DE102022130493A1 - Computer-implemented method, apparatus and computer program for selecting signal paths of a graph-based network to be validated - Google Patents

Abstract

Embodiments of the present invention relate to a computer-implemented method, to a device and to a computer program for selecting signal paths of a graph-based network to be validated. The method comprises obtaining (110) information about a plurality of input signals and a plurality of output signals of the graph-based network to be validated. The method further comprises calculating (120), for the plurality of input signals and the plurality of output signals, a number of internal signal sections between the respective input signal and the respective output signal, if a signal path exists between the respective input signal and output signal. The method further comprises calculating (130), by means of a statistical function, for the plurality of input signals and/or the plurality of output signals, a statistical value that represents the signal paths between an input signal and the plurality of output signals or the signal paths between an output signal and the plurality of input signals. The method further comprises selecting (140), based on particularly high and/or particularly low statistical values, one or more input signals and/or output signals for which the particularly high or low statistical values were calculated. The method further comprises providing (190) information about signal paths to be validated, wherein the signal paths were selected based on the selected one or more input signals and/or output signals.

Description

Embodiments of the present invention relate to a computer-implemented method, a device and a computer program for selecting signal paths of a graph-based network to be validated.

Complex signal networks are a way of representing software functions with their internal dependencies, as is necessary, for example, in the documentation for the approval of vehicles in the USA. The complex signal networks are used to summarize data dependencies between signals, and in particular between input signals and output signals, in a model. Applications include the aforementioned analysis for the purpose of approval, but also the improvement of the software architecture, the safeguarding of influences on functional safety and the test management of new software functions.

If a large signal network is to be evaluated in terms of its connections and dependencies, sufficient validation of the information it contains can or should be carried out in advance during the signal network creation process (regardless of the exact type of creation or specific further use). Complete protection of the signal network is not possible due to time and resource constraints and practical reasons, as several million connections would usually have to be secured.

The aim is therefore to create a sufficiently large and robust security width of the signal network so that the signal network can be trusted in its informative value during the subsequent analysis.

The computer-implemented method, the device and the computer program according to the independent claims take this need into account.

The present invention is based on the finding that a statistical evaluation of the path lengths of signal paths in a signal network to be validated offers a useful starting point for reducing the number of signal paths to be validated, which, with a significant reduction in the number of signal paths to be validated, brings with it a large and robust security range of the signal network. Based on the input signals and output signals of the signal network to be validated, input signals and output signals are selected for which the calculated statistical values are particularly high or particularly low. Based on these input signals and/or output signals, signal paths (linked to the input signals and/or output signals) can now be selected for validation. Additional techniques, such as calculating a path index or path coding, can further specify this preselection. The result is a selection of signal paths that brings with it a large and robust security range of the signal network, with a significantly reduced validation effort.

One aspect of the present invention relates to a computer-implemented method for selecting signal paths of a graph-based network to be validated. For example, the graph-based network can map a functionality of a control unit of a vehicle. The method comprises obtaining information about a plurality of input signals and a plurality of output signals of the graph-based network to be validated. The method further comprises calculating, for the plurality of input signals and the plurality of output signals, a number of internal signal sections between the respective input signal and the respective output signal, if a signal path exists between the respective input signal and output signal. The method further comprises calculating, by means of a statistical function, for the plurality of input signals and/or the plurality of output signals, a statistical value that represents the signal paths between an input signal and the plurality of output signals or the signal paths between an output signal and the plurality of input signals. The method further comprises selecting, based on particularly high and/or particularly low statistical values, one or more input signals and/or output signals for which the particularly high or low statistical values were calculated. The method further comprises providing information (for classification) about signal paths to be validated, wherein the signal paths were selected based on the selected one or more input signals and/or output signals. The signal paths selected here can now be used for validation.

In particular, a calculation of the coefficient of variation can be used as a statistical function. In other words, starting from an input signal or an output signal, the coefficient of variation can be calculated over the length of the signal paths (i.e. the number of internal signal sections) to the output signals or input signals connected to this input signal. The coefficient of variation has proven to be a suitable selection criterion in tests.

