Method of processing a measurement signal

Abstract

A method for processing a measurement signal, in particular of a steering system, is presented. The method comprises the steps of obtaining a measurement variable on the basis of the measurement signal, wherein the measurement variable comprises information about a physical variable and wherein the measurement variable is a superposition of an actual value of the physical variable and a measurement noise. Filter parameters of a filter are determined on the basis of the measurement variable and a mathematical model of the measurement noise. The measurement signal is filtered by means of the filter, whereby an estimated value of the physical variable is obtained, wherein the filter has the determined filter parameters. The filter parameters are determined in such a way that a deviation between the estimated value of the physical variable and the actual value of the physical variable is approximated and minimized. Furthermore, a control unit for a steering system, a steering system, a computer program and a computer-readable data carrier are disclosed.

Description

Technical Field

[0001] This invention relates to a method for processing measurement signals. The invention also relates to a control unit for a steering system of a motor vehicle, a steering system for a motor vehicle, a computer program, and a computer-readable data carrier. Background Technology

[0002] In summary, the electromechanical steering system of a motor vehicle is designed to detect the torque applied by the driver on the steering wheel and, based on that torque, provide a matching auxiliary torque via an electric motor, which assists the driver in steering the motor vehicle.

[0003] Steering systems typically include sensors that acquire, for example, torque and / or steering angle, and convert these physical variables into electrical signals (measurement signals) for further processing. Due to this measurement process, in addition to information about the physical variables (the so-called useful signal component), the measurement signals also contain unwanted measurement noise.

[0004] The measured signal is used, for example, to determine the auxiliary torque via a controller or regulator. The auxiliary torque is then provided via an electric motor. According to a typical method, a specific frequency component in the measured signal is amplified to generate the auxiliary torque. However, this amplifies the measurement noise present in the measured signal along with the useful signal, which can potentially lead to unwanted noise in the steering system.

[0005] Therefore, in the prior art, measurement signals are filtered using low-pass filters, band-pass filters, or similar filters with time-constant transmission characteristics to reduce unwanted measurement noise. However, if the frequency ranges of the useful signal to be filtered and the measurement noise overlap, the filtering may impair steering feel. Furthermore, the use of measurement signals in the control loop poses a risk of degrading the robustness and stability of the control loop when using this type of filtering. Summary of the Invention

[0006] Therefore, the object of the present invention is to provide an alternative method for processing measurement signals, wherein measurement noise is reliably reduced across the entire frequency range without altering the frequency components of the useful signal.

[0007] According to the present invention, this objective is achieved by a method for processing measurement signals, particularly those of steering systems. The method includes the following steps: acquiring measurement variables based on the measurement signals, wherein the measurement variables include information about physical variables, and wherein the measurement variables are a superposition of the actual values ​​of the physical variables and measurement noise. Determining filter parameters for a filter based on a mathematical model of the measurement variables and the measurement noise. Filtering the measurement signals using the filter to obtain estimates of the physical variables, wherein the filter has the determined filter parameters. The filter parameters are determined such that the deviation between the estimated values ​​and the actual values ​​of the physical variables is approximated and minimized.

[0008] Physical variables are preferably useful signals, especially useful signals from the steering system.

[0009] Compared to existing technologies, the measurement signal is not simply filtered by a filter with time-constant transmission characteristics to reduce measurement noise. Instead, according to the method of the present invention, a filter with time-varying filter parameters is used to filter the measurement signal.

[0010] Here, filter parameters are determined at each data acquisition point to filter out as much measurement noise as possible without also filtering out the components of the actual values ​​of the physical variables. To this end, the deviation between the estimated and actual values ​​of the physical variables is minimized. Specifically, the quadratic deviation between the estimated and actual values ​​of the physical variables is minimized.

[0011] Measurement noise can be filtered out using a mathematical model of the noise characteristics, which allows the filter parameters to be intentionally adjusted for measurement noise at each data acquisition step.

[0012] Therefore, by means of the method according to the invention, measurement noise is filtered out over the entire relevant frequency range without altering the useful components. This maintains the stability and robustness of the control loop using the measurement signal. In the steering system, this also means that the steering feel remains unaffected.

[0013] The filter is preferably a filter with a finite impulse response. Such filters are also called "finite impulse response (FIR) filters". The advantage of such FIR filters is that they process measurement signals particularly quickly, i.e., they have low latency.

