noise band. In this case, the second filter property 112 causes filter 100 to exhibit first-order low-pass behavior.

[0027] Filter 100 is operated with the third filter property 114 when the filter input 102 moves steadily away from the noise band and exhibits highly dynamic behavior. In this case, the third filter property 114 causes filter 100 to behave like a pass-through filter. With the third filter property 114, filter 100 directly switches the filter input 102 to the filter input 102.

[0028] It may also be provided that the majority of filter properties include 108 further properties of signal processing structures or functions, such as a bandpass property, an inverter property, or differently configured filter properties of the same category.

[0029] Fig. Figure 2 shows a schematic representation of a state graph 200 for the filter 100. The state graph includes operation states 202, 204, 206, 208 and 210. These operation states 202 to 210 are represented by the filter 100 and the majority of filter properties 108.

[0030] In the first operating state 202, the filter input signal 102 remains in the noise band and the filter 100 is operated with the first filter property 110. The filter output value 104 is therefore set to the middle value of the noise band.

[0031] In the second operating state 204, the filter input signal 102 shows a tendency to leave the noise band, and the filter 100 is operated with the second filter property 112. The filter output signal 104 is therefore the low-pass filtered filter input signal 102.

[0032] In the third operating state 206, the filter input signal 102 shows a tendency to enter the noise band, and the filter 100 is operated with the second filter property 112. The filter output signal 104 is therefore the low-pass filtered filter input signal 102.

[0033] In the fourth operating state 208, the filter input signal 102 is outside the noise band and is steadily moving away from it, and the filter 100 is operated with the third filter property 114. The filter input signal 102 is therefore directly connected to the filter output signal 104, and consequently no filtering takes place.

[0034] In the fifth operating state 210, the filter input signal 102 no longer exhibits the dynamic behavior of the fourth operating state 206 and begins to approach the noise band, and the filter 100 is operated with the second filter property 112. The filter output 104 is therefore the low-pass filtered filter input 102.

[0035] Depending on the operating state, the mean value of the noise band is updated accordingly; consequently, the noise band moves with a signal waveform of the filter input quantity 102. This update occurs in operating states 206, 208, and 210. The state graph 200 shows the following properties of the filter 100. In the first operating state 202, a filter error e = |in - out|, calculated by subtracting the magnitude of the filter output quantity 104 from the filter input quantity 102, lies within the noise band. The filter error e is defined by the width of the noise band in the first operating state 202.In this example, the first operating state 202 can only be reached by starting from the third operating state 206 and the fifth operating state 210, in which the mean value of the noise band is updated to the current value of the filter output 104. This allows the input signal 102 to remain as close as possible to the center of the noise band in the first operating state 202. It can be provided that the width of the noise band is used as an application parameter of the filter 100. In the fourth operating state 208, the filter input 102 is set to the filter output 104, so the filter error e = 0. In the second operating state 204, the third operating state 206, and the fifth operating state 210, a check for a deviation between the filter input 102 and the filter output 104 is additionally performed. If this deviation is greater than a defined maximum filter error e... m, so the filter output size 104 will have a defined maximum filter error e m adapted. By appropriately selecting the coefficient of the low-pass filter property of the second filter property 112, it is possible for the filter error e to remain within the maximum filter error e m remains. In this case, checking the deviation is not necessary.

[0036] The following is an example of the structure and algorithm of filter 100. This structure is based on the state graph from Fig. 2 and the operational states 202 to 210 and shows, among other things, the transition conditions of the individual operational states 202 to 210 as well as an implementation of the filter properties 110 to 114. The setup can be implemented, for example, in embedded software.

[0037] Variables and parameters: - Filter input size 102: 1. u(k), u(k-1): the current or previous sample value of the filter input variable to be filtered 102. - Filter output size 104: 1. y(k), y(k-1): the current and previous sample values ​​of the filter output variable 104, respectively. - Internally defined application parameters: 1. ε : the width of the noise band 2. β ∈ (1,3) : Application parameter for a slowly varying filter input variable 102 3. a ∈ (0,1) : 1st order low-pass filter coefficient 4. e m : the permissible maximum filter error 5. d : Application parameter that is greater than a maximum difference in the filter input size in one sampling step. - Internal variables: 1. m(k), m(k-1): the current or previous mean value of the noise band. 2. D(k), D(k-1): the difference between the current or previous sample value of the filter input quantity 102 and the previous mean value of the noise band. Consequently, D(k) = u(k) - m(k-1) and D(k-1) = u(k-1) - m(k-1), respectively.

