阅读说明：本技术 用于重构预定的分布式实时仿真网络的计算机实现的方法 (Computer-implemented method for reconstructing a predetermined distributed real-time simulation network ) 是由 H·卡尔特 D·卢贝雷 于 2020-03-02 设计创作，主要内容包括：本发明示出且说明了一种用于重构(U)预定的分布式实时仿真网络(2)的计算机实现的方法(1),其中,仿真网络(2)具有多个网络节点(4、RK、R、IO)和多个数据连接(DV),其中,每个网络节点(4、RK、R、IO)具有至少一个数据连接接口以连接数据连接(DV),其中,网络节点(4、RK、R、IO)通过数据连接(DV)至少部分处于通信连接(KV)中,并且其中,在仿真网络(2)运行中在至少一个网络节点(4、RK、R、IO)上实施仿真应用(5)。用所述方法可以自动找到实时仿真网络(2)的一种结构,在该结构中,以如下方式减少并且尽可能避免处于临界状态的通信连接(KV),即,检测仿真网络(2)的拓扑,使得存在有关网络节点(4、RK、R、IO)和在网络节点(4、RK、R、IO)之间的数据连接(DV)的拓扑信息,特别是为仿真网络(2)的网络节点(4、RK、R、IO)确定节点数据率的预期值(E-KDR)和/或节点延时的预期值(E-KL)。(The invention shows and describes a computer-implemented method (1) for reconstructing (U) a predefined distributed real-time simulation network (2), wherein the simulation network (2) has a plurality of network nodes (4, RK, R, IO) and a plurality of data connections (DV), wherein each network node (4, RK, R, IO) has at least one data connection interface for connecting a data connection (DV), wherein the network nodes (4, RK, R, IO) are at least partially in a communication connection (KV) via the data connections (DV), and wherein a simulation application (5) is implemented on at least one network node (4, RK, R, IO) during operation of the simulation network (2). In this way, a structure of the real-time simulation network (2) can be automatically found in which the communication connections (KV) in a critical state are reduced and avoided as far as possible, in that the topology of the simulation network (2) is detected in such a way that topology information is available about the network nodes (4, RK, R, IO) and the data connections (DV) between the network nodes (4, RK, R, IO), in particular an expected value (E-KDR) of the node data rate and/or an expected value (E-KL) of the node delay for the network nodes (4, RK, R, IO) of the simulation network (2).)

1. Computer-implemented method (1) for reconstructing (U) a predefined distributed real-time simulation network (2), wherein the simulation network (2) has a plurality of network nodes (4, RK, R, IO) and a plurality of data connections (DV), wherein each network node (4, RK, R, IO) has at least one data connection interface for connecting a data connection (DV), wherein the network nodes (4, RK, R, IO) are at least partially in a communication connection (KV) via the data connections (DV), and wherein a simulation application (5) is implemented on at least one network node (4, RK, R, IO) during operation of the simulation network (2),

It is characterized in that the preparation method is characterized in that,

detecting the topology of the emulated network (2) such that topology information (6) exists about the network nodes (4, RK, R, IO) and the data connections (DV) between the network nodes (4, RK, R, IO),

determining an expected value (E-KDR) of a node data rate and/or an expected value (E-KL) of a node delay for a network node (4, RK, R, IO) of the simulation network (2),

an expected value (E-DVDR) of the data transmission rate is determined for the data connection (DV),

determining the communication connections (KV) between the network nodes (4, RK, R, IO) of the simulation network (2),

for the communication connection (KV), the expected value (E-KVDR) of the data rate of the communication connection and/or the expected value (E-KVL) of the delay of the communication connection is determined on the basis of the expected value (E-KDR) of the data rate of the nodes and/or the expected value (E-KL) of the delay of the nodes and/or the expected value (E-DVDR) of the data rate of the data connection (DV) participating in the communication connection,

determining a limit value (G-KVDR) for the data rate of the communication link and/or a limit value (G-KVL) for the delay of the communication link for the communication link (KV), and a limit value (G-DVDR) for the data rate for the data link (DV),

in an evaluation step (B), a communication link (KV) in a critical state is determined in that the determined expected value (E-KVDR) of the data rate of the communication link and/or the expected value (E-KVL) of the delay of the communication link and/or the expected value (E-DVDR) of the data rate of the data transmission rate are compared with a limit value (G-KVDR) of the data rate of the communication link and/or a limit value (G-KVL) of the delay of the communication link and/or a limit value (G-DVDR) of the data rate of the associated communication link (KV) and

In a reconstruction step (U), the predetermined simulation network (2) is reconstructed such that the communication links (KV) in the critical state are reduced.

2. Method (1) according to claim 1, characterized in that the topology of the emulated network (2) is obtained by invoking an information service implemented on a network node (4, RK, R, IO) of the emulated network (2), which information service, when invoked, returns information about which network nodes (4, RK, R, IO) the information service is directly connected to, in particular which data connection (DV) the information service is directly connected to, or by reading a file with topology information (6) of the predefined emulated network (2).

3. Method (1) according to claim 1 or 2, characterized in that in the reconstruction step (U) the communication connections (KV) in the critical state in the predetermined simulation network (2) are reduced in such a way that

-maintaining the topology of the simulation network (2) at least partially in the predetermined simulation network (2)

-functionally expanding and/or functionally shrinking the network nodes (4, RK, R, IO), and/or

-functionally expanding and/or functionally contracting said data connection (DV), and/or

-the communication connections (KV) between the network nodes (4, RK, R, IO) are directed differently, and/or

-said predetermined simulation network (2) being at least partially in case of a topology change of said simulation network (2)

-expanding at least one additional network node (4, RK, R, IO) and expanding at least one additional data connection (DV) and/or shrinking at least one existing network node (4, RK, R, IO) and shrinking at least one existing data connection.

4. Method (1) according to any one of claims 1 to 3, characterized in that the expected value (E-DVDR) of the data transmission rate of the data connection (DV) is obtained by summing the data rates (E-KDR) of the connected network nodes (4, RK, R, IO).

5. Method (1) according to any one of claims 1 to 4, characterized in that the channel capacity of the data connection (DV) is used as a limit value (G-DVDR) for the data transmission rate of the data connection (DV).

6. The method (1) according to any one of claims 1 to 5, characterized in that the expected value (E-KDR) of the node data rate and/or the expected value (E-KL) of the node latency of a network node (4, RK, R, IO) of the simulation network (2) is determined based on a hardware specification of the network node (4, RK, R, IO) of the simulation network (2), in particular without taking into account a simulation application (5) of the respective network node (4, RK, R, IO), in particular without taking into account a possible hardware parameterization of the network node (4, RK, R, IO).

