Patent Application
20170351562
The present invention relates to a method for processing at least one information item in a networked system and to a networked system, and to a calculation unit and a computer program for carrying out the method.
When different information sources are networked, for example for the so-called “internet of things,” the result is that data and/or information items that are obtained from external sources are normally regarded as valid directly after they are transferred, and are checked for validity and plausibility only later on in processing.
For example, according to the CAN protocol it is a principle of information transfer between two control units of a motor vehicle that the transmitting control unit vouches for the validity of the transmitted data and/or information items.
Control units of this kind process large volumes of signals that derive from external sources, for example sensors or interfaces. Those signals are combined during processing steps, and the result is then used for decisions upon software execution or in order to apply control to actuating members. Digital signals (logical zero=invalid, logical one=valid) or discrete status variables (invalid, initialized, . . . , valid), which characterize the validity status, can be used partly in parallel with the variables used for that purpose in the software.
During initialization of the control units and in the event of a fault in the external sources (e.g. due to a sensor failure, delays in the interface, etc.), in real-time systems initial and substitute values are typically used for further data processing.
It is often not possible to determine, based on the input variables of a control unit, whether or not a signal or a value ascertained in that control unit is trustworthy at the current time, since no information regarding their validity exists for those variables.
A need therefore exists to demonstrate ways in which the validity of at least one information item can be taken into consideration upon processing thereof.
According to the present invention, an example method for processing at least one information item in a networked system, and a networked system, as well as a calculation unit and a computer program for carrying out the method are provided. Advantageous embodiments are described herein.
In order to process at least one information item in a networked system to yield a result, the at least one information item is transferred from at least one source of the system to a receiver of the system. An interim data set having at least two variables is used, of which a first variable is allocated to a value of the at least one information item, and a further variable is allocated to the validity of the value. At least one evaluating step is carried out in the receiver, using the value and the validity, in order to determine a resultant data set having a resultant value and a resultant validity as a result of the processing of the at least one information item. The resultant validity thus makes available a value with which the validity of the resultant value can be assessed. This allows the result to be assessed as to whether or not it is valid, i.e. correct or trustworthy. The validity indicates, for example whether what was intended to be calculated or measured was in fact calculated or measured, or whether or not the information items obtained are also in fact correct.
Preferably, in a comparing step, the validity is compared with a minimum value, and the interim data set and/or the resultant data set are determined as a function of the comparison result. This makes it possible to react effectively, in the context of processing, to information items having insufficient validity.
Preferably a substitute value is used as a value of the at least one information item for the interim data set if the validity is less than or equal to the minimum value. It is thereby possible, for example, to prevent possibly erroneous values from leading to an undesired result, or to prevent possibly faulty values from being unrestrictedly further processed or forwarded as valid values. The substitute value can advantageously be a reliable standard value for the respective source. Alternatively, it can also be the most recent valid value, or an average of several most recent valid values.
The substitute value is preferably a value from a different source having a higher validity. If the validity is too low, information values of further sources are therefore evaluated so as therefore to find a value having a higher validity.
According to a further embodiment the resultant value is determined only when the value of the validity is greater than or equal to the minimum value. This ensures that values that are to be categorized as very uncertain, because of their low validity, cannot be further processed and thus lead to incorrect results.
According to an embodiment, a confidence value relating to the respective source is allocated by the receiver to the at least one source. A confidence value that does not derive from the source itself, and is thus source-independent, is thus used. Malfunctions of the source itself thus cannot lead to an erroneously high confidence value.
According to a further embodiment, the confidence value is used in order to determine the resultant validity. The confidence value is thus incorporated into the resultant validity, so that the resultant validity depends not only on the validity of the value but also on the confidence value of the source. A particularly accurate resultant validity is thus furnished.
According to a further embodiment, the validity can have values between one and zero. The value zero can be allocated to an absence of validity, and the value one to a validity verging on certainty. Validity values less than one and greater than zero characterize a lower or unknown validity of the data or of the result. This permits particularly simple processing of the values for the validity.
The confidence value(s) of the source can furthermore analogously have values between one and zero. The value zero can be allocated to an absence of validity, and the value one to a confidence verging on certainty. This permits particularly simple processing of the confidence values.
In terms of the validity of the result, this means that low validity values close to zero occur when data sources with an unknown origin for the original data or an unclear data source state are used, or because the data source itself has evaluated the validity of the transmitted values to be less than one. High validity values of one or close to one, on the other hand, occur when known sources with a good knowledge of the state, i.e. a high confidence value for the data source, are used, and because the data source itself has evaluated the validity of the transmitted values to be one or very close to one.