[0038] The transition conditions to the corresponding operational states 202 to 210 and the events triggered accordingly: - Operational status 202: 1. Conditions: |D(k-1)| < ε and |D(k)| < ε 2. Events: y(k) = y(k-1), then m(k) = m(k-1) - Operational status 204: 1. Conditions: (|D(k-1)| ≤ ε and |D(k)| ≥ ε) or (|D(k)| > ε and |D(k-1)| ≤ β*ε and ||D(k)| - |D(k-1)|| ≤ ε) 2. Events: y(k) = a*u(k)+(1- a)*y(k-1)+sgn(y(k-1)-u(k))*min(e m -(1-a)*|y(k-1)-u(k)|,0), where sgn(.) is the sign function and min(.) takes the smaller value of the two expressions, then m(k) = m(k-1) - Operational status 206: 1. Conditions: |D(k-1)|≥ε and |D(k)|<ε 2. Events: y(k) = a*u(k)+(1-a)*y(k-1)+sgn(y(k-1)-u(k))*min(e m -(1-a)*|y(k-1)-u(k)|,0), then m(k) = y(k) - Operational status 208 1. Conditions: (|D(k-1)|≤|D(k)| and |D(k-1)|> β*ε) or (|D(k-1)|> ε and |D(k)|-|D(k-1)|>ε) 2. Events: y(k) = u(k), then m(k) = u(k)-sgn(D(k))* d - Operational state 210: 1. Conditions: (|D(k-1)|>|D(k)| and |D(k)|≥ε and |D(k-1)|>β*ε) or (|D(k-1)|-|D(k)|>ε and |D(k)|2ε) 2. Events: y(k) = a*u(k)+(1-a)*y(k-1)+sgn(y(k-1)-u(k))*min(e m -(1-a)*|y(k-1)-u(k)|,0), then m(k) = y(k).

[0039] Filter 100 is defined by the application parameters e, β, a, e mand d parameterized. The application parameters of the noise band width ε and the application parameter β are used as parameters for determining the input signal characteristic of the filter input variable 102. The application parameter a is used as the coefficient of the low-pass filter property of the second filter property 112, thereby parameterizing this filter property. The application parameter e m The maximum filter error is also used as an application parameter of the second filter property 112. The application parameter d is used as a parameter of the third filter property 114 to update the mean value of the noise band. The filter 100 is adapted to different scenarios and application areas using the application parameters.

[0040] Fig. Figure 3 represents a flowchart 300 of a procedure for determining a filter output variable 104 of a filter 100. In step 302, the filter input variable 102 is provided to the filter 100 and the input signal characteristic is determined.

[0041] The input signal characteristic is determined depending on the filter input parameter 102 at a current and at least one previous time.

[0042] It may be provided that the input signal characteristic is determined depending on at least one application parameter, for example the width of the noise band ε and / or an application parameter for a rate of change of the filter input parameter 102.

[0043] In step 304, depending on the input signal characteristic of the filter input quantity 102, a predefined filter property of the filter 100 is selected from the majority of predefined filter properties 108.

[0044] Preferably, the majority of the filter features 108 comprise the first filter feature 110, the second filter feature 112, and the third filter feature 114, wherein the first filter feature 110 of the filter 100 sets the filter output 104 to the midpoint, in particular to the midpoint of a noise band, of the filter input 102. The second filter feature 112 operates the filter 100 as a low-pass filter. The third filter feature 114 operates the filter 100 as a pass-through filter.

[0045] Preferably, the specified filter property of filter 100 is parameterized by at least one application parameter. It is possible that this at least one application parameter is a low-pass filter coefficient and / or a maximum permissible filter error.

[0046] It can be provided that if the filter input parameter 102 remains within the width of the noise band for the input signal characteristic, the first operating state 202 of the filter 100 is determined, with the filter 100 operating with the first filter property 110. If the filter input parameter 102 shows a tendency to leave the width of the noise band for the input signal characteristic, the second operating state 204 of the filter 100 is determined, with the filter operating with the second filter property 112. If the filter input parameter 102 shows a tendency to enter the width of the noise band for the input signal characteristic, the third operating state 206 of the filter 100 is determined, with the filter operating with the second filter property 112.If the filter input parameter 102 is outside the width of the noise band for the input signal characteristic and continuously moves away from it, the fourth operating state 208 of the filter 100 is determined, with the filter operating with the third filter property 114. If the filter input parameter 102 is outside the width of the noise band for the input signal characteristic and begins to approach the noise band, the fifth operating state 210 of the filter 100 is determined, with the filter 100 operating with the second filter property 112.

[0047] In step 306, the filter output size 104 is determined by the filter 100 with the specified filter property.

[0048] The filter 100 is used, for example, in a steer-by-wire steering system to filter a torsion bar torque as filter input size 102.