When selecting the input signals and/or output signals, a statistical value can be considered particularly high or particularly low if the statistical value is among the n highest or m lowest statistical values, where n and m are predefined positive integers. Alternatively or additionally, a statistical value can be considered particularly high if the statistical value reaches at least 95% of the highest statistical value and/or a statistical value can be considered particularly low if the statistical value reaches at most 105% of the lowest statistical value. Both are suitable approaches for identifying particularly high or low statistical values.

Depending on the use case, it may be useful to restrict the set of input and output signals in advance. For example, the majority of input signals and the majority of output signals can be signals selected by a user from the totality of input signals and output signals of the graph-based network. This allows an a priori restriction of the signal paths to be validated.

Based on the selected input and/or output signals, signal paths can now be determined that provide the basis for the further selection filters. The method can include determining a plurality of signal paths based on the selected one or more input signals and/or output signals. The signal paths of the plurality of signal paths start with one of the selected input signals or terminate with one of the selected output signals. The method can now include selecting the signal paths for the information about signal paths to be validated from the plurality of signal paths. In other words, based on the selected input and/or output signals, signal paths are identified that are related to the selected input and/or output signals. This provides the basis for the subsequent selection.

The further selection can now be based on a metric. The method can include calculating, for the plurality of signal paths, a metric that characterizes the respective signal path. The metric can be based on at least one of the number of internal signal sections of the respective signal path, a number of internal signal inputs that contribute to the signal path, and a number of internal signal outputs. The method can include sorting the plurality of signal paths according to the metric and selecting the signal paths among the plurality of signal paths for the information about signal paths to be validated based on the sorting. The metric allows the identification of similar signal paths, among which one or a few can be selected for validation.

To further support the selection, a partially manual method can also be used, which is supported by an automated, graphical coding of the signal paths. The method can include creating a graphical representation for the sections of the respective signal paths between the input signal and the output signal (hereinafter also referred to as path coding). The graphical representation of the sections can be based at least on a ratio between internal signal inputs and internal signal outputs of the respective section. Furthermore, the graphical representation of the sections can be based on the absolute number of internal signal inputs and/or internal signal outputs. The method can also include calculating a graphical overview of the plurality of signal paths, sorted according to the metric based on the graphical representation of the sections, and providing the information about the signal paths to be validated with the graphical overview of the plurality of signal paths. Such a representation can be used by an expert to select suitable signal paths for validation from the remaining signal paths.

Additionally or alternatively, signal sub-paths that are used particularly frequently can also be selected automatically. The method can include selecting, from the plurality of signal paths, signal sub-paths that are contained in a plurality of signal paths. The method can include selecting the signal paths from the plurality of signal paths for the information about signal paths to be validated based on the selected signal sub-paths. For example, based on the selected signal sub-paths, one or more signal paths that include the selected signal sub-paths can be determined.

One aspect of the present invention relates to a computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out the aforementioned method.

One aspect of the present invention relates to a device comprising one or more processors and one or more memory devices configured to carry out the aforementioned method.

• 1a zeigt ein Flussdiagramm eines Beispiels eines computerimplementierten Verfahrens zur Auswahl von Signalpfaden eines zu validierenden graphenbasierten Netzwerks;
• 1b zeigt ein schematisches Diagramm einer Vorrichtung zur Ausführung eines computerimplementierten Verfahrens zur Auswahl von Signalpfaden eines zu validierenden graphenbasierten Netzwerks;
• 2 zeigt ein schematisches Diagramm einer Hitze-Karte (Heat-Map) -Darstellung einer Verbindungsmatrix eines Signalnetzwerks;
• 3 zeigt ein Diagramm eines einfachen Beispiels zur Bestimmung der Parameter eines Pfadindex;
• 4 zeigt ein Kennfeld des Pfadindex;
• 5 zeigt ein schematisches Diagramm einer Unterteilung von Knoten in unterschiedliche Kategorien;
• 6 zeigt ein schematisches Diagramm einer Pfadkodierung für eine beispielhafte Auswahl von Signalpfaden mit der gleichen Quelle; und
• 7 zeigt eine schematische Darstellung einer Verwendungshäufigkeit von Signalknoten in den Signalpfaden.

