EP1891316A1

EP1891316A1 - Method and device for correcting the signal of a sensor

Info

Publication number: EP1891316A1
Application number: EP06755183A
Authority: EP
Inventors: Joachim Berger; Roland Klatt; Reinhold Danner; Heinrich Barth; Andreas Pfaeffle; Rainer Strohmaier; Thomas Bleile; Michael Scheidt; Matthias Illing; Wolfgang Dressler; Torsten Handler; Christof Rau; Thomas Pauer; Gottfried Flik; Udo Schulz; Wolfgang Fischer; Matthias Schueler; Jan Bahlo; Jean-Pierre Hathout
Original assignee: Robert Bosch GmbH
Current assignee: Robert Bosch GmbH
Priority date: 2005-06-06
Filing date: 2006-05-15
Publication date: 2008-02-27
Also published as: DE102005025884A1; WO2006131435A1; JP2008542776A; US20090222231A1; KR20080015430A; CN101189423A

Abstract

The invention relates to a method and a device for correcting the signal of a sensor (1) which allow as exact a drift compensation of a characteristic line of the sensor (1) as possible. At least one characteristic value of the signal of the sensor (1) is compared with a reference value. The signal of the sensor (1) is corrected depending on the result of comparison. The invention is characterized by producing a value derived from the signal of the sensor (1) for the at least one characteristic value of the signal of the sensor (1) as the reference value.

Description

Method and device for correcting a signal of a sensor

State of the art

The invention is based on a method and a device for correcting a signal of a sensor according to the preamble of the independent claims.

It is thus known, for example, that the drift occurring during a service life of an internal combustion engine in a hot-film air mass meter is corrected by comparing the signal of the hot-film air mass meter with a mass air value modeled as a reference value from a boost pressure, a charge air temperature and an engine speed.

Since the boost pressure sensor for determining the charge pressure, the temperature sensor for determining the charge air temperature and the speed sensor for determining the engine speed are each subject to tolerances, the achievable with the known method accuracy of the drift compensation is lower than the new part tolerance of the non-polluted air mass meter.

Furthermore, it is known from DE 100 63 439 A1, for a sensor designed, for example, as a hot-film air mass meter, in addition to a signal range check

To carry out on-board diagnostics with regard to specifiable plausibility criteria relating to the offset drift and / or the sensitivity drift of the sensor.

Advantages of the invention The inventive method and the erfϊndungsgemäße device for correcting a signal of a sensor having the features of the independent claims have the advantage that at least one characteristic size of the signal of the sensor is compared with a reference value and the signal of the sensor is corrected depending on the comparison result, as the reference value on from the signal of the

Sensor derived value for the at least one characteristic size of the signal of the sensor is formed. In this way, it is possible to dispense with the use of substitute signals for modeling the signal of the sensor or of the at least one characteristic variable as well as also to model the signal of the sensor itself and, using only the signal from the sensor for forming the reference value, an increased accuracy of the sensor Drift compensation can be achieved.

The measures listed in the dependent claims advantageous refinements and improvements of the main claim method are possible.

It is particularly advantageous if the reference value is formed in a predetermined operating state of the sensor, in particular within a predetermined time after initial startup of the sensor. In this way, the accuracy of the drift compensation of the signal of the sensor can be increased. In the most favorable case, the accuracy of the drift compensation is influenced only by the new part tolerance of the non-contaminated sensor.

A further advantage results if the sensor detects an operating variable of a drive unit, in particular an internal combustion engine, and if the formation of the reference value and / or the formation of the at least one characteristic variable of the signal of the sensor for comparison with the reference value in at least one predetermined Operating state of the drive unit, in particular in an idle state, is performed. In this way, the accuracy of the drift compensation can be further increased, in particular by taking into account the time constants present in the measured value acquisition by the sensor.

It is particularly advantageous if an air mass measuring device, in particular a hot-film or ultrasonic air mass meter, is selected as the sensor. In this way, as accurate as possible drift compensation can be carried out for such an air mass measuring device. AIs at least one characteristic size of the signal of the sensor are particularly suitable a time average and / or a signal amplitude of the signal of the sensor. From these two variables, an offset and a sensitivity of a sensor characteristic curve for the conversion of the sensor signal into the measured variable to be detected can be corrected in a simple and reliable manner.