7. Method (1) according to one of claims 1 to 6, characterized in that the expected value of the node data rate (E-KDR) and/or the expected value of the node delay (E-KL) of a network node (4, RK, R, IO) of the emulated network (2) is determined in the worst case in such a way that the maximum value of the node data rate and/or of the node delay is used for the expected value of the node data rate (E-KDR) and/or the expected value of the node delay (E-KL).

8. Method (1) according to one of the claims 1 to 7, characterized in that average values of the actual node data rates and/or node delays of the network nodes (4, RK, R, IO) of the simulation network (2) are determined from a plurality of real-time simulation networks configured to be operational and are selected as expected values of the node data rates (E-KDR) and/or expected values of the node delays (E-KL) of the network nodes (4, RK, R, IO) of the simulation network (2).

9. Method (1) according to any one of claims 1 to 5, characterized in that the expected value (E-KDR) of the node data rate and/or the expected value (E-KL) of the node delay of a network node (4, RK, R, IO) of the simulation network (2) is determined taking into account a simulation application (5) of the respective network node (4, RK, R, IO), in particular taking into account a possible hardware parameterization of the network node (4, RK, R, IO).

10. Method (1) according to claim 9, characterized in that the expected value of the node data rate (E-KDR) and/or the expected value of the node latency (E-KL) of the network node (4, RK, R, IO) is determined taking into account the calculation step size (TS) of periodic tasks and/or the assumed invocation rate and processing duration of aperiodic tasks, the size of the packets calculated and sent in the tasks, the configuration of the I/O functions, in particular the invocation rate and size of the I/O packets processed.

11. The method (1) according to any one of claims 1 to 5, characterized in that the expected value (E-KDR) of the node data rate of the network node (4, RK, R, IO) and/or the expected value (E-KL) of the node delay and/or the expected value (E-DVDR) of the data transmission rate of the data connection (DV) of the emulated network (2) is determined by measurements in the emulated network (2).

12. The method (1) according to any one of claims 1 to 5 or according to claim 11, characterized in that the expected value (E-KVDR) of the data rate of the communication connection and/or the expected value (E-KVL) of the delay of the communication connection is determined by measurements during operation of the simulation network (2).

13. Method (1) according to claim 12, characterized in that the expected value (E-KVL) of the communication connection delay is measured, in particular during operation of the simulation network (2), i.e. when the simulation application (5) is implemented on the network nodes (4, RK, R, IO), in such a way that a synchronized clock time is set for all network nodes (4, RK, R, IO) of the simulation network (2), each transmitting network node (4, RK, R, IO) sets a transmission time stamp for the data transmitted by it, and the last receiving network node (4, RK, R, IO) calculates the expected value (E-KVL) of the communication connection delay of the respective communication connection (KV) from the reception time determined by it and by evaluating the transmission time stamps.

14. Method (1) according to claim 12, characterized in that the expected value (E-KVL) of the communication connection delay is measured in such a way that a synchronized clock time is set for all network nodes (4, RK, R, IO) of the simulation network (2), an echo function is implemented in the network nodes (4, RK, R, IO) of the simulation network (2), and the network node (4, RK, R, IO) of a communication connection (KV) receiving in operation of the simulation network (2) or the network node (4, RK, R, IO) transmitting an echo request to the network node (4, RK, R, IO) transmitting accordingly or the network node (4, RK, R, IO) receiving in operation of the simulation network (2) of the same communication connection (KV), and the network node (4, RK, R, IO) sending the echo request determines an echo cycle time after receiving the echo signal and thereby an expected value (E-KVL) of the communication connection delay, in particular wherein the simulation network (2) is not running during execution of the echo method, i.e. the simulation application (5) is not implemented on the network node (4, RK, R, IO).

15. The method (1) according to any one of claims 1 to 14, characterized in that the method is implemented on a computer (7) which is connected to the simulation network (2) via a data connection (DVR) or the method (1) is implemented on a network node (4, RK, R, IO) of the simulation network (2) which is configured as a computational network node (RK).

16. Computer program product comprising instructions which, when the program is run by a computer, cause the computer to carry out the method according to any one of claims 1 to 15.

17. A computer-readable storage medium comprising instructions which, when the program is run by a computer, cause the computer to carry out the method according to any one of claims 1 to 15.

Technical Field

The invention relates to a computer-implemented method for reconstructing a predefined distributed real-time simulation network, wherein the simulation network has a plurality of network nodes and a plurality of data connections, wherein each network node has at least one data connection interface for connecting the data connections, wherein the network nodes are at least partially in communication connection via the data connections, and wherein a simulation application is implemented on at least one network node during the operation of the simulation network.

The invention relates to the field of control device development, in particular to the development of control devices which are used in large numbers in the automotive field, but are also used in space and aviation and for controlling other technical processes. Such control devices are today mostly small computers with I/O interfaces (Input/Output), which are often equipped with real-time-capable operating systems that allow complex tasks to be implemented on the control device, mostly in connection with control technology. The control system development is a core component of the technical development of a wide range of plant engineering systems known from industrial practice, for example and in particular from the technical field mentioned at the outset.

Background

The testing of a series of control devices used in the end product is the end of a plurality of upstream development steps of the control device application (often closed-loop control or open-loop control) to be implemented on the control device, wherein these development steps are usually described by a so-called V model or also by a V cycle. For this purpose, a distributed real-time simulation network as described at the beginning is required. During the initial development of applications that are important for the functionality of many technical devices, mathematical modeling, for example closed-loop control algorithms, is carried out on a computer having a mathematical graphical modeling environment, wherein a controller can be understood as a component of a control device. Furthermore, the environment of the control device is also mathematically modeled, since the interaction of the controller on the control device with the process to be controlled is of interest. In mathematical consideration of this functionality, real-time simulation is mostly unnecessary (off-line simulation).

In a next step, the previously designed closed-loop control algorithm is transmitted by means of a Rapid Control Prototype (RCP) to a hardware system that is capable of operation, mostly in real time, and is connected to the actual physical process, i.e., for example, to the motor vehicle motor, via a suitable I/O interface. Such real-time-capable hardware is generally independent of the series of control devices to be used later, and in this case involves the proof of the functional capability of the principle of closed-loop control designed previously in practice.