Examples of embodiments are explained in more detail below with reference to the accompanying figures. They show:

• 1a shows a flowchart of an example of a computer-implemented method for selecting signal paths of a graph-based network to be validated;
• 1b shows a schematic diagram of an apparatus for carrying out a computer-implemented method for selecting signal paths of a graph-based network to be validated;
• 2 shows a schematic diagram of a heat map representation of a connection matrix of a signal network;
• 3 shows a diagram of a simple example of determining the parameters of a path index;
• 4 shows a map of the path index;
• 5 shows a schematic diagram of a division of nodes into different categories;
• 6 shows a schematic diagram of a path coding for an exemplary selection of signal paths with the same source; and
• 7 shows a schematic representation of a usage frequency of signal nodes in the signal paths.

Some examples will now be described in more detail with reference to the accompanying figures. However, other possible examples are not limited to the features of these detailed embodiments. These may include modifications of the features as well as equivalents and alternatives to the features. Furthermore, the terminology used herein to describe particular examples is not intended to be limiting of other possible examples.

The present invention is particularly concerned with the identification of sample validation paths in complex signal networks.

The development of control units or other devices is now software-driven, with many control unit functions being developed as software functions. The behavior of the software and existing data dependencies can be modeled using signal networks. To efficiently validate these signal networks, the following procedure can be used. First, the signal network(s) of interest are created. The data dependencies are extracted from software components, for example by means of (static or dynamic) code analysis or manually or otherwise. Furthermore, connection knowledge from the software function(s) is stored. Next, a subset of signal paths is selected. The selected signal paths from the network are then validated between predefined sources (start of the path, in connection with the 1a and 1b also called input signals) and predefined sinks (end of the path, in connection with the 1a and 1b (also called output signals). In this case, a functional documentation or directly the software code, which represents the target, is compared with the signal network, which represents the actual.

The present invention deals with the selection of the subset of signal paths. This bridges the gap between an existing signal network (with potentially millions of signal paths) and the path validation (which is limited to an affordable scale of hundreds of connections). In the proposed concept, a qualified selection of validation paths is carried out that allow sufficient security for the analysis use case of the signal network.

1a shows a flowchart of an example of a computer-implemented method for selecting signal paths of a graph-based network to be validated, by which the proposed concept is implemented.

1b shows a schematic diagram of a corresponding device 10, with one or more processors 14, one or more storage devices 16, and further optional components, wherein the device 10 is designed to carry out the computer-implemented method of 1a to execute. In 1b the one or more processors 14 are coupled to an interface 12 and one or more storage devices 16. The functionality of the processor is determined by software. Therefore, the device 10 can include machine-readable instructions, for example in the one or more storage devices, which are executed by the processor. The device also represents an aspect of the proposed concept. The device 10 can be a computer-implemented device and can be included in a computer system 100.

In the following, the proposed concept is explained with reference to the procedure of 1a However, it is clear that features discussed in connection with the method also apply to the corresponding device and to a corresponding computer program.

The proposed concept starts with the creation of an initial list of sources (hereinafter also referred to as input signals) and sinks (hereinafter also referred to as targets) that describe the endpoints of the analysis for the corresponding use case. Accordingly, the method of 1a obtaining 110 information about a plurality of input signals and a plurality of output signals of the graph-based network to be validated.

The proposed concept follows a graph-based approach, i.e. the network to be validated is modeled using a graph. The graph-based network comprises a plurality of nodes and a plurality of edges. The nodes of the graph-based network comprise the input signals and output signals as input nodes and output nodes. The nodes also comprise other nodes that represent neither an input signal nor an output signal. These nodes represent processing steps that are carried out within the network and thus internal signals (intermediate results). The edges represent the data dependencies between the nodes of the network. Signal paths are spanned between an input node (which represents an input signal/source) and an output node (which represents an output signal/sink), and comprise the input node, the output node, as well as the edges and internal nodes between the input node and the output node. In many cases there are several signal paths between an input node and an output node. In this case, for example, one of the signal paths between the input and output nodes can be selected based on a metric, such as the shortest one.

This graph-based network maps a certain functionality. In the following, the proposed concept is used to validate a vehicle ECU. Consequently, the graph-based network can map a functionality of a vehicle ECU. However, the proposed concept can also be applied to other use cases, such as ECUs in industrial plants, other software-based sensor data processing use cases, or even use cases that are not based on an ECU application, such as Software in the Loop (a software development concept from control engineering).