The correction of the signal of the sensor can be carried out particularly simply by forming, depending on the result of the comparison, at least one correction value with which the signal of the sensor is corrected.

In order to determine a correction value that is as reliable and error-free as possible, it can be provided in an advantageous manner that the at least one correction value is formed only in the case of a signal of the sensor recognized to be plausible, in particular as a function of its time course.

The correction of the signal of the sensor can be carried out particularly simply by forming the at least one correction value as a correction value for an offset and / or as a correction value for a sensitivity of the signal of the sensor.

In particular, in the case of a non-linear characteristic, it is advantageous if the at least one correction value is formed differently in different regions of the signal size. In this way, a drift compensation which is as accurate as possible can be realized even in the case of a non-linear sensor characteristic, specifically for several ranges of this characteristic, in particular for the entire characteristic curve.

drawing

An embodiment of the invention is illustrated in the drawing and explained in more detail in the following description. Show it

FIG. 1 shows a block diagram of a detail of a drive unit designed as an internal combustion engine,

FIG. 2 shows a reference characteristic curve and a drift characteristic deviating therefrom of an air mass meter, - A -

FIG. 3 shows a functional diagram for explaining the method according to the invention and the device according to the invention, and

4 shows a flow chart for an exemplary sequence of the method according to the invention.

Description of the embodiment

In FIG. 1, 5 designates, by way of example, a drive unit embodied as an internal combustion engine with a cylinder block 40 to which fresh air is supplied via an air supply 35. The internal combustion engine 5, for example, a gasoline engine or a

Drive diesel engine. In the air supply 35, an air mass meter 1, for example in the form of a Heißfileinuftmassenmessers or an ultrasonic air mass meter is arranged. Furthermore, a rotational speed sensor 45 is arranged in the area of the cylinder bank 40, which detects a motor rotational speed nmot in a manner known to those skilled in the art at predetermined, in particular equidistant, sampling times and forwards the corresponding measured values to a controller 50. Depending on the air mass flow in the air supply 35, the air mass meter 1 likewise generates a signal S in the form of time-discrete measured values in a manner known to those skilled in the art, these measured values in turn being detected in particular at equidistantly spaced points in time. The signal S of the air mass meter 1 is also forwarded to the controller 50. Further for the

Operation of the internal combustion engine in a manner known to those skilled in the known or required components, which are not required for understanding the invention, are not shown in Figure 1 for reasons of clarity.

The controller 50 converts the signal S of the air mass meter 1 into the physical quantity of the air mass flow LMS by means of a characteristic curve. FIG. 2 shows two such characteristics stored in the controller 50. In this case, the air mass flow LMS is plotted against the signal S of the air mass meter 1. The two characteristics shown are linear in this example. This represents a simplification of the actual relationship between the signal S and the air mass flow LMS, which is less true in the case of the air mass meter 1 as an ultrasonic air mass meter and less true in the case of forming the air mass meter 1 as a hot film air mass meter, but below Explanation of the method and apparatus of the invention should be based. In this case, R denotes a reference characteristic with a first offset value Ol and a first offset value Curve slope or sensitivity Yl / Xl. Furthermore, in the diagram according to FIG. 2 a drift characteristic D is shown which has a second offset 02 and a second slope or sensitivity Y2 / X2, where Ol ≠ 02 and Yl / Xl ≠ Y2 / X2. In this case, it should be assumed in this example that the reference characteristic R is the image of the signal S of the air mass meter 1 in the air mass flow LMS in a

Represents new condition of the air mass meter 1, in which the air mass meter 1 is not dirty. In contrast, the drift characteristic D describes the image of the signal S of the air mass meter 1 in the air mass flow LMS at a later time at which the air mass meter 1 already has some pollution, which leads to a larger compared to the reference curve offset, d. H. 02> Ol and which leads to a lower sensitivity or slope compared to the reference characteristic R, d. H. ie Y2 / X2 <Yl / Xl. The drift characteristic D thus results due to the contamination of the air mass meter 1. Additionally or alternatively, the drift characteristic D may also result from the aging of the air mass meter 1 and the concomitant wear.