A networked system, for example an electrical system of a motor vehicle, has as a calculation unit a control unit that is configured to carry out, in particular by programmed execution, a method according to the present invention. At least one source can be embodied as a sensor of a motor vehicle. The processing reliability of information items in a vehicle electrical system can be appreciably increased with such a system, so that the electrical system is, for example, less sensitive to electromagnetic interference signals and/or to a failure or malfunctions of components of the electrical system, e.g. sensors. Improvements are therefore achieved specifically in the safety-relevant sector of motor vehicle electrical systems.
Implementation of the method in the form of software is also advantageous, since this entails particularly low cost, especially if an executing control unit is also used for further tasks and is therefore present in any case. Suitable data media for furnishing the computer program are, in particular, diskettes, hard drives, flash memories, EEPROMs, CD-ROM, DVDs, and many others. Downloading of a program via computer networks (internet, intranet, etc.) is also possible.
Further advantages and embodiments of the present invention are described below and are shown in the figures.
It is understood that the features discussed above and those yet to be explained below are usable not only in the respective combination indicated but also in other combinations or in isolation, without departing from the scope of the present invention.
The present invention is schematically depicted in the figures on the basis of an exemple embodiment, and is described below with reference to figures.
In the present exemplifying embodiment, sources 41, 42, 43, . . . , 4n can be intelligent or smart sensors or interfaces of an electrical system of a motor vehicle, while receiver 6 can be a control unit or a portion of a control unit of the motor vehicle.
From the respective sensors 41, 42, 43, . . . , 4n, respective information items I1, I2, I3, . . . , In are transferred to control unit 6, the respective information items I1, I2, I3, . . . , In each having a value x1, x2, x3, . . . , xn, for example a measured value of a sensor.
According to the embodiment shown, the respective validities v1, v2, v3, . . . , vn relating to the values x1, x2, x3, . . . , xn of information items I1, I2, I3, . . . , In are also transferred from the respective sensors 41, 42, 43, . . . , 4n to control unit 6. Information items I1, I2, I3, . . . , In can have a vector format.
While values x1, x2, x3, . . . , xn corresponding to sensors 41, 42, 43, . . . , 4n can assume any values, for example in accordance with the measurement range of a sensor, the validities v1, v2, v3, . . . , vn can assume, for example, values between zero and one. The value zero can be allocated to an unknown or low validity v1, v2, v3, . . . , vn, and the value one to a validity v1, v2, v3, . . . , vn verging on certainty. The validity v1, v2, v3, . . . , vn indicates here whether, for example, what was intended to be measured was in fact measured, or whether or not the information items I1, I2, I3, . . . , In obtained are in fact correct or derive from a trustworthy source.
Sensors 41, 42, 43, . . . , 4n can be embodied, for example, as intelligent or smart sensors or interfaces, and can have a self-diagnosis function or plausibility checking function. Sensors 41, 42, 43, . . . , 4n can thus, for example, detect their own malfunction or recognize implausible values. Alternatively or additionally, sensors 41, 42, 43, . . . , 4n can be embodied for a measured data evaluation that checks a series of measured values in terms of plausibility, for example whether the measured values fluctuate excessively within a predetermined time interval, or whether multiple values are plausible with respect to one another.
In the present exemplifying embodiment a CAN bus system is used to transfer the information items I1, I2, I3, . . . , In and validities v1, v2, v3, . . . , vn from sensors 41, 42, 43, . . . , 4n to control unit 6.
In addition to processing of the information items, the latter can also be displayed to a user or driver of the vehicle, for example on a display device. If the information items are made available on a display device, then in addition to the useful information item I the allocated validity value of the information item can additionally be displayed. If the information is organized in a sequenced list, the presentation can be organized by descending validity, so that information items having the highest validity are displayed first, and those having the lowest validity are displayed last.
Vehicle electrical system 2, sensors 41, 42, 43, . . . , 4n, and/or control unit 6 have hardware and software components for this purpose and in order to process information items I1, I2, I3, . . . , In.
Control unit 6 creates and uses an interim data set ZD having at least two variables. Of the two variables, a first variable is allocated to the value x1, x2, x3, . . . , xn of the respective information item I1, I2, I3, . . . , In, while a further variable is allocated to the validity v1, v2, v3, . . . , vn of the respective value x1, x2, x3, . . . , xn. The interim data set ZD can have a vector format.
Control unit 6 is embodied or configured to execute an evaluating step so that, using values x1, x2, x3, . . . , xn and their validities v1, v2, v3, . . . , vn, a resultant data set ED having a resultant value xE and a resultant validity xE are determined as a result. The resultant data set ED can have a vector format.