In the present example, the sources/input signals can correspond to the input sensor signals (and input bus messages) of a control unit, i.e. sources = [input sensor data 1, input sensor data 2, .... X]. The output signals/sinks can correspond to the output actuator signals (and bus messages and internal error memories) of the same control unit, i.e. sinks = [output actuator 1, output actuator 2, .... Y]. Depending on the application, a pre-selection of sources and sinks can be made. In other words, the majority of input signals and the majority of output signals can be signals selected by a user from the totality of the input signals and output signals of the graph-based network (e.g. of the control unit). The individual sources and sinks correspond to findable nodes (signals) from the existing signal network, i.e. input or output nodes of the signal network.

To clarify the dimensions, size arrangements for the validation of a vehicle control unit are given below. Such a signal network, as examined here, comprised a six-digit number of nodes, a seven- to eight-digit number of edges, and based on this a number of signal paths (i.e. tuples of edges strung together between connected nodes) that was several orders of magnitude higher than the number of edges.

Such a signal network can be generated, for example, from a graph database such as Neo4J. A query of the database can be performed and a connection matrix [sources x sinks] can be generated. This allows an automated creation of the connection matrix of all source-sink combinations from the signal network. This connection matrix represents the existing connections between the sources and sinks. In addition to the actual existence of a connection between source and sink, i.e. between input signal and output signal, this connection matrix can also provide information about the length of the connection. Therefore, the method of 1a calculating 120, for the plurality of input signals and the plurality of output signals, a number of internal signal sections between the respective input signal and the respective output signal, if a signal path exists between the respective input signal and output signal. The length of the signal path corresponds to the number of internal nodes between the input node (ie the source/input signal) and the input node (ie the sink/output signal). Other calculation methods are also possible, for example based on the edges between the input node and the output node. The signal paths and the length of the signal paths can be calculated by the graph database, for example using the Dijkstra algorithm.

This length of the signal paths can now be displayed in the form of a heat map. In the heat map, as shown in 2 is shown, only the shortest signal path between a source and a sink is shown. The heat map of 2 represents over 100,000 signal paths. In other words, the connection matrix can be converted into a heat map representation of the individual signal path connection lengths, where the connection length indicates how many signals (internal nodes) lie between source and sink per path.

2 shows a schematic diagram of a heat map 200 (heat map) representation of a connection matrix of a signal network. The sinks/output signals are plotted on the X-axis, and the sources/input signals/sources on the Y-axis. The input signals in the example of the control unit are sensor signals or input bus messages (ie messages via a vehicle bus that are directed to the control unit), and the output signals are actuator control signals or output bus messages. In the heat map representation of 2 Dark areas represent long connections (ie, long signal paths with many internal nodes), light areas represent short connections (ie, short signal paths with few internal nodes), and empty areas represent non-existent connections.

A pre-selection is now made based on this heat map (but not necessarily based on the graphical representation of the heat map). This can be based on which lines are particularly uniform or particularly different. After this pre-selection, for example, there are still thousands of signal paths left.

In the present concept, a statistical approach is chosen to select the rows and/or columns from the connection matrix. The method of 1a includes calculating 130, by means of a statistical function, for the plurality of input signals and/or the plurality of output signals, a statistical value that represents the signal paths between an input signal and the plurality of output signals or the signal paths between an output signal and the plurality of input signals. In other words, a statistical value is calculated for each input signal and/or each output signal, which is based on the signal paths (and in particular their lengths) to the (approximately all) output signals or input signals. In an exemplary implementation, the coefficient of variation was used, ie the statistical function can correspond to the calculation of the coefficient of variation. However, other statistical functions are also possible, such as a different representation of the variation within the respective column/row. In 2 a row-by-row/column-by-column calculation of the coefficient of variation was carried out in the heat map. The coefficient of variation corresponds to the relative ratio of the variation (σ) of the connection length and the mean value of the connection length (ie length of the signal path). This is used for pre-selection from the heat map in order to select rows/columns that are particularly uniform or different, ie whose coefficient of variation is as small or large as possible.

In 2 such a preselection is shown. Row 210 is a row with a low coefficient of variation (for the same input signal/source). Row 220 is a row with a high coefficient of variation (same input signal). Column 230 is a column with a high coefficient of variation (for the same output signal/sink). Column 240 is a column with a low coefficient of variation (same output signal).