The signal S of the air mass meter 1, depending on the number of cylinders of the cylinder bank 40 and the engine speed nmot pulsations, which are superimposed on the time average of the signal S of the air mass meter 1. Due to soiling of the air mass meter 1, over the service life of the air mass meter 1, offset and sensitivity or slope drifts of the characteristic of the air mass meter 1, which maps the signal of the air mass meter 1 into the physical quantity of the air mass flow. These offset and sensitivity drifts lead to a shift in the average time value of the air mass flow LMS resulting from the aforementioned characteristic curve and to a change in its pulsation amplitude.

The goal is to convert the signal S of the air mass meter 1 at any time as accurately as possible into the air mass flow LMS, ie to determine the current drift characteristic D as far as possible at any time. For this purpose, the controller 50 comprises a device 10 according to the functional diagram according to FIG. 3. In this case, the device 10 can be implemented in the controller 50 in software and / or hardware, for example. The device 10 can also be identical to the controller 50, ie form the controller 50 or a corresponding control unit. This control unit may be identical to or different from an engine control unit. The device 10 comprises a reference value formation unit 30 with an evaluation unit 55, a first controlled switch 60 and a second controlled switch 65. The device 10 further comprises an operating state detection unit 95 which detects the engine speed nmot detected by the rotational speed sensor 45 and the time t detected by a time detection unit 90 have been supplied for the first time commissioning of the air mass meter 1. In this case, the time t may also correspond to the time which has elapsed since the first startup of the internal combustion engine 5, when this time coincides with the time of the initial startup of the air mass meter 1. The time acquisition unit 90 may be part of the device 10 or, as shown in FIG. The first controlled switch 60 and the second controlled switch 65 are each actuated by the operating state detection unit 95 in their switch position. This activation takes place depending on the time t and the engine speed nmot, which characterize the operating state of the internal combustion engine 5. The device 10 further comprises a current drift characteristic D, which is identified by the reference numeral 110. The signal S of the air mass meter 1 is fed both to the evaluation unit 55 and the drift characteristic 110 on the input side. The drift characteristic D is corrected by a correction unit 25 of the device 10. This is done by means of a first correction value KO for the offset of the drift characteristic 110 and a second correction value KS for the slope or sensitivity of the

Drift characteristic 110. At the output of the drift characteristic 110 then results in the air mass flow LMS, which is discharged from the device 10 for internal and / or external further processing. The correction unit 25 can receive the output signal of a first comparison unit 15 via a third controlled switch 100 and the output signal of a second comparison unit 20 via a fourth controlled switch 105. The two comparison units 15, 20 are also part of the device 10. In the first comparison unit 15, the output of a first reference value memory 70 is compared with the output of a first comparison value memory 80 and in the second comparison unit 20, the output of a second reference value memory 75 with the output of second comparison value memory 85 compared. Both

Reference value memory 70, 75 and both comparison value memory 80, 85 are arranged in the device 10 in the example according to FIG. The first controlled switch 60 connects a first output 115 of the evaluation unit 55 either to an input of the first reference value memory 70 or to an input of the first comparison value memory 80. The second controlled switch 65 connects a second output 120 of the evaluation unit. Unit 55 either with an input of the second reference value memory 75 or with an input of the second comparison value memory 85. The control of the third controlled switch 100 and the fourth controlled switch 105 is dependent on the operating condition of the internal combustion engine 5 by the Betriebszustandserfassungsein- unit 95th

The first controlled switch 60 is connected by the operating state detection unit 95 for connecting the first output 115 of the evaluation unit 55 to the input of the first reference value memory 70 when the time t is less than a predetermined limit time tgrenz and the engine speed nmot is less than a predetermined engine speed nmotgrenz. Otherwise, the operating state detection unit 95 controls the first controlled switch 60 to connect the first output 115 of the evaluation unit 55 to the input of the first comparison value memory 80. In a corresponding manner, the second controlled switch 65 is actuated by the operating state detection unit 95 to connect the second output 120 of the evaluation unit 55 to the input of the second reference value memory 75 when t <tlimit and nmot <nmotgrenz. Otherwise, the second controlled switch 65 is driven by the operating state detection unit 95 to connect the second output 120 of the evaluation unit 55 to the input of the second comparison value memory 85.