In a further step, within the scope of the automatic generation of the series code, it is then possible to implement closed-loop control on the target processor to be actually used in the series control. The target hardware is therefore close to the series control device in this step, but not identical to the series control device. In a further step, the hardware in the context of a loop test (HIL) is checked for a series of control devices which are usually present only in a later development phase. The series of control devices in which the entity is present in this step is connected to the functional simulator by means of the control device interface of the entity. The simulator simulates the required variables of the series control device to be tested and exchanges the input and output variables with the series control device.

The series of control devices thus tested in the context of the HIL simulation are ultimately installed in a "real" target system, i.e. for example in a motor vehicle, and are tested in a real physical environment which was previously simulated only in the simulation environment.

It is clear that to implement the different development steps, both in the form of rapid control prototypes and in the form of hardware-in-the-loop simulations, powerful real-time capable hardware must be present, which usually form a distributed real-time simulation network with a plurality of network nodes and a plurality of data connections. The network node may for example relate to a computing node, i.e. a small computer with a real-time operating system. It is also possible to refer to I/O nodes with which measurement data of a physical process are detected, for example, in a measurement technique, and then the digitized values are forwarded in the form of data packets into an emulation network, or control signals, also analog, are output in order to influence the physical process, for example, by actuating actuators. Communication nodes also belong to the network nodes, which are often connected to at least two further nodes and are used less frequently for signal generation or signal processing, but rather for signal conversion (for example, converting data according to a specific protocol) or for serialization of parallel data streams (routers), etc.

The network node has at least one data connection interface, via which the network node is physically connected to at least one further network node by means of a data connection. The network nodes are at least partially in communication connection via a data connection, wherein a communication connection may comprise a plurality of data connections when the communication connection is operated via a plurality of network nodes. Not every node has to exchange data with any other node on a mandatory basis, even though this is physically possible based on the physical architectural design of the emulated network. The network nodes of the distributed real-time simulation network, which for the sake of simplicity are also just nodes, are usually not at a great distance from one another, but can for example even be installed in a common housing of the simulator. By "distributed" is meant here that the network nodes exchange information via a data connection and this cannot normally be done by accessing a common memory. Real-time capability is an important feature because real-time simulation networks are typically connected with real-world physical processes. If only a simulation network is mentioned in the following, partly for simplicity, this always refers to a real-time simulation network.

The simulation application is implemented on at least one network node while the simulation network is running. This usually takes place on the compute node, more precisely with a real-time capable operating system, so that a sampling system (Abtastsystem) of closed-loop control techniques can be implemented. Digital sampling systems need to be able to perform calculations reliably and hence also complete calculations within a certain fixed time grid. In highly dynamic closed-loop control, for example, complete calculation steps of the simulation application or of tasks of the simulation application, i.e. partial functions, may need to be carried out in microseconds. For this purpose, the respective measurement data must be transmitted from the I/O network node to the network node implementing the simulation application within the microsecond grid via the data connections of the simulation network, and the respective manipulated variables of the network node implementing the simulation application must accordingly also be calculated within this time grid and transmitted to the I/O network node. Simulation applications implemented on network nodes may have different tasks that must be implemented at different computation steps.

Simulation application refers herein to an algorithm implemented on a network node. It is entirely possible to distribute the overall task of a closed-loop control technique to a plurality of network nodes for the computation, so that each of these computation network nodes implements a simulation application, wherein the different simulation applications then together form the overall function to be achieved.

In the case of RCP applications, the control device to be tested is simulated by means of a real-time simulation network, wherein the simulation network is then connected to the process to be actually controlled, i.e. via an I/O network node. In the case of HIL simulation, the environment of a series of control devices which has been developed in practice is simulated with a real-time simulation network, which is then in turn physically connected to the series of control devices via I/O computing nodes. In any case, the real-time simulation network is therefore connected to the actual technical physical process via the I/O interface of the particular network node and influences this technical physical process during operation.

In practice, the design of such a real-time simulation network described is critical, since the data connections and the communication connections implemented in the simulation network must be selected such that the data transmission rate achieved in practice does not exceed the channel capacity, i.e. the maximum achievable data transmission rate over the communication path. It may happen that the total amount of traffic is not well balanced in such real-time simulation networks, that certain communication connections operate within or even beyond the limits of their data transmission capacity, while other data connections are rarely fully utilized. The same applies to the delay times, which are also referred to as delays (Latenz) and which have to be accepted on a particular communication connection. In the event of unfavorable network design simulations, only long delays can be achieved by certain communication connections, while short delays can be achieved on other communication connections, so that a better distribution of the data flows or a better selection of the communication connections is also required here. The design of the described real-time simulation network requires a great deal of experience from the associated application engineers. The aim is to utilize the existing resources of the simulation network as fully as possible, since unused resources in the form of idle hardware may be associated with high costs. Reconstructing the simulation network according to the "Trial and Error" principle is not efficient and Error-prone.

Disclosure of Invention

The object of the present invention is therefore to specify a computer-implemented method for reconfiguring a system for a real-time simulation network for the development of control devices, with which a configuration for the real-time simulation network can be found in which communication links in critical states are reduced and avoided as far as possible.

The method according to the invention for solving the previously derived and explained tasks is firstly characterized in that the topology of the simulated network is detected, so that topology information is available about the network nodes and the data connections between the network nodes. The method steps are used for accurately detecting or obtaining the topology of the predetermined real-time simulation network.

An expected value of the node data rate and/or an expected value of the node delay is then determined for the network nodes of the emulated network. An expected value of the data transmission rate is also determined for the data connection. These different expected values relate to assumptions about the node data rate, node delay and/or data transfer rate over the data connection that may occur. Furthermore, a communication connection between the network nodes of the emulated network is determined. The aforementioned method steps can be performed in a different order.

The expected value of the data rate of the communication connection and/or the expected value of the delay of the communication connection is now determined for the determined communication connection on the basis of the expected values of the data rates of the network nodes participating in the communication connection and the nodes of the data connection and/or the expected values of the node delays and/or the expected values of the data transmission rates. Since the communication connections in the simulation network are the actually used data transmission paths between the network nodes via which the communication takes place, the communication connection data rates and the communication connection delays relate to the parameters of real interest in the simulation network.

Furthermore, a limit value for the data rate of the communication connection and/or a limit value for the delay of the communication connection is determined for the communication connection. A limit value for the data transmission rate is determined for the data connection. These limit values relate to the value of the data rate or delay of the communication connection and the data transmission rate over the data connection, which are also considered to be acceptable values. The limit value for the data transmission rate may relate, for example, to the channel capacity, i.e. the maximum possible data transmission rate of the communication connection or of a part of the communication connection, or may relate to a certain percentage of the capacity when a particular capacity should be planned for safety. However, the limit value for the data transmission rate can also be derived from a comparison of expected values of the data transmission rate or from an average value formed therefrom, if it is required that the data transmission rates on the different communication connections should deviate from one another only to a certain extent. The specific method of determining the different expected values and the different limit values of the data rate, the latency and the data transmission rate of the communication connection is not critical to the implementation of the general inventive concept, but is critical: first-regardless of the specific details-the desired values and limit values are determined.