It has been shown that particularly high and particularly low values of the coefficient of variation of the respective column or row are good starting points for further validation. Therefore, a bookend analysis (according to rows and columns with particularly high and particularly low coefficients of variation) can be carried out to preselect paths from the heat map. Several complete rows or columns are taken from the heat map. The method of 1a therefore includes selecting 140, based on particularly high and/or particularly low statistical values, one or more input signals and/or output signals (ie, one or more rows and/or columns) for which the particularly high or low statistical values were calculated. In this case, a statistical value can be considered particularly high if the statistical value reaches at least 95% of the highest statistical value. A statistical value can be considered particularly low if the statistical value reaches at most 105% of the lowest statistical value. For example, a statistical value can be considered particularly high or particularly low if the statistical value is among the n highest or m lowest statistical values, where n and m are predefined positive integers.

The selection can be made both automatically and manually (using a graphical representation of the heat map). In other words, the pre-selection can also be an expert-based and purely visual selection of the path packages. Alternatively, when pre-selecting the rows and/or columns, a specification can be made as to how many Variation coefficient variants should be selected (e.g. by appropriate specification of n and m). This is done, for example, depending on the available validation resources.

The one or more input signals result in a plurality of signal paths that are defined by the respective input-output signal tuples. Consequently, the method can comprise determining 150 a plurality of signal paths based on the selected one or more input signals and/or output signals. The signal paths of the plurality of signal paths start with one of the selected input signals or terminate with one of the selected output signals. In other words, based on the selected rows and/or columns of the connection matrix, signal paths are selected that start at the input signal of the respective selected row or terminate at the output signal of the respective selected column.

These signal paths can now be further analyzed in order to make a specific selection from this pre-selection. For example, a measure can be used for this, which is also referred to below as a path index. This can later be used to order the paths in the final selection. The method of 1a can optionally include calculating 160, for the plurality of signal paths, a metric that characterizes the respective signal path. Three criteria are of particular interest here - the length of the signal path, and the number of internal signal inputs and the number of internal signal outputs. Accordingly, the metric can be based on at least one of the number of internal signal sections of the respective signal path (ie, the length of the signal path), a number of internal signal inputs that contribute to the signal path, and a number of internal signal outputs. The metric thus allows a conclusion to be drawn about the behavior of the internal signal nodes in the signal path. The following formula can be used, for example: $$\text{Pfadinex}=\sqrt[n]{\frac{{\displaystyle \sum \text{Entry}\ddot{\text{a}}\text{nge}}}{{\displaystyle \sum \text{Issue}\ddot{\text{a}}\text{nge}}}}$$

Where n (which is different from the previously used n) is the length of the signal path (ie the number of internal nodes/signals), and ΣInputs and ΣOutputs are the number of internal signal inputs and internal signal outputs of the internal nodes. 3 shows a diagram of a simple example of determining the parameters of the path index. 3 shows the nodes source 310 (input signal), internal node 320 with one internal input signal (in addition to the signal path) and three internal output signals, internal node 330 with two internal input signals and four internal input signals, and sink node 340 (output signal) with five internal input signals. Internal nodes are framed by a bracket 350. The sum of the internal signal inputs of the internal nodes is 3, the sum of the internal signal outputs of the internal nodes is 7, and n = 2 is the number of internal nodes.

4 shows a map of the path index. In 4 From top to bottom, the ratio of the sum of the internal signal inputs and the sum of the internal signal inputs is entered, from a ratio of 6000/1 to 1/6000. From left to right, the resulting path index after calculating the n-th root is shown.

The path index allows classification and sorting of different paths next to each other. Accordingly, the method may comprise sorting 162 the plurality of signal paths according to the metric. Signal paths with a similar ratio between internal signal inputs and internal signal outputs, obtained as in 4 shown, a similar measure, and are grouped together by the sorting (and thus categorized). The length of the signal paths can also have an influence on the sorting, so that signal paths of similar length may be grouped together by the sorting. With the help of the path index, distribution paths can be distinguished from consumer paths and short paths from long paths.

Based on the generated 150 signal paths, for example based on the sorting according to the metric, signal paths can now be selected among the majority of signal paths for information about signal paths to be validated based 170 as in 1a is shown.