The predetermined time tgrenz can be suitably applied, for example, on a test bench, such that for times t <tgrenz is not yet to be expected with a polluted air mass meter 1. Tgrenz can be derived in particular from empirical values of air mass meters of the same type. The limit nmotgrenz for the engine speed can also be suitably applied, for example, on a test bench, such that engine speeds nmot <nmotgrenz characterize an idling state of the internal combustion engine 5. In principle, the limit nmotgrenz for the engine speed should be applied in an advantageous manner so that the time constant of the air mass meter 1 in the air mass detection, which may be up to 15ms, for example, is taken into account. The limit nmotgrenz for the engine speed can be applied so that for engine speeds nmot <nmotgrenz the air mass detection by the air mass meter 1 due to the time constant of the air mass meter 1 is not or only slightly falsified, the falsification of the air mass measurement for engine speeds nmot> nmotgrenz, however, an undesirable high level. In this way, it is ensured that the first reference value memory 70 and the second reference value memory 75 are only described or overwritten in an operating state of the internal combustion engine 5 in which substantial air pollution of the air mass meter 1 is not to be expected. In addition, it is ensured that the first reference value memory 70 and the second reference value memory 75 are only written or overwritten in an operating state of the internal combustion engine 5 in which the measurement result of the air mass meter 1 is not limited by a rotational speed which is too high or exceeds the limiting rotational speed nmotgrenz engaging engine speed nmot is corrupted.

The third switch 100 is closed by the operating state detection unit 95 for connecting the output of the first comparison unit 15 with the correction unit 25, if nmot <nmotgrenz and t> tgrenz. Otherwise, the third controlled switch 100 is opened by the operation state detection unit 95. The fourth controlled switch 105 is operated by the operating state detecting unit 95 to connect the

Output of the second comparison unit 20 is closed with the correction unit 25 when nmot <nmotgrenz and t> tgrenz. Otherwise, the fourth controlled switch 105 is opened by the operation state detection unit 95.

The first comparison value memory 80 and the second comparison value memory 85 are only written or overwritten in the operating states in which the first reference value memory 70 and the second reference value memory 75 can not be written or overwritten because of the switch position of the first controlled switch 60 and the second controlled switch 65. Alternatively, it may also be provided that the first comparison value memory 80 and the second comparison value memory 85 are in principle described or overwritten in any state of the internal combustion engine 5. An update of the two correction values KO and KS in the correction unit 25 only takes place as long as the two controlled switches 100, 105 are in their closed position, as shown in FIG. If the two switches 100, 105 are opened, there is no actualization of the correction values K0, KS by the correction unit 25. The correction of the drift characteristic 110 always takes place with the last updated correction values K0, KS. As shown in FIG. 3, the two switches 60, 65 are driven synchronously by the operation state detection unit 95. The same applies to the two controlled switches 100, 105. By means of the two controlled switches 100, 105 it is ensured that the correction unit 25 only then the two correction values KO, KS are updated when the engine speed nmot <nmotgrenz and the time t> tgrenz. In this case, the drift characteristic curve 110 may initially be predetermined in the form of the reference characteristic curve R and stored in the device 10 in accordance with the manufacturer's instructions of the air mass meter 1 or based on a calibration measurement. A correction of this drift characteristic 110 then takes place only after expiration of the predetermined time tgrenz after initial startup of the air mass meter 1 or the internal combustion engine 5 and under the condition that the engine speed nmot is below the predetermined limit rpm nmotgrenz, that is, the correction by not one high speed greater than or equal to the limit speed nmotgrenz is falsified. In other words, also in the correction of the drift characteristic 110, the

Time constant in the air mass detection by the air mass meter 1 is taken into account in order to avoid errors in the correction of the drift characteristic 110.

The evaluation unit 55 evaluates the signal S of the air mass meter 1 with regard to at least one characteristic variable of this signal S. In the present example, the evaluation unit 55 evaluates the signal S of the air mass meter 1 with regard to two characteristic quantities of the signal S. In this case, the evaluation unit 55 determines, as a first characteristic variable of the signal S, a time average of this signal S and outputs it as a moving average value at its first output 115. Furthermore, the evaluation unit 55 determines as the second characteristic variable of the signal S the currently actual value of the signal amplitude of the signal S and outputs it at its second output 120.