In a subsequent evaluation step, a communication link in a critical state is determined in that the determined expected value of the data rate of the communication link and/or the expected value of the delay time of the communication link and/or the expected value of the data transmission rate of the communication link are compared with a limit value of the data rate of the communication link and/or a limit value of the delay time of the communication link and/or a limit value of the data transmission rate of the associated communication link. From a comparison of the expected values and the corresponding limit values on the different communication connections, it can be estimated which communication connections can be evaluated as being in a critical state. The evaluation is carried out by means of a numerical algorithm, which can be defined very simply but also very complex. The rule may be, for example, that an approach of the expected value of the data transmission rate to the limit value of the data transmission rate exceeds a certain share, for example exceeds 80%, resulting in the communication connection being evaluated as critical. The delay can be treated similarly, i.e. if the expected value of the delay of the communication connection exceeds a certain percentage, for example more than 80%, of the limit value of the delay of the corresponding communication connection, the communication connection is likewise evaluated as critical.

Finally, in a reconstruction step, the predefined simulation network is reconstructed in such a way that communication connections in critical states are reduced. Reconfiguring a simulation network means, for example, setting up new data connections between communicating network nodes, removing existing data connections if necessary, distributing the functions of the network nodes differently, transferring simulation applications to other network nodes or dividing and distributing simulation applications to different network nodes, offloading network nodes which aggregate multiple communication paths by implementing alternative communication paths by the simulation network, etc.

The previously described method for reconstructing a predefined distributed real-time simulation network can also be carried out iteratively in order to thus step up a better and more uniform load level of the simulation network.

With the execution of the described method, the resources of the real-time simulation network can be used as efficiently as possible, since over-dimensioning can be partially avoided and a uniform degree of loading of the simulation network is sought.

The reconstruction can be carried out completely automatically depending on the technical implementation of the simulation network. This can be achieved, for example, particularly simply when a full mesh is provided as an initial situation of the real-time simulation network in which virtually every network node is connected to any other network node via a data connection, which is at least technically generally reasonable in this respect. It may for example make sense to connect I/O network nodes to each other with data connections, whereas it may make sense to connect computing network nodes all to each other. The data connection in the real-time simulation network can then be abandoned within the scope of the execution of the method, so that the simulation network is relieved in terms of data connection. This reconfiguration may be irreversible, for example, when using fuse or anti-fuse techniques that may be found in a one-time programmable FPGA (field programmable gate array). The reconfiguration may also have reversible features, for example when using a multi-time programmable architecture, such as an FPGA that can be configured multiple times.

However, the reconstruction can also be carried out partially manually, the reconstruction information being generated completely by an automated evaluation step. The reconstruction can then be done partly manually, i.e. removing data connections, displacing simulation applications, displacing communication connections, etc. The specific instructions for what action should be taken are generated automatically by executing the computer-implemented method. In an embodiment of the computer-implemented method, it is provided, in particular, that, in order to reduce the delay on a communication connection in which one reconfiguration step is preceded by one or more communication network nodes, such as routers, between the computing network node and the I/O network node, the reconfiguration step is predefined by the method in such a way that, after the completion of the reconfiguration step, there are a reduced number of data connections of the communication network nodes (for example routers) or no data connections of the communication network nodes. After one of the two last-mentioned reconstruction steps, for example, a reduced or nonexistent transmission delay that can be allocated to one or more communication network nodes can be ascertained during the data transmission from the I/O network node to its respective computing network node. In a preferred embodiment of the method, provision is made for the topology of the simulation network to be determined by invoking information services implemented on the nodes of the simulation network, which information services, when invoked, return information about which nodes they are directly connected to, in particular via which data connections they are directly connected to. These information services may be very basic and return corresponding connection information to the location where the call was made. The topology of the real-time simulation network can be unambiguously determined without difficulty when each network node knows with which other network node it is directly connected. Alternatively, provision is made for the topology of the simulation network to be determined by reading a file with topology information of a predetermined simulation network.

In an advantageous further development of the method, provision is made in the reconfiguration step for the communication connections in the predetermined simulation network to be reduced in a critical manner in such a way that the network nodes are functionally expanded and/or functionally reduced and/or the data connections are functionally expanded and/or functionally reduced and/or the communication connections are routed differently between the network nodes in the predetermined simulation network at least in part while maintaining the topology of the simulation network. The functional expansion or reduction of the network nodes can be achieved, for example, by dividing the simulation application differently between the network nodes. Another possibility to reduce the functionality of the network node may be, for example, to reduce the data rate at which information should be transmitted. It is also possible, for example, to reduce the resolution of the values to be transmitted, thus resulting in a reduction in the byte length of the information to be conveyed. Another possibility is, for example, lossless data compression, wherein it is to be taken into account here that this may lead to greater delays of the network nodes. An expansion or reduction of the functionality of the data connection can be achieved by using other transmission media with different channel capacities.

Alternatively or additionally, communication connections in critical states in the predetermined simulation network can be reduced by the predetermined simulation network expanding at least one additional network node and expanding at least one additional data connection and/or reducing at least one existing network node and reducing at least one existing data connection at least partially in the event of a simulation network change.

A preferred embodiment of the method is characterized in that the expected value of the data transmission rate of the data connection is obtained by summing the data rates of the connected network nodes. The data rate of the connected network node can be obtained in different ways, as will be explained below.

In a further preferred embodiment of the method, it is provided that the channel capacity of the data connection is used as a limit value for the data transmission rate of the data connection, i.e. the maximum data transmission rate that can be achieved by the data connection.

Three variants for designing the method described so far are shown in the following, which variants differ in the way in which the expected value of the node data rate and/or the expected value of the node delay of the network node and/or in the way in which the expected value of the data transmission rate of the data connection is determined.