[0014] However, filters with an infinite impulse response can also be considered. Such filters are also called "infinite impulse response (IIR) filters".

[0015] One aspect of the invention provides that the physical variable is steering column torque and / or the measured variable is a measurement by a torque sensor. Herein and hereinafter, "steering column torque" should be understood as the torque acting in the steering column. Therefore, the steering system may include a torque sensor coupled to the steering column and designed to measure the torque acting in the steering column.

[0016] Alternatively or additionally, the physical variables can be the steering angle and / or steering angle rate. Therefore, the measured variable can be the value measured by the steering angle sensor.

[0017] At least one actuator of the steering system is preferably controlled based on estimates of physical variables. Therefore, the actuator is controlled based on the most probable estimate of the actual values ​​of the physical variables.

[0018] For example, the electric motor of the steering system is controlled based on estimates of physical variables. In particular, the electric motor is an auxiliary motor of the steering system, which is designed to generate auxiliary torque to assist the driver when steering the motor vehicle.

[0019] In another embodiment of the invention, the measurement noise in the mathematical model is modeled as a Gaussian process. Here, the measurement noise can be described sufficiently accurately using knowledge of the process average and process autocorrelation.

[0020] According to another embodiment of the invention, in the mathematical model, the measurement noise is uncorrelated with the physical variables. This means that the cross-correlation between the physical variables and the measurement noise is zero at any point in time. Therefore, the measurement noise can be decoupled from the physical variables in the mathematical model.

[0021] In another embodiment of the invention, the measurement noise is modeled as white noise in the mathematical model. Therefore, at least within the relevant frequency range, the spectral power density of the measurement noise is substantially constant, and in particular, constant. Furthermore, the measurement noise at one point in time is uncorrelated with the measurement noise at any other point in time.

[0022] Another aspect of the invention provides that the measurement noise is modeled in a mathematical model as time-independent or having a known time-series dependence. More precisely, in the mathematical model, the characteristic statistical variables of the measurement noise are constant over time, or their time-series dependence is known. In particular, the mean and / or autocorrelation of the measurement noise are constant over time or their time-series variation is known.

[0023] Preferably, the filter parameters are determined recursively, particularly by a least-squares recursive method. In other words, instead of re-determining the filter parameters using stored data from previous data acquisition steps at each data acquisition step, the filter parameters are recursively determined based on previous filter parameters and newly acquired data. This saves processing time and / or resources.

[0024] In one embodiment of the invention, the measured values ​​are considered only within a predefined time window, according to the filter parameters. Therefore, data points of the measured variables from too far in the past are no longer considered. The filter parameters are thus adaptively adjusted to account for any possible time-varying nature of the physical variables.

[0025] According to the invention, this objective is also achieved by a control unit for a steering system of a motor vehicle, wherein the control unit is designed to perform the above-described method.

[0026] Specifically, the physical variable is the steering column torque and / or the measured variable is the value measured by the torque sensor of the steering system.

[0027] The above explanations, which relate to the advantages and characteristics of control units, also apply to control units, and vice versa.

[0028] According to the invention, this objective is also achieved by a steering system for a motor vehicle. The steering system includes a sensor designed to acquire measured variables, including information about physical variables. Furthermore, the steering system includes the aforementioned control unit.

[0029] Specifically, the physical variable is the steering column torque and / or the measured variable is the value measured by the torque sensor of the steering system. The sensor can be designed accordingly as a torque sensor.

[0030] The above explanations, which relate to the advantages and characteristics of steering systems, also apply to steering systems, and vice versa.

[0031] According to the present invention, this objective is also achieved by a computer program having program code means so as to execute the steps of the above-described method when the computer program is executed on a computer or a corresponding processing unit, especially the processing unit of the aforementioned control unit.

[0032] "Program code device" is understood herein and hereinafter as computer-executable instructions in compiled and / or uncompiled form as program code and / or program code modules, which are available in any programming language and / or machine language.

[0033] The above explanations, which are relevant to the methods used to describe the advantages and characteristics of computer programs, also apply to computer programs, and vice versa.

[0034] According to the present invention, this objective is also achieved by a computer-readable data carrier having the aforementioned computer program stored thereon.