A process can be used for selection, which can be automated or partially manual. A graphical approach, i.e. a representation-based approach, can be used to select the signal paths. The remaining paths (from source to sink) can be path-coded automatically. The path coding is a graphical representation that is shown in the 5 and 6 Path components (ie internal nodes) are represented differently, which have different ratios between internal input signals and internal output signals. Internal nodes are divided into different categories and coded graphically, for example color-coded or with different patterns (as in 5 to 7 shown).

5 shows a schematic diagram of a division of nodes (path components) into the different categories. Input signals (sources) and output signals are shown with white boxes, they are in 5 not shown, but in the 6 and 7 . In addition, there are the categories distributor, mass distributor, balanced, mass balanced, consumer and mass consumer. In the example, distributors are internal nodes that have more than 2 internal signal outputs more than internal signal inputs, but a total of a maximum of 12 internal signal outputs (parameter A in 5 ). Mass distributors are internal nodes that have more than 2 internal signal outputs than internal signal inputs, but a total of more than 12 internal signal outputs. Balanced are internal nodes that have approximately the same number of internal signal inputs as internal signal outputs (±2, parameter D in 5 ), but a maximum of 20 internal signal inputs or internal signal outputs (parameter C in 5 ). Mass balanced are internal nodes that have approximately the same number of internal signal inputs as internal signal outputs (±2), but more than 20 internal signal inputs or internal signal outputs. Consumers are internal nodes that have more than 2 internal signal inputs more than internal signal outputs, but a maximum of 15 internal signal outputs in total (parameter B in 5 ). Mass consumers are internal nodes that have more than 2 internal signal inputs than internal signal outputs, but a total of more than 15 internal signal outputs. However, a different path coding is possible, depending on the application, and in particular depending on the number of internal nodes. The limit values (parameter AD) are set differently depending on the signal network in question, so that an even distribution of the path coding classes can occur.

The procedure of 1a optionally includes creating 180 a graphical representation for the sections of the respective signal paths between the input signal and the output signal. The sections (internal nodes) are graphically coded differently, for example as described above. The graphical representation of the sections, as explained above, can be based at least on the relationship between internal signal inputs and internal signal outputs of the respective section. In addition, the graphical representation of the sections, as in 5 shown, further based on the absolute number of internal signal inputs and/or internal signal outputs. The method may include calculating a graphical overview of the plurality of signal paths, sorted according to the metric, based on the graphical representation of the partial paths. The graphical overview may be stored in a memory of the computer system executing the method (such as in a storage device 16), and/or output as a signal (such as via the interface 12), such as via a screen or via a computer network (and a browser).

In the graphical representation, the paths can be aligned with each other (left, right, center), coded graphically (color or pattern) and counted statistically. From the path coding, you can read which parts of the path are particularly frequently crossed in the application-based connection matrix. The assignment of different graphic categories and sorting by path index allows "common sections" of the validation paths to be visualized.

6 shows a schematic diagram of a path coding for an exemplary selection of signal paths with the same source. 6 shows 10 lines 600-690 as an example, which graphically represent ten different signal paths. The signal paths extend from the source on the left edge (shown as a white box) to the respective sink (also shown as a white box). In 6 In lines 600-620, 660-690 the sink is not shown because the representation is cut off here. The different lines correspond to different paths. For example, the same input signals/sources can be shown aligned to the left, the same output signals/sinks are shown aligned to the right.

This graphical representation can help an expert to select signal paths that seem particularly relevant, for example because they include sections that are shared by a particularly large number of other signal paths.

This selection can also be supported or carried out automatically. In order to select 170 the signal paths, the path parts that occur most frequently can be extracted from all the path codings carried out. This creates a high level of efficiency and breadth in the subsequent validation/validation. The method can also (automatically) select 164, from the plurality of signal paths, signal sub-sections that are contained in several signal paths. The selection 170 of the signal paths from the plurality of signal paths for the information about signal paths to be validated can now be carried out based on the selected signal sub-sections. Here, the path coding enables the frequency of use of the same signal nodes to be counted.

Such an evaluation is in 7 shown. 7 shows a schematic representation of a frequency of use of signal nodes in the signal paths. In 7 is a series of 700 most frequently used signal nodes are shown, which occur in as many signal paths as possible in this order. Line 710 below shows in what percentage of the paths the respective signal nodes are used. Line 720 shows in how many paths (absolute) the signal nodes are used. The path shown allows the coverage of at least 61 total paths in the validation.