According to a first variant, it is provided that the expected value of the node data rate and/or the expected value of the node delay of a network node of the simulation network are determined on the basis of the hardware specification of the network node of the simulation network, in particular without taking into account the simulation application of the respective network node, in particular without taking into account a possible hardware parameterization of the relevant network node. For this purpose, the expected value can only be estimated relatively roughly. In particular, it is not taken into account how many different tasks and with which sampling times are used to calculate which simulation applications, with which frequencies data are acquired and/or with which channels data are transmitted and how, for example, the I/O network nodes (sampling rate, resolution) are parameterized. This variant of determining the expected value is therefore suitable for determining the expected value of the node data rate and/or the expected value of the node delay of a network node of the simulated network in the worst case in such a way that the maximum value of the node data rate and/or the node delay is used for the expected value of the node data rate and/or the expected value of the node delay. In any case, this will allow one to see if the design of the real-time simulation network is functional even under conditions that are generally considered to be highly unfavorable.

In a further embodiment of the method described above, it is provided that an average value of the actual node data rates and/or an average value of the node delays of the network nodes of the simulation network is determined from a plurality of known real-time simulation networks configured to be functional and that these average values are selected as expected values of the node data rates and/or expected values of the node delays of the network nodes of the simulation network. In this design of the first variant for determining the expected value, application-specific information, i.e. information relating to the operation of the particular predefined real-time simulation network, is also not used, but only hardware information about the simulation network is used. But using empirical values from other simulated networks having the same network nodes. This requires the presence of corresponding information about the real-time simulation network configured to be operational. In one embodiment of the computer-implemented method, for example, it is provided that the above-mentioned empirical values of the application-specific call rates including aperiodic tasks and/or of the application-specific processing durations including aperiodic tasks are used to determine the expected value/values.

According to a second variant for determining expected values of the node data rate and/or expected values of the node delay of the network nodes of the simulation network, these expected values are determined taking into account the simulation application of the respective network node, in particular taking into account a possible hardware parameterization of the network nodes. In order to carry out such a method, it is not necessary to also implement the simulation application, but it is only important to know the relevant parameters of the simulation application and/or of the hardware parameterization.

In an advantageous embodiment of the method, it is therefore provided that the expected value of the node data rate and/or the expected value of the node delay of the network node is determined taking into account the calculation step size of the periodic task and/or the assumed call rate and processing duration of the aperiodic task, the size of the calculated and transmitted data packets in the task, the configuration of the I/O function, in particular the call rate and the size of the processed I/O data packets.

A periodic task is a function implemented within a simulation application at specific constant time intervals, computing step sizes. This is necessary, for example, when numerical methods for solving equations, even differential equations, are implemented in the context of such periodic tasks. Such calculation steps are typically in the millisecond range, but in demanding, very dynamic tasks, the calculation steps may also be in the microsecond range. While periodic tasks are run through a fixed time grid, aperiodic tasks are triggered by several effects that are not within the foreseeable time grid, but are triggered, for example, by other effects. In practical implementations, the task of associating a function within the simulation application with an interrupt and then just performing it event-driven or aperiodically as needed. Since it is not easily foreseen how frequently such aperiodic tasks are called, in this design of the method it works with the assumed calling rate.

It is also necessary to take into account how large data packets are to be transmitted when carrying out periodic or aperiodic tasks, since this is also important for the data transmission rate which is then to be generated, which is caused by the network node and can therefore be transmitted via the data connection and therefore also via the communication connection. Network nodes implemented as I/O network nodes do not usually execute freely programmable simulation applications, but rather execute I/O functions, such as analog/digital conversion, which are determined by the hardware capacity, but can also be parameterized within certain limits. I/O network nodes can often be predetermined, for example by parameterization, at which rate they detect and/or output data and at which digital resolution they execute, which affects the size of the processed I/O data packets.

The second variant for determining the expected value is in principle much more precise than the pure processing of the hardware specification with the network node and the data connection. However, no consideration is given here, and also possible delays which are usually present in real simulation networks may not be taken into account. For example, collisions between data transmissions of different network nodes using the same data connection may occur. This may result in that data packets intended to be transmitted by the network node must be retransmitted with a certain time delay, which is not taken into account in purely theoretical observations, for example, according to the second variant of determining the expected value.

In order to also take these effects into account, according to a third variant for determining an expected value of the node data rate of a network node and/or an expected value of the node delay and/or an expected value of the data transmission rate of a data connection of the simulation network, it is provided that these expected values are determined by means of measurements in the simulation network. In a preferred embodiment, the expected value of the communication connection data and/or the expected value of the communication connection delay are determined by measurements during the operation of the simulation network, for which purpose a simulation application needs to be executed on the respective network node during the measurements.

According to a preferred embodiment, provision is made for the network nodes of the simulation network to be set with synchronized clock times for measuring the expected value of the delay of the communication connection. Each transmitting network node transmits data with a transmission time stamp, i.e. a data packet with a transmission time stamp, which is transmitted by the network node. The last receiving network node in the communication connection can then determine the expected value of the communication connection delay of the respective communication connection from the reception time of the received data or data packets and by evaluating the transmission time stamp. Since this occurs when the simulation application is implemented on the network node, additional time delays are also taken into account during the actual operation of the simulation network, for example by collisions on the communication link.

Alternatively, the expected value of the delay of the communication connection can also be measured in such a way that a synchronized clock time is set for all network nodes of the simulation network and an echo function is implemented in the network nodes of the simulation network. This echo function is also known per se in the field of data transmission with the concept "Ping". The expected value of the communication connection delay is then measured in such a way that the receiving or transmitting network node of the communication connection transmits an echo request to the network node of the communication connection which is transmitting or receiving accordingly in the operation of the simulation network, and the network node which transmits the echo request determines the echo cycle time after receiving the echo signal and thus the expected value of the communication connection delay. Such a measurement can be performed in particular when the simulation network is not running during the execution of the echo method, i.e. when no simulation application is implemented on the network node. Alternatively, it is also possible to run the simulation network during the execution of the echo method, i.e. to implement a simulation application on the network node.

In a preferred embodiment of the method and in all variants of the method described above, it is provided that the method is implemented on a computer, which is connected to the simulation network via a data connection, or that the method is implemented on a network node of the simulation network, which is designed as a computing node.

The object set forth at the outset is likewise achieved by a computer program product comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method according to the preceding description. The object is also achieved by a computer-readable storage medium, which contains instructions that, when the program is executed by a computer, cause the computer to carry out the method described above. When reference is made here to a computer, this may be, as already explained above, a computer (Rechner) connected to the simulation network via a data connection, or may also be a network node of the simulation network which is designed as a computing node.