The evaluating step can encompass, for example, carrying out calculation operations using scalar variables, for example an addition involving the values x1, x2, and their validities v1, v2, of sensors 41, 42:
I
1
⊕I
2=(x1,v1)⊕(x2,v2)=(x1+x2,v1*v2).
In the context of addition, the values x1, x2 are thus added, and their validities v1, v2, whose value is between zero and one, are multiplied, so that the value of the resultant validity vE is always between zero and one.
The calculation operations can also encompass a multiplication involving the values x1, x2, and their validities v1, v2, of sensors 41, 42:
I
1
I
2=(x1,v1)(x2,v2)=(x1*x2,v1*v2).
In the context of multiplication, the values x1, x2 are thus multiplied, and their validities v1, v2, whose value is between zero and one in this case as well, are multiplied, so that here as well, the value of the resultant validity vE is always between zero and one.
According to further preferred embodiments, the information items I1, I2, I3, . . . , In can additionally have one or more further elements and can have, for example, the following vector format:
I
n=(xn,vn,qn,sx,nx,rx),
where qn can be a confidence value relating to the respective source, e.g., relating to a known source, a sensor value, an information item from a third-party source (e.g., government information, company information), etc., or a value between zero for an unknown source and one for a trustworthy source. A confidence value q1, q2, q3, . . . , qn is thus allocated by the receiver of the respective source 41, 42, 43, . . . , 4n and is used to determine the resultant validity vE.
The evaluating step can encompass, for example, any calculation operations, for example any mathematical functions relating to the values x1, x2, . . . , xn and their validities v1, v2, . . . , vn and confidence values q1, q2, . . . , qn of sources 41, 42, 43, . . . , 4n. The result represents a resultant value xe as a function fx of input values x1, x2, . . . , xn and of their validities v1, v2, . . . , vn confidence values q1, q2, . . . , qn, and a resultant validity ve as a function fv of the input validities v1, v2, . . . , vn and confidence values q1, q2, . . . , qn:
F(I1,I2, . . . ,In)=F((x1,v1,q1),(x2,v2,q2), . . . ,(xn,vn,qn))==(fx(x1,v1,q1,x2,v2,q2, . . . ,xn,vn,qn),fvx(x1,v1,q1,x2,v2,q2, . . . ,xn,vn,qn))
In addition, sx can assume, for example, values which denote respective information items I1, I2, I3, . . . , In that are obtained for the first time, obtained several times, confirmed by further sources, calculated, and/or valid.
In addition, nx can denote the number of utilized sources 41, 42, 43, . . . , 4n which was used in order to determine the resultant validity vE.
Lastly, rx can denote the number of sub-steps, having evaluating steps and optionally further processing steps, that were carried out starting from information items I1, I2, I3, . . . , In up to the current result.
Receiver 6 can furthermore be embodied to carry out a comparing step, for example before the evaluating step. In the comparing step, a value of one of the validities v1, v2, v3, . . . , vn is compared with a minimum value. The evaluating step for determining the resultant value xE is performed only when the value of the validity v1, v2, v3, . . . , vn is greater than or equal to the minimum value. The comparing step ensures that, for example, a value x1 that is to be categorized as very uncertain, based on its low validity v1, is not further processed and thus cannot lead to incorrect results.
In this case a fault signal can be generated, or the method is continued or repeated using a substitute value (e.g., standard value, average, moving average of earlier values having sufficient validity, etc.), as described below. For example, receiver 6 can be embodied to carry out a supplementing step, for example before the evaluating step but after the comparing step, the supplementing step being executed, for example, when the value of one of the validities v1, v2, v3, . . . , vn is less than or equal to the minimum value, i.e. is to be categorized as very uncertain. In the supplementing step, further information items I1, I2, I3, . . . , In from further sources 41, 42, 43, . . . , 4n are read in by receiver 6, or transferred to receiver 6 from the further sources 41, 42, 43, . . . , 4n, and then evaluated by receiver 6 in order to determine a further validity v1, v2, v3, . . . , vn for the values x1, x2, x3, . . . , xn that were read in or transferred.
What can possibly be achieved by way of the supplementing step is that the value of the further validity v1, v2, v3, . . . , vn is greater than the threshold value, so that the resultant value xE is then also determined.
The evaluating step and the optional comparing steps and supplementing steps can be sub-steps of a method, having multiple steps, for processing information items 41, 42, 43, . . . , 4n in a networked system 2 to yield a result. Interim resultant data sets having interim resultant values and interim resultant validities can be created during each sub-step and are updated again in the subsequent sub-step.