Depending on the switch position of the first controlled switch 60 then the current sliding temporal average value of the signal S in the first reference value memory 70 or in the first

Comparison value memory 80 filed. Accordingly, depending on the position of the second controlled switch 65, the current value for the signal amplitude of the signal S is stored in the second reference value memory 75 or in the second comparison value memory 85. The first comparison unit 15 compares the sliding mean value of the signal S stored in the first reference value memory 70 with the sliding time average stored in the first comparison value memory 80, for example by subtraction or by division, and passes the comparison result, ie the difference or the quotient in the case of the closed one third switch 100 to the correction unit 25 on. Accordingly, the second comparison unit 20 compares the value for the signal amplitude in the second reference value memory 75 with the value for the signal amplitude in the second comparison value. value memory 85, for example by subtraction or by quotient and forwards the comparison result in the form of the difference or the quotient to the correction unit 25, if the second controlled switch 105 is in its closed position.

Initially, the first reference value memory 70 and the first comparison value memory 80 may be assigned the same value, so that the first comparison unit 15 outputs the value zero at its output as a comparison result in the case of subtraction. Accordingly, the second reference value memory 75 and the second comparison value memory 85 may initially be assigned the same value, with the result that the second comparison unit 20 has at its

Output at quotient forming the value one. In this case, it can generally be provided that in the case in which the respective two input variables are the same size, the first comparison unit 15 outputs the value zero at its output and the second comparison unit 20 outputs the value one at its output. The correction unit 25 receives the value zero from the first comparison unit 15 and from the second comparison unit

20 is one, it does not update the two correction values K0, KS. This corresponds to a state with opened switches 100, 105. In this case, the correction value KO for the offset can initially be set to the value zero and the correction value KS for the slope or the sensitivity can initially be set to the value 1. In this case, the correction of the drift characteristic 110 takes place by adding the offset of the drift characteristic 110 with the first correction value KO and the correction of the slope of the drift characteristic 110 by multiplication with the second correction value KS. Alternatively, the correction of the offset can also be done in any other way, for example by multiplication, by division or by subtraction, as well as the correction of the slope of the drift characteristic 110 can alternatively be done in any other form, for example by addition, by subtraction or by division. However, the manner of correcting the offset and the slope of the drift characteristic 110 should be predetermined and maintained in an advantageous manner. Depending on the selected correction operation, ie addition, subtraction, division or multiplication, the correction values K0, KS are to be initialized, in order not to initially modify the drift characteristic 110.

The output of the first reference value memory 70 is identified in FIG. 3 by Rl, the output of the first comparison value memory 80 by V1, the output of the second reference value memory 75 by R2 and the output of the second comparison value memory 85 by V2. The following is an example assuming that the first 15 equal to the difference Δ = Rl - Vl forms and with closed third controlled switch 100 to the correction unit 25 passes. Furthermore, it should be assumed that the second comparison unit 20 forms the quotient Q = R2 / V2 and forwards it to the correction unit 25 as the result of the comparison in the case of the closed fourth controlled switch 105. The correction unit 25 is formed from the difference Δ, the

Quotient Q and the first offset, oil of the reference characteristic of the air mass meter by means of an equation system and the first correction value KO for the offset of the drift characteristic 110 and the second correction value KS for the slope of the drift characteristic 110. The equation system is as follows:

κs = λ-Q

KO = (I - -) (Rl - Ol) - A

The drift characteristic 110 is then corrected by means of the first correction value KO and the second correction value KS in such a way that the current offset of the drift characteristic 110 is added to the first correction value KO by a new offset for the drift characteristic

110 and that the current slope of the drift characteristic 110 is multiplied by the second correction value KS to form a new slope for the drift characteristic 110. In this way, after the correction by means of the two correction values K0, KS, a new drift characteristic 110 exists, which converts the signal S of the air mass meter 1 into the physical quantity of the air mass flow LMS.

Alternatively, in the case of a linear reference characteristic, the first offset value Ol can also be determined via a measurement in the control unit follow-up in the new state of the air mass meter 1, in which no air mass flow is present. The first offset value Ol is stored in an offset value memory 1000 of the device 10 and supplied from there to the correction unit 25. The output of the first reference value memory 70 is also supplied to the correction unit 25.