Drawings

In detail, there are now a number of possibilities for designing and extending the design of the method according to the invention. Reference is hereby made, on the one hand, to those claims which follow claim 1 and, on the other hand, to the following description of the embodiments in conjunction with the drawing. In the figure:

FIG. 1 is a schematic diagram of an HIL simulator with which a real-time simulation network is implemented, wherein a control device is connected to the HIL simulator;

FIG. 2 schematically shows a simulation network with different network nodes and with a plurality of data connections;

FIG. 3 is the simulation network of FIG. 2, which is examined and reconstructed according to a first variant of the method for reconstructing the simulation network;

FIG. 4 is a table diagram of a first variant of a method for reconstructing a simulation network according to FIG. 3;

FIG. 5 shows the simulation network according to FIG. 2, which is examined and reconstructed according to a second variant of the method for reconstructing the simulation network;

FIG. 6 is a table diagram of a second variant of the method for reconstructing the simulation network according to FIG. 3;

FIG. 7 shows the simulation network according to FIG. 2, which is examined and reconstructed according to a third variant of the method for reconstructing the simulation network; and is

Fig. 8 is a table diagram of a third variant of the method for reconstructing the simulation network according to fig. 7.

Detailed Description

The figure generally illustrates a computer-implemented method 1 for reconstructing a U-predetermined distributed real-time simulation network 2, which is in the following also referred to simply as "simulation network", which is generally used for control device development.

Fig. 1 and 2 show typical situations known from the prior art when developing a control device. In fig. 1, a hardware-in-the-loop simulator 3 is shown, with which a predetermined distributed real-time simulation network 2 is implemented. The network node 4 of the emulation network 2 is here also present as a plug-in card for the HIL emulator 3. The network nodes 4 are at least partly interconnected by a plurality of data connections DV. For this purpose, each network node 4 has at least one data connection interface for connecting a data connection DV. The network node 4 is at least partially in the communication link KV via a data connection DV. A communication connection is thus a path within the emulated network 2 through which two network nodes actually exchange information. While the simulation network 2 is running, a simulation application 5 is implemented on at least one of the network nodes 4. In the exemplary embodiment shown, a technical mathematical vehicle model is used here.

The network nodes 4 have different functions. There is a network node 4 designed to compute a network node RK. These computing network nodes are small computers on which the real-time operating system runs. The other network nodes 4 implement a communication network node R. These communication network nodes are used, for example, for the serialization of parallel data streams. The other network nodes 4 are in turn designed as I/O network nodes IO. With the I/O network node, measurement data can be received from an external physical process or signals influencing an external physical process can be output by the I/O network node. The external physical process is given in the exemplary embodiment shown in fig. 1 by two control devices 8. The control device 8 is connected to the simulation network 2 in terms of signaling, which is only briefly illustrated in fig. 1. While the simulation network 2 is operating, the simulation network 2 interacts with the connected control devices 8. The control devices 8 check their functional capability by means of the HIL simulator 3. The exemplary embodiment according to fig. 1 is merely exemplary, and completely different device-technical configurations are also possible.

In the figures, the different components of the simulation network 2 shown are labeled with the common abbreviations (e.g. DV, RK, R, IO, etc.) and the following numbers (e.g. DV1, RK2, R1, IO3), respectively, already mentioned above. In the following, when the exact elements of the emulated network 2 are not important, the identifiers of the different network nodes 4 of the emulated network 2 are used in part in a generic way. If a particular element is referred to, for example a particular computing node in the illustrated simulation network, then for example the computing node RK2 is referred to and not just any computing node RK.

The simulation network 2 is shown more accurately in fig. 2, although the illustration is also only schematic. The simulation network 2 has two computing network nodes RK1, RK2, two communication network nodes R1, R2 in the form of routers and five I/O network nodes IO1, IO2, IO3, IO4, IO 5. The network nodes 4 are generally connected to one another in a specific manner via a plurality of data connections DV, namely data connections DV0, DV 1. Not every network node 4 is connected to any other network node 4, so that only selected data connections DV exist between the network nodes 4.

The network nodes 4 communicate with one another via a specific communication link KV depending on the particular application carried out with the simulation network 2, of which only two communication links KV1, KV2 are shown here for the sake of clarity. In the embodiment shown in fig. 2, the I/O network node IO1 sends measurement data to the computing network node RK1 via the communication network node R1. The I/O network node IO4 likewise transmits the measurement data via the communication network node R2 to the computing network node RK 2. The communication connections KV are thus communication connections which are actually formed between the network nodes 4 which communicate with one another, and these communication connections can thus utilize a plurality of data connections DV.

In the design and implementation of such a simulation network 2, the structure of the simulation network 2 must be chosen with great care so that the simulation network 2 does not reach its functional limits during operation. This may be the case, for example, when the data connection DV is loaded by the network node 4 in aggregate at a high data transmission rate, so that the data connection reaches its channel capacity. Overload may also occur when the delay in transmitting data packets between the network nodes 4, the so-called delay, is greater than expected and also required.

The computer-implemented method 1 for reconstructing U-subscription distributed real-time simulation networks 2 shown in fig. 3 to 8 enables systematic reconstruction of subscription real-time simulation networks 2, so that communication connections in critical states are automatically identified and at least reduced or even completely eliminated.

Fig. 3, 5 and 7 each show a simulation network 2 having identical components, i.e. identical network nodes 4 and identical data connections DV between the network nodes 4. The computing network nodes RK1, RK2 are each provided for implementing the simulation application 5. The simulation applications 5 on the computing network nodes RK1, RK2 are functionally different from one another and together implement the overall application of the simulation network 2.

The following method steps are common in the three different variants of the method 1 for reconstructing the real-time simulation network 2 shown in each case by U, which variants are shown in fig. 3 and 4, fig. 5 and 6, and fig. 7 and 8. The topology of the simulation network 2 is first detected in each case, so that topology information exists about the network nodes 4 and the data connections DV between the network nodes 4. This is shown in the tables according to fig. 4, 6 and 8, respectively, in which the network node 4 is completely detected in each case. The information of which network node 4 is connected to which data connection DV is not shown in detail here for the sake of clarity. At least computing network nodes RK1, RK2, communication network nodes R1, R2 and I/O network nodes IO 1. All data connections DV0 to DV8 were also detected separately.

And then determining an expected value E-KDR of the node data rate and an expected value E-KL of the node delay for the network nodes RK, R and IO of the simulation network 2. The node data rate is here a data rate that can be generated, received or relayed by the respective network node RK, R, IO. The node delays relate to time delays with which the network nodes RK, R, IO delay the sending out, receiving and forwarding of data packets. Furthermore, an expected value E-DVDR of the data transmission rate is determined for the data connection DV. The expected values of the aforementioned parameters are listed in the second, third and fourth columns of the table shown, respectively.