By iteratively repeating the path coding and selecting frequently occurring path sections for validation, the efficiency of validation can be further increased for the corresponding use case, since each selected path section occurs more than once. This means that its validation (or falsification) helps to improve the quality of several other existing connections. This leads to a robust increase in the scope of assurance.

All the previously used features can be used to (automatically or semi-automatically) select 170 the signal paths for the information on signal paths to be validated among the plurality of signal paths. The method of 1a further comprises providing 190 the information about signal paths to be validated, for example as a signal via a screen or via a computer network (via the interface 12). The signal paths were selected fundamentally based on the selected one or more input signals and/or output signals (for example based on the rows and/or columns of the heat map). Various techniques were presented to further refine the selection, based on the path index and the path coding, for example. Therefore, the information about the signal paths to be validated can be provided with the graphical overview of the plurality of signal paths, ie with the path coding. Either the result of the selection can now be validated, or an expert can now use the graphical representation to select the specific signal paths to be validated.

The one or more interfaces 12 of the device 10 may correspond to one or more inputs and/or outputs for receiving and/or transmitting information, which may be in digital (bit) values according to a specific code within a module, between modules, or between modules of different units. The one or more interfaces 12 may, for example, comprise interface circuits configured for receiving and/or transmitting information.

The one or more processors 14 of the device 10 can be implemented by one or more processing units, one or more processing devices, any means of processing, such as a processor, a computer, or a programmable hardware component that can be operated with appropriately adapted software. In other words, the described function of the one or more processors 14 can also be implemented in software that is then executed on one or more programmable hardware components. Such hardware components can include a general-purpose processor (such as a central processing unit), a digital signal processor (DSP), a microcontroller, etc.

The one or more storage devices 16 of the apparatus 10 may include at least one member of the group of computer-readable storage media, such as a magnetic or optical storage medium, e.g., a hard disk drive, a flash memory, a floppy disk, a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a network storage.

The aspects and features described in connection with a particular one of the previous examples may also be combined with one or more of the further examples to replace an identical or similar feature of that further example or to additionally introduce the feature into the further example.

Examples may further be or relate to a (computer) program having a program code for carrying out one or more of the above methods when the program is executed on a computer, a processor or another programmable hardware component. Steps, operations or processes of various of the methods described above may thus also be carried out by programmed computers, processors or other programmable hardware components. Examples may also cover program storage devices, e.g. digital data storage media, which are machine-, processor- or computer-readable and encode or contain machine-executable, processor-executable or computer-executable programs and instructions. The program storage devices may be, e.g. digital memories, magnetic storage media such as for example, magnetic disks and tapes, hard disk drives or optically readable digital data storage media. Further examples may also cover computers, processors, controllers, field-programmable logic arrays ((F)PLAs = (Field) Programmable Logic Arrays), field-programmable gate arrays ((F)PGA = (Field) Programmable Gate Arrays), graphics processor units (GPU = Graphics Processor Unit), application-specific integrated circuits (ASIC = application-specific integrated circuit), integrated circuits (IC = Integrated Circuit) or system-on-a-chip (SoC = System-on-a-Chip) programmed to carry out the steps of the methods described above.

It is further understood that the disclosure of multiple steps, processes, operations, or functions disclosed in the specification or claims should not be construed as being necessarily in the order described, unless explicitly stated in the individual case or required for technical reasons. Therefore, the foregoing description does not limit the performance of multiple steps or functions to a particular order. Furthermore, in further examples, a single step, function, process, or operation may include and/or be broken down into multiple sub-steps, functions, processes, or operations.

If some aspects in the previous sections were described in connection with a device or a system, these aspects are also to be understood as a description of the corresponding method. For example, a block, a device or a functional aspect of the device or system can correspond to a feature, such as a method step, of the corresponding method. Accordingly, aspects described in connection with a method are also to be understood as a description of a corresponding block, a corresponding element, a property or a functional feature of a corresponding device or a corresponding system.