[0035] The above explanations, which relate to the advantages and characteristics of computer-readable data carriers, also apply to computer-readable data carriers and vice versa. Attached Figure Description

[0036] Further advantages and features of the invention will become apparent from the following description and accompanying drawings, with reference to these drawings. In the drawings:

[0037] - Figure 1 A steering system according to the present invention is schematically illustrated;

[0038] - Figure 2 A flowchart illustrating a method for processing measurement signals according to the present invention is shown schematically; and

[0039] - Figure 3 A diagram illustrating the function of a computer program according to the present invention is shown. Detailed Implementation

[0040] Figure 1 A steering system 10 for a motor vehicle is schematically shown, wherein the steering system 10 is implemented as an electromechanical assisted steering system (column drive EPS) with steering column auxiliary equipment.

[0041] The steering system 10 includes a steering wheel 12, which is connected to a pinion 18 via the upper portion of a steering column 14 and through a steering intermediate shaft 16. The pinion 18 meshes with a rack 20 such that a torque is applied to the rack when the driver rotates the steering wheel 12.

[0042] A torque and / or steering angle sensor 22 is arranged on the steering column 14. This torque and / or steering angle sensor is designed to measure steering torque and / or steering angle. In particular, this sensor is therefore a steering torque and steering angle sensor, also known as a "torque and angle sensor (TAS)," and can provide steering angle in addition to steering torque.

[0043] In addition, an electric motor 24 is provided, which is connected to the steering intermediate shaft 16 via a gear 26 to transmit torque.

[0044] Gear 26 Figure 1 The gear 26 is designed as a worm gear. However, alternatively, gear 26 can be designed as a spur gear, bevel gear, or any other suitable type of gear.

[0045] In any case, the torque provided by at least the electric motor 24 is transmitted to the steering intermediate shaft 16 via the gear 26 to perform steering assistance.

[0046] Electric motor 24 and torque and / or steering angle sensor 22 (only when Figure 1 (The diagram illustrates that each of the components is connected to the control unit 28 of the steering system 10 by means of transmitting signals.)

[0047] Generally, the control unit 28 is designed to determine the torque to be applied based on measurement data from the steering system 10, especially measurement data from the torque and / or steering angle sensor 22, and to transmit the corresponding control command to the electric motor 24, so that the electric motor 24 provides the torque to be applied.

[0048] It should be noted that the steering system 10 with steering column assistance described above is merely an example for illustration. The following explanation (which may have been slightly modified) also applies to any other type of steering system, especially steering systems with pinion drive (single pinion drive EPS), steering systems with dual pinion drive (dual pinion drive EPS), steering systems with concentric rack drive via recirculating ball nuts, steering systems with belt drive, and so-called steer-by-wire steering systems (where there is no mechanically operated connection between the steering wheel 12 and the wheels of the motor vehicle).

[0049] Generally speaking, all different types of steering systems with electromechanical steering assistance devices share the following characteristics: the control unit 28 detects the torque applied by the driver to the steering wheel 12 and controls the electric motor 24 based on the torque to generate a specific auxiliary torque.

[0050] In this case, specific frequency components in the measurement signal are typically amplified to produce a specific steering feel. However, the measurement noise present in the measurement signal is also amplified at the same time, which may, for example, lead to unwanted noise in the steering system 10.

[0051] Noise measurement depends on frequency and can be characterized by the so-called signal-to-noise ratio (SNR). More precisely, the SNR at a specific frequency ω can be expressed by the actual physical variable P. phyVal Spectral power density of (ω) and interference P noisw It is characterized by the ratio of the spectral power density of (ω), that is, by...

[0052]

[0053] To characterize.

[0054] To reliably reduce measurement noise across the entire relevant frequency range, control unit 28 is designed to perform the following based on Figure 2 and Figure 3 Explanation of the methods and steps.

[0055] More precisely, the control unit 28 includes a processing unit 30 and a data carrier 32, wherein a computer program is stored on the data carrier 32, the computer program is executed on the processing unit 30 and includes program code means to cause the steering system 10 to perform the method steps explained below.

[0056] First, a measurement signal is generated by the torque and / or steering angle sensor 22 (step S1). Depending on the embodiment of the torque and / or steering angle sensor 22, the measurement signal contains information about the torque acting in the steering column 14 and / or information about the rotation angle of the steering column 14.

[0057] The measurement signal is relayed to the control unit 28, where it is further processed. The control unit acquires the measurement variable based on the measurement signal (step S2).