Furthermore, the communication connections KV, i.e. the data paths, are determined between the network nodes of the simulation network, via which the different network nodes 4 communicate end to end. The communication link KV is illustrated in fig. 3, 5 and 7 by dashed lines. The respective elements participating in the simulation network 2 are not listed individually in the table again, but the communication connections KV are each listed in a column of the respective table.

For the communication link KV, the expected value E-KVDR of the communication link data rate and/or the expected value E-KVL of the communication link delay is determined on the basis of the expected value E-KDR of the node data rate of the network nodes 4 participating in the respective communication link KV and the data link DV and/or on the basis of the expected value E-KL of the node delay and/or the expected value E-DVDR of the data transmission rate. These determined expected values are listed in the corresponding columns of the communication link KV.

For the communication link KV, a limit value G-KVDR for the data rate of the communication link and/or a limit value G-KVL for the delay of the communication link are then determined, and for the data link DV, a limit value G-DVDR for the data transmission rate is determined. These limit values may be predetermined, for example, by an engineer participating in the design of the simulation network. But this may also relate to predetermined values from different predetermined design patterns. In a safe design mode, for example, a smaller limit value can be predefined, whereas in a design mode in which the resources are utilized as completely as possible, a larger limit value is also acceptable.

In particular, it is not worth noting here how the different desired values and the different limit values are determined or selected in detail, it being important here first of all to determine or select or to predetermine these values or the selection of these values as a whole.

In an evaluation step BS, the communication link is then determined in a critical state in such a way that the determined expected value E-KVDR of the communication link data rate and/or the expected value E-KVL of the communication link delay and/or the expected value E-DVDR of the data transmission rate is compared with the limit value G-KVDR of the communication link data rate and/or the limit value G-KVL of the communication link delay and/or the limit value G-KVL of the associated communication link KV or the limit value G-DVDR of the data transmission rate of the associated or involved data link DV. The limit values G-KVDR for the data rate of the communication link and the limit values G-KVL for the delay of the communication link are also registered in the table in the column for the respective communication link KV.

The result of the evaluation step BS is symbolically marked in the table either by hooking (when the evaluation step indicates that the communication connection KV being checked is not in a critical state, i.e. does not exceed the corresponding limit value for the expected value) or by a lightning symbol (when the communication connection KV being checked is verified to be in a critical state, i.e. the expected value exceeds the corresponding limit value). If a communication link KV in the critical state is identified, the components of the simulation network 2 participating in the communication link in the critical state can also be found, which is shown in the column with BS in the corresponding table.

The three method variants shown have in common that, in the reconstruction step U, the predetermined simulation network 2 is reconstructed in such a way that the communication links KV in the critical state are reduced and, ideally, are completely eliminated. The corresponding reconstruction measures are noted in the table in the columns written with U or in the columns written with U1 and U2.

Having explained the common features of the different variants of the method 1 for reconstructing the U-predetermined distributed real-time simulation network 2, only the features of the characterizing part of the variants of the method 1 are explained next for the individual method variants.

The method 1 shown in fig. 3 and 4 is characterized in that the expected value E-KDR of the node data rate and the expected value E-KL of the node delay are determined for the network nodes RK, R, IO of the simulation network 2 on the basis of the hardware specifications of the network nodes RK, R, IO of the simulation network 2, in particular without taking into account the simulation application 5 which calculates the network nodes RK1, RK2 and without taking into account further hardware parameterizations of the network nodes RK, R, IO. Since the simulation application 5, which calculates the network nodes RK1, RK2, and the further hardware parameterization of the network nodes RK, R, IO are not taken into account, it is not known which communication connections KV exist between the network nodes RK, R, IO, and therefore a possible, reasonable communication connection KV is assumed. The communication connection between the I/O network nodes IO is not considered reasonable here. Thus, 9 communication connections KV1, KV9 are produced, which are shown in fig. 3 by dashed lines.

Since no (information) about the simulation applications 5 of the network nodes RK, R, IO and possible hardware parameterization can be used, the maximum values of the node data rate and of the node delay are used here to determine the respective expected values. In the communication link KV, the expected value E-DVDR of the data transmission rate of the data connection DV is obtained by summing the data transmission rates of the connected network nodes. The I/O network nodes IO1, IO2 and IO3 are pooled in the communication network node R1. In the most unfavorable case, all I/O network nodes IO1, IO2 and IO3 communicate with the computing network node RK1 via the communication network node R1. A data rate expected value of 1.4Gbps ("bps" which is then always referred to as "bits/second") is thus generated for the data connection DV1 as the sum of the expected node data rates of the I/O network nodes IO1, IO2 and IO 3. As already explained, the expected values E-KDR of the node data rates of the I/O network nodes IO1, IO2 and IO3 are maximum values. The expected value E-KVDR, which is the data rate of the communication link KV2, therefore likewise yields a maximum value of 1.4 Gbps.

The minimum channel capacity of the participating data connections is used as the limit value G-KVDR for the data rate of the communication connection KV 2. Since all participating data connections have a channel capacity of 1.2Gbps, this is also the limit value for the data transmission rate of the communication connection KV 2. The limit values G DVDR for the data transmission rates of the data connections DV0 to DV8 are selected accordingly, which therefore correspond to the respective channel capacities.

In the evaluation step BS, it is required that the expected value E-KVDR of the data rate of the communication connection should not exceed the limit value G-KVDR of the data rate of the communication connection. Communication connection KV2 does not satisfy this condition. In addition, further communication connections which are omitted in the table list for the sake of clarity do not satisfy this condition either. The reason for the classification of communication link KV2 as critical classification is data link DV1, as can be seen from a comparison of the expected value E-DVDR of the data transmission rate of the participating data link with the limit value G-DVDR of the data transmission rate of the participating data link. The reconstruction step U1 for eliminating the communication link KV2 in the critical state consists in functionally expanding the data link DV1 in such a way that it is replaced by a data link with a channel capacity of 2.5Gbps, which is twice as large as the channel capacity, and is not explained in detail here.

As a second reconstruction measure U2 suggests, the functionality of the I/O network nodes IO2 and IO3 is reduced, i.e. by limiting the maximum allowed data rate to 400Mbps each. This measure also does not require a change of the topology of the simulation network 2 shown in fig. 3.