Lastly, receiver 6 can be embodied to create a sequenced information item list. The sequenced information item list IL has a plurality of data sets, each data set having a value xn and the validity vn thereof:
IL={(x1,v1),(x2,v2), . . . ,(xn,vn)}
The data sets can be disposed in the sequenced information list IL, for example, in accordance with the magnitudes of the values of the respective validity v1, v2, v3, . . . , vn. The value x1 having the highest validity v1 is thus in first place, followed by the value x2 having the second-highest validity v2, and so forth.
A sequenced information list IL of this kind can be created in particular when, for one of the values xn, its validity vn is greater than or equal to the minimum value, and that value xn with its validity vn would therefore not be taken into account in further processing.
For example, the sequenced information list IL with its individual data sets (i.e. value xn and validity vn) can be displayed to a user in accordance with its validity, for example in a sequence that begins with the value xn having the highest validity vn, followed by the value x2 having the second-highest validity v2, and so forth.
A method sequence will now be explained with additional reference to
What is depicted is that at the beginning, sensor 41 requires a first minimum time period M1, for example until a first zero crossing or the like, in order to furnish the first information item I1. Sensor 41 accordingly outputs at the beginning, during the time span M1, the value x1=0 and validity v1=0. Also depicted is the fact that during a later time span T1, an uncertainty exists in the measured value conditioning, and sensor 41 accordingly reduces the validity to v1=0.3.
The values x1 and validities v1 are transferred from sensor 41 to control unit 6. In the control unit, an interim data set ZD is created therefrom for further processing. Control unit 6 is furthermore embodied to carry out a comparing step before the evaluating step. In the comparing step, the validity v1 is compared with a minimum value that, in the present example, is equal to more than 0.3. If the validity does not exceed the minimum value, then in accordance with the embodiment depicted here, calculation occurs in the control unit using a substitute value xE1=0. This can be, for example, a reliable standard value for the respective sensor. It is thereby possible to prevent possibly erroneous values from leading to an undesired result.
What is depicted is that at the beginning, sensor 42 requires a second minimum time period M2 in order to furnish the second information item I1. Sensor 42 accordingly outputs at the beginning, during the time span M2, the value x2=1 and validity v2=0. Also depicted is the fact that during a later time span T2, an uncertainty exists in the measured value conditioning, and sensor 42 accordingly reduces the validity to v2=0.9.
The values x2 and validities v2 are transferred from sensor 42 to control unit 6. In the control unit, an interim data set ZD is created therefrom for further processing. Control unit 6 is furthermore embodied to carry out a comparing step before the evaluating step. In the comparing step, the validity v2 is compared with a minimum value that, in the present example, is equal to more than 0.9. If the validity does not exceed the minimum value, then in accordance with the embodiment depicted here, calculation occurs in the control unit using a substitute value xE2=1. This can be, for example, a reliable standard value for the respective sensor. It can also be the most recent valid value or an average of several most recent valid values.
Control unit 6 is embodied or configured to perform an evaluating step in order to determine, using the values x1 and x2 or the respective substitute values xE1 and xE2 (the alternatives are represented by x(E)i) and their validities v1 and v2, as a result of multiplying the two information items I1 and I2, a resultant data set ED having a resultant value xE and a resultant validity vE. An evaluating step of this kind is illustrated in
Conventionally, the product x1*x2 would be calculated for this purpose. According to the embodiment of the invention which is presented, however, the validity is taken into consideration, as follows:
I
1
I
2=(x(E)1,v1)(x(E)2,v2)=(x(E)1*x(E)2,v1*v2).
It is evident that the resultant validity vE, constituting the product of validities v1, v2, is reduced, i.e. is less than one, during the first minimum time period M1, the second minimum time period M2, the first time interval T1, and the second time interval T2, while between the second minimum time period M2 and the first time interval T1, between the first time interval T1 and the second time interval T2, and after the second time interval T2, the resultant validity vE is one.
The resultant value xE is thus reliable and trustworthy between the second minimum time period M2 and the first time interval T1 and after the first time interval T1, and not trustworthy (ve=0) or partly trustworthy (0<ve<1) in the other time intervals.
What is therefore available with the resultant validity vE is a value with which the validity of the resultant value xE can be assessed. The resultant validity vE indicates the trustworthiness of the processing result in the control unit.
| Number | Date | Country | Kind |
|---|---|---|---|
| 102015200721.9 | Jan 2015 | DE | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2016/050008 | 1/4/2016 | WO | 00 |