[0058] In the following text, the measured variable is the torque T acting in the steering column 14. column,meas Let's take an example to illustrate. Therefore, the basic physical variable is the actual torque T acting in the steering column 14. column In this case, the measured torque is the actual torque T. column and measurement noise v meas The superposition of these factors makes the following formula applicable.

[0059] T column,meas =T column +v meas .

[0060] Measured variable T column,meas The torque can be measured directly by the torque sensor 22. For example, the torque T acting on the steering column 14 can be determined by the torsion angle of the torsion bar of the steering column 14. column,meas .

[0061] Now based on the measured torque T column,meas and measurement noise v meas The mathematical model is used to determine the filter parameters of the filter. p (Step S3)

[0062] The filter in the exemplary embodiments described below is a filter with a finite impulse response.

[0063] However, filters with infinite impulse response can also be considered.

[0064] Here and below, the variables indicated by underscores are vector-valued variables.

[0065] Generally speaking, the measured torque T is filtered by a filter. column,meas After filtering, the physical variable T was obtained. column The estimated value

[0066] Therefore, in step S3, the filter parameters are determined. p The method is to make the estimated value With physical variable T column The deviations should be as small as possible, as will be explained in more detail below.

[0067] At least one of the following assumptions, and in particular all of them, forms the basis of the mathematical model:

[0068] i) Measurement noise can be described accurately enough using a Gaussian process. That is, knowledge of the process mean and the autocorrelation function of the process is sufficient to characterize the measurement noise.

[0069] ii) Physical variable T column Actual value and measurement noise v meas Uncorrelated, i.e., cross-correlation function E{T column v meas} equals zero.

[0070] iii) The characteristic statistical variables of the measurement noise are constant over time, or their time-series dependencies are known. In particular, the mean and / or autocorrelation function of the measurement noise are constant over time or their time-series dependencies are known.

[0071] iv) Physical variable T column The actual value of is time-varying, and therefore can change over time. Specifically, the physical variable T column The spectral power density is also time-varying.

[0072] exist Figure 3 The function of the computer program is shown in more detail, and the determination of the filter parameters is demonstrated more precisely. It can be seen that the computer program includes a filter module 34 and an update module 36.

[0073] Update module 36 receives the measured variable T column,meas And based on the measurement noise v meas The mathematical model determines the updated filter parameters of filter module 34. p .

[0074] Filter module 34 consists of the filters described above, which have a parameter vector dependent on the filter parameters. p The impulse response h of n parameters filt ( p ), where n is a natural number greater than zero. The filter is given at time point t≥0 for the measured variable T via the following convolution. column,meas The time response y(t) of the stimulus

[0075] y(t)=h filt ( p )*T column,meas

[0076] This describes the filtering of the measured variables via filter module 34.

[0077] By using filter module 34, the measured variable T can be used more accurately with filters. column,meas Filtering is performed if the adjusted filter parameters from the update module 36 are used in the filter module 34. p Then the physical variable T can be obtained. column The estimated value This means that, for this purpose, the general time response y(t) of the filter corresponds to an estimate of the physical variable.

[0078] In each data acquisition step, the filter parameters of the filter are adjusted by the update module 36. p T =[p0,p1,…,p n-1 ].

[0079] Filter parameters p The adjustment is achieved by determining a set of filter parameters that minimizes the following quality function. p To carry out

[0080]

[0081] This depends on the estimate. With physical variable T column The deviation between them.

[0082] estimated value With physical variable T column The variance of the error between them is used as the quality function J.

[0083] The mass function J can be understood as the filter parameters. p The function is therefore J = J(p0, p1, ..., p n Determine filter parameters. pThis makes the mass function J take its minimum value.

[0084] For this purpose, solve the equation And it is usually satisfied as a condition for minimizing the mass function J.

[0085]

[0086] One possible implementation is represented by using a filter with a finite impulse response. The filter can then be specified at any sampling time point t in the processing unit 30 by the following formula. k Time response

[0087]

[0088] vector T column,meas This is given by the following formula.

[0089]

[0090] Where t k-m -t k-m-1 =t s And t k-m :=t k -m·t s , where t s This represents the sampling time of processing unit 30, and m is a natural number greater than or equal to zero.