[0049] In this example, filter 100 is implemented as a torsion bar torque filter within the context of a manual torque control system for a steer-by-wire steering system. Filter 100 is used, for example, to smooth or filter a torsion bar torque that is noisy due to quantization.

[0050] In Fig. Figure 4 shows a vehicle 400 with a steer-by-wire steering system comprising a steering system 402 and a hand torque actuator 404.

[0051] In this example, the vehicle 400 comprises two rear wheels 406 and two front wheels 408. The rear wheels 406 are not steerable in this example. The front wheels 408 are steerable by the steering mechanism 402.

[0052] In addition to the front wheels 408, or instead of the front wheels 408, the rear wheels 406 can also be steered by means of a steering mechanism of the rear axle.

[0053] The hand torque actuator 404 receives directional input from the driver and generates a steering feel for the driver. To generate this steering feel, the hand torque actuator 404 contains a motor 410 which transmits a corresponding motor torque to a rotor shaft 412. This torque is then transmitted via a gearbox 414 and a torsion bar 416 to a steering handle 418. A steering wheel is shown as an example of the steering handle 418.

[0054] Due to the fact that the Motor 410 is located relatively close to the driver, this can lead to noticeable acoustic and haptic feedback. This feedback can even become disturbing and unacceptable if the Motor 410 is controlled by highly noisy signals.

[0055] The filter 100 and the filtering process significantly improve the acoustic and haptic behavior without significantly impairing the remaining control quality.

[0056] In Fig. Figure 5 shows a control loop with a controller 502 and a filter 100 for controlling the manual torque adjuster 404. Control with the controller 502 is based on an actual torque 504, in this example the torsion bar torque at the torsion bar 416, and a predefined target torque 506, in this example the desired torsion bar torque.

[0057] The controller 502 determines a suitable motor torque 508 to regulate the target torque 504 even in the presence of disturbances 510, e.g., driver intervention or model deviations. As in Fig. As shown in Figure 5, the filter input signal 102 is based on the actual moment 504 and is fundamentally noisy 512.

[0058] The example specifically considers quantization noise. While higher-quality measurement technology may reduce the influence of noise, it can never completely compensate for it. Controller 502 could itself be designed to account for the influence of noise 512. However, this typically means that the control characteristics with respect to other criteria, such as performance, are negatively affected, potentially creating a conflict of objectives.

[0059] A common approach to handling noisy signals is the use of appropriate filters. A conventional filter has a low-pass characteristic to compensate for the high-frequency noise components, but this introduces an additional phase shift into the system. Therefore, a further compromise must be made regarding noise filtering and phase delay to prevent the additional phase from significantly degrading the system's stability.

[0060] To significantly reduce quantization noise without introducing a large phase delay, and also to limit filter error within a defined range, the case-dependent filter 100 is used. In this example, the filter output 104 is a feedback signal for the controller 502. The controller 502 uses this feedback signal and the setpoint torque 506 to determine the motor torque 508.

[0061] Typically, when the steering wheel of the steer-by-wire system is stationary, i.e., when the torsion bar torque is not changing significantly, filter 100 operates with the first filter characteristic 110. This characteristic applies when the torsion bar torque, i.e., the filter input variable 102, alternates around a mean value and between two quantization values. If the characteristic of the torsion bar torque changes, the input characteristic at filter 100 also changes. This switches the filter characteristics based on the input signal characteristic as described. The filter output variable 104 is thereby adjusted situationally.