FIG. 4 describes a flowchart for an exemplary sequence of the method according to the invention as performed by the device 10. After the start of the program, the operating state detection unit 95 at a program point 200 receives the current time t, which has elapsed since the first startup of the air mass meter 1 or of the internal combustion engine 5, from the time detection unit 90, which is the first time the air mass meter 1 or internal combustion engine is started up 5 with the Value t = 0 was initialized. Furthermore, the operating state detection unit 95 receives at program point 200 from the rotational speed sensor 45, the current engine speed nmot of the internal combustion engine 5. Then, a program point 205 is branched.

At program point 205, it is checked whether in the first reference value memory 70 and in the second reference value memory 75 a value has already been received and stored by the evaluation unit 55. This is checked by the first comparison unit 15 checking whether the difference Δ ≠ is zero and that the second comparison unit 20 checks whether the quotient Q ≠ 1. If this is the case, a branch is made to a program point 210, otherwise a branch is made to a program point 225.

At program point 225, the operating state detection unit 95 checks whether t <tlimit and nmot <nmotlimit. If this is the case, a branch is made to a program point 230, otherwise the program branches back to program point 200.

At program point 230, the operating state detection unit 95 causes the first controlled switch 60 to connect the first output 115 of the evaluation unit 55 to the first reference value memory 70 and the second controlled switch 65 to connect the second output 120 of the evaluation unit 55 to the second reference value memory 75 leads to a description of the first reference value memory

70 with the current sliding time average of the signal S of the air mass meter 1 and the second reference value memory 75 with the current signal amplitude of the signal S at the subsequent program point 235. Subsequently, the program branches back to the program point 200.

At program point 210, the operating state detection unit 95 checks whether nmot <nmotlimit. If this is the case, a branch is made to a program point 215, otherwise a branch is made back to program point 200. In order to branch to program point 215, it does not necessarily have to be necessary for additional t to be greater than or equal to tgrenz. The correction of the drift characteristic 110 can also be carried out already for times t <tgrenz.

At program point 215, the operating state detection unit 95 initiates a closure of the two controlled switches 100, 105. A branch is then made to a program point 220. At program point 220, the correction unit 25 determines the first correction value KO and the second correction value KS from the supplied input quantities Δ, Q in the described manner and with these corrects the third characteristic curve 110 in the manner described. Afterwards the program is left.

According to a development of the invention, it may be provided that the correction values K0, KS are formed only in the case of a signal S of the air mass meter 1, which is recognized as being plausible, in particular as a function of its time profile. For this purpose, the evaluation unit 55 carries out a plausibility check of the signal S. In this case, the evaluation unit 55 can check, for example, whether there is a non-uniform amplitude change of the signal S, for example due to a leak in one of the cylinders of the cylinder bank 40. Such a non-uniform change in amplitude can be detected by the evaluation unit 55 when the amplitude of the signal S within a two cycle revolutions comprehensive cycle of the cylinder has a fluctuation range which is above a predetermined value, which can be applied for example on a test bench so suitable that it can distinguish the amplitude change of the signal S due to a leak in one of the cylinders of the cylinder bank 40 from a comparatively smaller change in amplitude, which results without cylinder leakage, solely due to installation tolerances and aging influences. The evaluation unit 55 then outputs a plausibility signal P to the operating state detection unit 95 as a function of this plausibility check. If the plausibility information P is set, it indicates a plausible signal S, otherwise, that is, if the signal S is reset, this indicates an implausible signal S. In the case of an implausible signal S, the operating state detection unit 95 initiates an opening of the two controlled switches 100, 105 in order to prevent erroneous correction of the drift characteristic 110. If, on the other hand, the plausibility information P is set, then the opening or closing state of the two controlled switches 100, 105 depends in the manner described above on the time t and the engine speed nmot, or only on the engine speed nmot.