Fig. 5 and 6 show a second variant of a method 1 for reconstructing a predefined distributed real-time simulation network 2. This second variant of the method 1 is characterized in that the expected value E-KDR of the node data rate and/or the expected value E-KL of the node delay of the network nodes RK, R, IO of the simulation network 2 are determined taking into account the simulation application 5 of the respective network node and also taking into account the possible hardware parameterization of the network nodes RK, R, IO. The consideration of this information makes it possible to estimate the different expected values much more accurately, since there is now a basis for the rate of generation of data and thus for the total amount of data on the data connection DV. It is therefore also known who the sender and receiver of the data packet are, so that the communication link KV can also be determined precisely.

The simulation network according to fig. 5 is identical in structure to the simulation network according to fig. 3, but the data in the tabular overview according to fig. 4 and in the tabular overview according to fig. 6 are very different due to the different processing. Some of them are defined differently, and therefore, the examples cannot be compared simply. Embodiments can therefore be viewed separately. The topology information 6 is substantially similar to the previously described embodiments. But there are much fewer communication connections KV, i.e. in total only 6 communication connections KV1 to KV 6. In the tabular overview, the communication link KV4 is not shown for spatial reasons.

In particular, when the simulation application 5 is considered, it is provided that the expected value E-KDR of the node data rate and/or the expected value E-KL of the node delay are determined taking into account the calculation step TS of the periodic task and the assumed call rate and processing duration of the aperiodic task. The size of the calculated and transmitted data packets, the configuration of the I/O functions of the I/O network nodes IO and the scheduling and size of the processed I/O data packets in the tasks of the simulation application 5 are also taken into account. The result is that it can be ascertained that a much smaller expected value of the node data rate E-KDR and also a much smaller expected value of the data connection data rate E-DVDR result, since the relevant network nodes RK, R, IO are not actually fully utilized in their function. The expected value E-DVDR of the data transmission rate of the data connection DV is likewise determined, as explained with reference to fig. 3 and 4, i.e. by observing the expected value E-KDR of the node data rate, at which the participating network node 4 feeds into the participating data connection DV. The expected value E-KVDR of the communication connection delay is also determined by summing the delays of the participating network nodes 4.

Based on the data thus determined, the communication connection KV3 in the critical state is determined. The expected value E-KVDR of the data rate of the communication link is slightly greater than the corresponding limit value G-KVDR of the data rate of the communication link. As a solution provided here, which is also implemented automatically if necessary, the node data rates of the I/O network nodes IO2 and IO3 are each reduced by 25%, in such a way that the analog/digital conversion performed by the I/O network nodes is set to a lower resolution, i.e. from 16Bit to 12Bit (see corresponding entries in column U1). This measure does not cause structural changes in the topology of the simulation network 2.

Other proposals for the reconfiguration U2 are that the I/O network node IO3 is not connected via the communication network node R1 to the computing network node RK2, but rather via an assigned data connection DV6, which data connection DV6 connects the I/O network node IO3 to the communication network node R2. The additional data transmission rate thus generated at the data connection DV3 does not lead to an inadmissible utilization of the data connection DV3, and the corresponding limit value G-KVDR of the data rate of the communication connection is then not yet reached.

Fig. 7 and 8 show a further third variant of a computer-implemented method 1 for reconstructing a U-subscription distributed real-time simulation network 2. The simulation network 2 shown in fig. 7 corresponds virtually completely to the simulation network 2 shown in fig. 5, including the simulation application 5 implemented on the computing network nodes RK1, RK 2. However, a computer 7 is additionally shown, on which the method 1 is implemented by means of programming techniques. The computer 7 is connected to the emulation network 2 via a data connection DVR. In the simulation network 2 shown in fig. 3 and 5, the method 1 can also be implemented on a computer 7 as shown in fig. 7. The computer 7 is not shown in the other figures in order not to unnecessarily complicate the illustration.

The variant of the method 1 discussed in fig. 7 and 8 is characterized in that the expected values of the node data rates E-KDR and the expected values of the node delays E-KL of the network nodes RK, R, IO are determined by measurements in the simulation network 2. The expected value E-DVDR of the data transmission rate of the data connection is determined by calculation from the expected value described earlier (as already explained in the foregoing), but it is therefore also based on the performed measurement.

In the present case, the measurements are performed while the simulation network 2 is running. This has the advantage that effects such as data collisions or data collisions on the data connection DV or the communication connection KV are also taken into account, which generally lead to a delay in the data transmission and also to a higher data throughput, since the data packets may have to be transmitted several times. A particularly practical estimate of the full utilization of the simulation network 2 is thus obtained.

The expected value E-KVDR of the data rate of the communication connection and the expected value E-KVL of the delay of the communication connection are also determined by measurements, currently by measurements during operation of the simulation network 2. The measurement of the expected value of the communication link is shown in the table of fig. 8, since the expected value for the communication link KV is not always summed up from the expected values of the nodes involved.

For the reasons already mentioned, the expected value E-KVL of the communication connection delay is measured during the operation of the simulation network 2, i.e. when the simulation application is implemented on the computing network nodes RK1, RK2, in such a way that a synchronous time is set for all network nodes RK, R, IO of the simulation network 2, each transmitting network node RK, R, IO sets a transmission time stamp for data or data packets transmitted by said network node and the last receiving network node RK, R receives a time from its determined reception time and calculates the expected value E-KVL of the communication connection delay of the respective communication connection KV2 by evaluating said transmission time stamps. In the present case, the deviation between the expected value determined in accordance with fig. 5 and the measured expected value measured in accordance with fig. 7 is small, so that it is determined that the same critical communication link KV can also be eliminated by the same measures.

In order to measure the expected value E-KVDR of the data rate of the communication connection, the communication network nodes R1, R2 are each equipped with an observation application which statistically detects the data traffic and then sends corresponding information for evaluation to the computer 7, where the corresponding expected value is determined by evaluating the statistical data. An advantage of this solution is that no additional infrastructure is required, but it is to be taken into account that embodiments of the viewing application may have an impact, albeit a minor impact, on the real-time data transmission depending on the implementation.

List of reference numerals

1 method

2 real-time simulation network

3 HIL simulator

4 network node

5 simulation application

6 topology information

7 computer

8 control device

RK computing network node

R communication network node, router

IO I/O network node

Data connection of DV between network nodes

Data connection of DVR between computer and emulation network

KV communication connection

Expected value of E-KDR node data rate

Expected value of E-KL node delay

Expected value of data transmission rate of E-DVDR data connection

Limit value for the data transmission rate of a G-DVDR data connection

Expected value of data transmission rate of E-KVDR communication connection

Limit value for data transmission rate of G-KVDR communication connection

Expected value of E-KVL communication connection delay

Limit value of G-KVL communication connection delay

Calculating step length of TS periodic task