Claims

[1] Method (300) for determining a filter output variable (104) of a filter (100) for filtering a torsion bar torque of a steer-by-wire steering system of a vehicle depending on a filter input variable (102) to be filtered, wherein a predetermined filter property of the filter (100) is selected from a plurality of predetermined filter properties (108) depending on an input signal characteristic of the filter input variable (102) (304), wherein the input signal characteristic is determined depending on the filter input variable (102) at a current and at least one previous time point (302), and wherein the filter output variable (104) is determined by the filter (100) with the filter property (306). [2] Method (300) according to claim 1, characterized by , that the majority of the filter properties (108) comprise a first filter property (110), a second filter property (112) and a third filter property (114), wherein: the first filter property (110) of the filter (100) the filter output variable (104) is placed on a midpoint, in particular on a midpoint of a noise band, the filter input variable (102); the second filter property (112) of the filter (100), the filter (100) is operated as a low-pass filter; the third filter property (114) of the filter (100), the filter (100) is operated as a through-pass filter. [3] Method (300) according to any one of the preceding claims, characterized by , that the specified filter property of the filter (100) is parameterized depending on at least one application parameter. [4] Method (300) according to claim 3, characterized by , that at least one application parameter for the specified filter property characterizes a low-pass filter coefficient and / or a maximum permissible filter error. [5] Method (300) according to any one of the preceding claims, characterized by, that the input signal characteristic is determined depending on at least one application parameter. [6] Method (300) according to claim 5, characterized by , that the application parameter for the input signal characteristic of the filter input quantity (102) characterizes a width of a noise band of the filter input quantity (102), in particular in the form of a quantization noise of the filter input quantity (102) and / or a rate of change of the filter input quantity (102). [7] Method (300) according to claim 6, characterized by , that if the filter input size (102) is for the input signal characteristic: in the width of the noise band, a first operational state (202) of the filter (100) is determined, whereby the first filter property (110) for the filter (100) is determined; a tendency to leave the width of the noise band is shown, a second operational state (204) of the filter (100) is determined, whereby the second filter property (112) for the filter (100) is determined; a tendency to enter the width of the noise band, a third operational state (206) is determined of the filter (100), whereby the second filter property (112) for the filter (100) is determined; is located outside the width of the noise band and continuously moves away from it, a fourth operational state (208) of the filter (100) is determined, whereby the third filter property (114) for the filter (100) is determined; when the filter is outside the width of the noise band and begins to approach the noise band, a fifth operational state (210) of the filter (100) is determined, whereby the second filter property (112) for the filter (100) is determined. [8] Method according to any one of the preceding claims, characterized by , that the torsion bar torque (504) is measured, the filter input variable (102) to be filtered is determined depending on the measured torsion bar torque (504), and depending on the filter output variable (104) and a target torque (506) for the torsion bar torque, a motor torque (508) for the steer-by-wire steering system is determined. [9] Device for determining a filter output variable (104) of a filter (100) for filtering a torsion bar torque of a steer-by-wire steering system of a vehicle depending on a filter input variable (102) to be filtered, wherein the device is designed to: to select a predefined filter property of the filter depending on an input signal characteristic of the filter input quantity (102) from a plurality of predefined filter properties (108); to determine the input signal characteristic depending on the filter input variable (102) at a current and at least one previous time point; to determine the filter output variable (104) by the filter (104) with the filter property. [10] Device according to claim 9, characterized by , that the plurality of filter properties (108) comprises a first filter property (110), a second filter property (112) and a third filter property (114), wherein the device is designed to: by means of the first filter property (110) of the filter (100) to place the filter output quantity (104) on a midpoint, in particular on a midpoint of a noise band, of the filter input quantity (102); by means of the second filter property (112) of the filter (100), to operate the filter (100) as a low-pass filter; by means of the third filter property (114), to operate the filter (100) as a through-element. [11] Device according to one of claims 9 and 10, characterized by, that the device is designed to parameterize the specified filter property of the filter (100) depending on at least one application parameter. [12] Device according to any one of claims 9 to 11, characterized by , that at least one application parameter for the specified filter property characterizes a low-pass filter coefficient and / or a maximum filter error. [13] Device according to any one of claims 9 to 12, characterized by that the device is designed to determine the input signal characteristics depending on at least one application parameter. [14] Device according to claim 13, characterized by, that at least one application parameter for the input signal characteristic of the filter input quantity (102) characterizes a width of a noise band of the filter input quantity (102), in particular in the form of a quantization noise of the filter input quantity (102) and / or a rate of change of the filter input quantity (102). [15] Device according to any one of claims 9 to 14, characterized by , that the device is designed to ensure that when the filter input parameter (102) is used for the input signal characteristic: remaining in the width of the noise band, to determine a first operational state (202) of the filter (100), as well as to determine the first filter property (110) for the filter (100); a tendency to leave the width of the noise band, to determine a second operational state (204) of the filter (100), and to determine the second filter property (112) for the filter (100); a tendency to enter the width of the noise band, to determine a third operational state (206) of the filter (100), and to determine the second filter property (112) for the filter (100); is located outside the width of the noise band and continuously moves away from it, to determine a fourth operational state (208) of the filter (100), and to determine the third filter property (114) for the filter (100); is located outside the width of the noise band and begins to approach the noise band, to determine a fifth operating state (210) of the filter (100), as well as to determine the second filter property (112) for the filter (100). [16] vehicle, characterized by, that the vehicle comprises the device according to one of claims 9 to 15, wherein the device is configured to measure the torsion bar torque (504), to determine the filter input parameter (102) to be filtered depending on the measured torsion bar torque (504), and to determine a motor torque (508) for the steer-by-wire steering system depending on the filter output parameter (104) and a target torque (506) for the torsion bar torque (504).

Images