In the above, it has been assumed by way of example that the drift characteristic 110 is linear. In general, however, the drift characteristic 110 will not be linear, but can be approximated by a linear characteristic, in particular in the case of the ultrasonic air mass meter, to a rough approximation. In the case of a Heißfileinuftmassenmessers is such Linearization of the drift characteristic 110 may no longer be productive, so that the drift characteristic 110 must be linearized differently in this case, at least in different areas. In this case, it can be provided that the evaluation unit 55 additionally checks in which region of the characteristic line the received signal S of the air mass meter 1 is located, wherein this information can also be communicated to the operating state detection unit 95 by means of a signal B. In this case, an arrangement with a first reference value memory, a first reference value memory, a first comparison unit and a second reference value memory, a second reference value memory, a second one would then be for each of the said ranges of the signal magnitude, which are mapped different linearized by the drift characteristic 110 Comparing unit and a correction unit to be provided, which corrects the respective linearized region of the signal magnitude in the drift characteristic 110, each with a correction value for offset and a correction value for slope. The operating state detection unit 95 then has to switch over between the individual arrangements with the two reference value memories, the two comparison value memories, the two comparison units and the correction unit depending on the currently present signal range, the operating state detection unit 95 being notified of the current signal range by the signal B as described , The location of the switch to be attached is indicated by reference numeral 125 in FIG. 3 and is located between the first controlled switch 60 and the first reference value memory 70, between the first controlled switch 60 and the first comparison value memory 80, between the second controlled switch 65 and the second reference value memory 75 and between the second controlled switch 65 and the second comparison value memory 85. The activation of these additional switches 125 by the operating state detection unit 95 is also indicated by dashed lines in FIG.

It goes without saying that the correction of the drift characteristic curve 110 by the correction unit 25 or the correction of a region of the drift characteristic curve 110 by thecorrespondingly assigned correction unit for a currently received signal value of the

Air mass meter 1 is performed only when previously the corresponding comparison value memory 80, 85 have been filled depending on this current signal value S, by the comparison units 15, 20 corresponding comparison results .DELTA.Q have been formed and these from the associated correction unit 25 in corre sponding correction values KO, KS were converted. It may also be provided for this purpose that a suitable timing of the storage of the comparison values in the comparison value memory 80, 85, the comparison units 15, 20 and the associated correction unit 25 is performed for example by the operating state detection unit 95, wherein the comparison value memory 80, 85 are overwritten at a first time clock a subsequent second time clock, the comparison units 15, 20 the

Comparison results Δ, Q determine and deliver and at a subsequent third timing cycle, the correction unit 25, the correction values KO, KS determined and for correction to the drift characteristic 110 passes. This timing cycle from the overwriting of the comparison value memories 80, 85 to the correction of the drift characteristic 110 should be within the time interval between two directly successively determined measured values of the

Run off air mass meter 1.

The method described and the device described have been described by way of example with reference to the drift compensation of an air mass meter 1. In a corresponding manner, any other sensors of the internal combustion engine 5, for example a pressure sensor, a temperature sensor or a rotational speed sensor can be compensated in their drift, but also sensors that are not installed in an internal combustion engine 5 and physical variables, such as pressure, for example. Temperature, mass flow, speed or the like capture.

Depending on the sensor used, at least one characteristic variable of the signal of the sensor is compared with a reference value and the signal of the sensor is corrected as a function of the result of the comparison. In this case, a value derived from the signal of the sensor for the at least one characteristic variable of the signal of the sensor is formed as the reference value. As characteristic quantities of the signal of the air mass meter 1, the temporal mean value and the signal amplitude were selected in the example described above. If, for example, the characteristic curve of the sensor only depends on one size, for example, it always has a fixed offset value and only drifts with respect to the slope or always has a fixed slope and drifts only in relation to the offset, then it is sufficient if as a reference value a derived from the signal of the sensor value for a single characteristic size of the signal of the sensor is formed, for example, only the time average or only the signal amplitude. In particular, in the case of non-linear sensor characteristics, it may also be necessary to form as a reference value a value derived from the signal of the sensor for more than two characteristic quantities of the signal of the sensor. In addition to the temporal chen mean value and the signal amplitude also still, for example, the second time derivative of the signal belong.

In FIG. 2, such a non-linear characteristic X is shown in dashed lines, which is divided into four linearized regions. Thus, depending on its size, the signal S can lie in one of these four ranges. The four areas are defined as follows:

0 <= S <Sl S1 <= S <S2 S2 <= S <S3

S3 <= p.

Each of these four areas is assigned an arrangement of a first reference value memory, a first comparison value memory, a first comparison unit, a second reference value memory, a second comparison value memory, a second comparison unit and a correction unit as shown in FIG. 3 and via the switching points indicated in FIG 125 switchable.