List of reference symbols

10

contraption

12

interface

14

processor

16

Storage device

100

Computer system

110

Obtaining information about input signals and output signals

120

Calculating the number of internal signal sections between input signal and output signal

130

Calculating a statistical value

140

Selecting input signals/output signals

150

Determine signal paths based on the selected input signals/output signals

160

Calculating a measure

162

Sorting the signal paths according to the measure

164

Selecting signal sections

170

Selecting signal paths

180

Creating a graphical representation

190

Providing information about signal paths to be validated

200

Heat map

210

Line with low coefficient of variation

220

Line with high coefficient of variation

230

Column with a high coefficient of variation

240

Column with a low coefficient of variation

310

source

320, 330

Internal nodes

340

depression

350

bracket

600-690

Path-encoded signal paths

700

Graphically coded sections

710

Percentage of usage

720

Absolute frequency of use

Claims (13)

Computer-implemented method for selecting signal paths of a graph-based network to be validated, the method comprising: Obtaining (110) information about a plurality of input signals and a plurality of output signals of the graph-based network to be validated; calculating (120), for the plurality of input signals and the plurality of output signals, a number of internal signal sections between the respective input signal and the respective output signal, if a signal path exists between the respective input signal and output signal; calculating (130), by means of a statistics function, for the plurality of input signals and/or the plurality of output signals, a statistical value that represents the signal paths between an input signal and the plurality of output signals or the signal paths between an output signal and the plurality of input signals; and selecting (140), based on particularly high and/or particularly low statistical values, one or more input signals and/or output signals for which the particularly high or low statistical values were calculated; and providing (190) information about signal paths to be validated, wherein the signal paths were selected based on the selected one or more input signals and/or output signals. The procedure according to Claim 1 , where the statistical function corresponds to the calculation of the coefficient of variation. The procedure according to one of the Claims 1 or 2 , where a statistical value is considered to be particularly high or particularly low if the statistical value is among the n highest or m lowest statistical values, where n and m are predefined positive integers, or where a statistical value is considered to be particularly high if the statistical value is at least 95% of the highest statistical value and/or where a statistical value is considered to be particularly low if the statistical value is at most 105% of the lowest statistical value. The procedure according to one of the Claims 1 until 3 , wherein the plurality of input signals and the plurality of output signals are signals selected by a user from among the total of the input signals and output signals of the graph-based network. The procedure according to one of the Claims 1 until 4 , further comprising determining (150) a plurality of signal paths based on the selected one or more input signals and/or output signals, wherein the signal paths of the plurality of signal paths begin with one of the selected input signals or terminate with one of the selected output signals, and selecting (170) the signal paths for the information on signal paths to be validated among the plurality of signal paths. The procedure according to Claim 5 , further comprising calculating (160), for the plurality of signal paths, a metric characterizing the respective signal path, sorting (162) the plurality of signal paths according to the metric, and selecting (170) the signal paths among the plurality of signal paths for the information about signal paths to be validated based on the sorting. The procedure according to Claim 6 , wherein the measure is based on at least one of the number of internal signal sections of the respective signal path, a number of internal signal inputs contributing to the signal path, and a number of internal signal outputs. The procedure according to one of the Claims 6 or 7 , further comprising creating (180) a graphical representation for the partial sections of the respective signal paths between input signal and output signal, wherein the graphical representation of the partial sections is based at least on a ratio between internal signal inputs and internal signal outputs of the respective partial section, calculating a graphical overview of the plurality of signal paths, sorted according to the metric based on the graphical representation of the partial sections, and providing the information about the signal paths to be validated with the graphical overview of the plurality of signal paths. The procedure according to Claim 8 , whereby the graphical representation of the sections is further based on the absolute number of internal signal inputs and/or internal signal outputs. The procedure according to one of the Claims 5 until 9 , further comprising selecting (164), from among the plurality of signal paths, signal sub-paths included in a plurality of signal paths, and selecting (170) the signal paths from among the plurality of signal paths for the information on signal paths to be validated based on the selected signal sub-paths. The procedure according to one of the Claims 1 until 10 , where the graph-based network maps a functionality of a vehicle control unit. Computer program comprising instructions which, when the program is executed, are a computer to cause it to perform the procedure according to one of the Claims 1 until 11 to execute. Apparatus (10) comprising one or more processors (14) and one or more memory devices (16) designed to carry out the method according to one of the Claims 1 until 11 .

Images