[0091] When using a filter with a finite impulse response, the solution to the equation... satisfy

[0092]

[0093] The following formula gives a set of filter parameters that satisfy this condition. p

[0094] p =E{ T column，meas T T column，meas} -1 E{ T column，meas T column}

[0095] Filter parameters p It also depends on the actual value T column This actual value cannot be determined through measurement. If we assume a physical variable... T column With measurement noise v meas If they are uncorrelated, then they follow the cross-correlation function.

[0096] E{ T column,meas T column}=E{ T column,meas T column,meas}-E{ v v}.

[0097] Conversely, based on assumptions i), ii), and iii), the autocorrelation function E{ of the measurement noise is known. v v}, such that a set of filter parameters that minimizes the quality function J can be determined by the following equation. p

[0098] p =E{ T column,meas T T column,meas} -1 (E{ T column,meas T column,meas}-E{ v v}).

[0099] Alternatively, this can be achieved by measuring the variable T. column,meas The autocorrelation matrix R TT and measurement variable T column,meas Or the autocorrelation r of noise v TT and r vv To represent, specifically to represent as

[0100]

[0101] First item The second term corresponds to the solution without noise. This indicates the adjustment of the solution due to additive noise.

[0102] Preferably, in each time step, the filter parameters are not determined again based on previously stored measurement signal data using the above equations. p Instead, the filter parameters are calculated recursively, particularly through a least-squares recursive method, based on the above equations. This means determining a new set of filter parameters based on the filter parameters determined in the previous time step and the current values ​​of the measured variables.

[0103] Alternatively, in the above method, the measured variables can only be considered within a predetermined time window, so that data points of the measured variables that are too far in the past are no longer considered. The filter is thus adaptively adjusted, thereby taking into account the time-varying nature of the physical variables.

[0104] After adjusting the filter parameters using the above method, the measured signal is filtered using the filter to obtain the result that is similar to the actual value T. column The estimate with the smallest bias (Step S4)

[0105] Based on the estimated value At least one actuator of the steering system 10 is controlled (step S5). Specifically, based on an estimated value... Control the electric motor 24.

[0106] By means of the above method, measurement noise is reliably reduced across the entire relevant frequency range without adversely affecting the robustness and stability of the steering feel or control loop.

Claims

1. A method for processing measurement signals of a steering system (10), the method comprising the following steps: - Obtain the measurement variable based on the measurement signal, wherein, The measured variable includes information about the physical variable, and wherein the measured variable is the superposition of the actual value of the physical variable and the measurement noise; - Filter parameters are determined based on a mathematical model of the measured variables and the measured noise, wherein, in the mathematical model, the measured noise and the physical variables are uncorrelated; and - The measured signal is filtered by the filter to obtain an estimate of the physical variable, wherein the filter has determined filter parameters; The filter parameters are determined in such a way that the deviation between the estimated value and the actual value of the physical variable is approximated and minimized.

2. The method as described in claim 1, wherein, The filter is a filter with a finite impulse response.

3. The method as described in claim 1 or 2, wherein, The physical variable is the steering column torque and / or the measured variable is the measurement value of the torque sensor (22).

4. The method of claim 3, wherein, At least one actuator (24) of the steering system (10) is controlled based on the estimated value of the physical variables.

5. The method as described in claim 1 or 2, wherein, The measurement noise is modeled as a Gaussian process in the mathematical model.

6. The method as described in claim 1 or 2, wherein, The measurement noise is modeled in the mathematical model as either time-independent or having a known temporal dependence.

7. The method as described in claim 1 or 2, wherein, The filter parameters are determined recursively.

8. The method as claimed in claim 1 or 2, wherein, The measured variables are considered only within a predefined time window to determine the filter parameters.

9. The method of claim 7, wherein, The filter parameters are determined using a recursive least-squares method.

10. A control unit for a steering system (10) of a motor vehicle, wherein, The control unit (28) is designed to perform the method according to any one of claims 1 to 9.

11. A steering system for a motor vehicle, the steering system having a sensor (22) and a control unit (28) according to claim 10, the sensor being configured to acquire a measurement variable including information about a physical variable.

12. A computer program product having program code means for performing the steps of the method as claimed in any one of claims 1 to 9 when the computer program product is executed on a computer or a corresponding processing unit.

13. The computer program product according to claim 12, wherein, The processing unit is the processing unit (30) of the control unit (28) as described in claim 10.

14. A computer-readable data carrier having stored thereon a computer program product as described in claim 12 or 13.

Images