It has been described above by way of example that the reference value memories 70, 75 are only described if t <tgrenz. Additionally or alternatively, the reference value memory 70, 75 but also described in another predetermined operating state of the air mass meter or overwritten. Such a predetermined operating state is characterized by the fact that the air mass meter 1 is not polluted in this operating state and is free of aging influences or wear. This may also be the case after maintenance of the air mass meter 1. Thus tgrenz can also be interpreted as a time limit after a corresponding maintenance of the air mass meter 1. A predefined operating state of the air mass meter 1 without soiling and aging effects or wear can also be achieved by plausibility of the air mass meter 1, for example with the aid of a redundant air mass meter or in any other manner known to the person skilled in the art, for example also by modeling the air mass meter signal from other operating variables the internal combustion engine 5 are detected, wherein a writing or overwriting of the reference value memory 70, 75 should then be possible in such a predetermined operating state of the air mass meter 1, if the condition for the engine speed nmot <nmotgrenz is satisfied. Also, the drive unit 5 does not have to be designed as an internal combustion engine as in the manner described, but may for example be designed as a hybrid drive from the engine and electric motor or as an electric motor or in any other known in the art, a sensor of this drive unit in the manner described in its drift can be compensated.

Furthermore, in advance, for example, the plausibility of the signal S was described as a function of its time course. However, the plausibility check can also take place in other ways known to the person skilled in the art, for example in that a characteristic size of the signal of the sensor, for example the time average or the signal amplitude, is made plausible. In this way too, in the case of a nonuniform change in amplitude due to, for example, a leak in one of the cylinders of the cylinder bank 40, an implausible characteristic variable of the signal S, for example the time average or the signal amplitude would result. Ie. the characteristic quantity would in this case deviate impermissibly from an expected value. The temporal mean value of the signal S would therefore, for example, inadmissible of an expected time average or the signal amplitude of the signal S would deviate inadmissibly from an expected signal amplitude.

The reference value memory 70, 75 and the comparison value memory 80, 85 may be formed for example as EEPORM.

Claims

claims

1. A method for correcting a signal of a sensor (1) in which at least one characteristic quantity of the signal of the sensor (1) is compared with a reference value and the signal of the sensor (1) is corrected as a function of the result of the comparison, characterized that as a reference value from the signal of

Sensor (1) derived value for the at least one characteristic size of the signal of the sensor (1) is formed.

2. The method according to claim 1, characterized in that the reference value in a predetermined operating state of the sensor (1), in particular within a predetermined time after initial startup of the sensor (1) is formed.

3. The method according to any one of the preceding claims, characterized in that by the sensor (1) an operating variable of a drive unit (5), in particular an internal combustion engine, is detected and that the formation of the reference value and / or the formation of the at least one characteristic Size of the signal of the sensor (1) for comparison with the reference value in at least one predetermined operating state of the drive unit (5), in particular in an idle state, is performed.

4. The method according to any one of the preceding claims, characterized in that as the sensor (1) an air mass measuring device, in particular a Heißfϊlm- or ultrasound air mass meter, is selected.

5. The method according to any one of the preceding claims, characterized in that a time average and / or a signal amplitude is selected as at least one characteristic size of the signal of the sensor (1).

6. The method according to any one of the preceding claims, characterized in that depending on the comparison result, at least one correction value is formed, with which the signal of the sensor (1) is corrected.

7. The method according to claim 6, characterized in that the at least one correction value is formed only in the case of a signal of the sensor (1) which, in particular as a function of its time course, is recognized as being plausible.

8. The method according to claim 6 or 7, characterized in that the at least one correction value as a correction value for an offset and / or as a correction value for a sensitivity of the signal of the sensor (1) is formed.

9. The method according to any one of claims 6 to 8, characterized in that the at least one correction value is formed differently in different areas of the signal size.

10. Device (10) for correcting a signal of a sensor (1), with at least one comparison unit (15, 20), which compares at least one characteristic quantity of the signal of the sensor (1) with a reference value, and with a correction unit (25). which corrects the signal of the sensor (1) as a function of the result of the comparison, characterized in that means (30) are provided for forming a reference value which uses as reference value a value derived from the signal of the sensor (1) for the at least one characteristic value Form the size of the signal from the sensor (1).