CALCULATION SYSTEM AND METHOD FOR OPERATION OF A CALCULATION SYSTEM

Description

[0001] The invention relates to a computing system comprising at least one computing unit of at least one motor vehicle and a cloud comprising at least one computing unit. The invention also relates to a method for operating such a computing system.

[0002] Modern vehicles incorporate an increasing number of functions, some of which are complex to implement, such as advanced driver assistance systems (ADAS) and / or in-vehicle infotainment systems (IVI). These and other modern functions implemented in vehicles require high computing power and / or a large amount of input data to solve the complex and / or time-consuming calculations. Solutions for this typically utilize external server facilities connected via a wireless communication link.These server facilities can be used to retrieve information, such as required, up-to-date input data. They can also provide at least some of the computing power needed to perform a function implemented by an application. The server facility can also be a cloud or a part of a cloud.

[0003] An example of computationally intensive functions within a motor vehicle is the calculation of routes, particularly for determining travel times, to points of interest (POIs) within a navigation system. Here, users, i.e., drivers of the motor vehicle, often want to see immediately how long it will take them to reach the respective POI. However, this calculation requires route planning using a multitude of input data, which, in addition to digital map data, includes, for example, current traffic information and the like. Especially when the motor vehicle is to be guided automatically to a POI, aspects of trajectory planning can also be involved. If input data and / or applications are used for this purpose, the calculation becomes complex.If subprocesses of applications on a server are used, the corresponding function in the vehicle is only available if the communication connection to the server also exists, as otherwise only the information available on board the vehicle could be used.

[0004] DE 101 05 898 A1 relates to a method for transmitting route guidance elements from a vehicle navigation device to a central unit. The vehicle navigation device is said to lack sufficient resources for route calculation at the central unit, so communication with the central unit, where the routes are calculated, is established via a communication device. DE 10 2012 107 886 A1 relates to a method for the electronic recognition of traffic signs, in which data of a recognized basic pattern is wirelessly transmitted via a communication device in the vehicle to an external server, where pattern recognition is performed by the external server using a pattern recognition algorithm. As a result of the pattern recognition, the data of the recognized traffic sign is transmitted to the vehicle's communication unit and, in particular, a corresponding traffic sign is displayed on a display unit.

[0005] Documents DE 10 2012 008978, US 2013 / 304863, US 2016 / 091328 and DE 10 2015 200422 disclose similar systems and procedures.

[0006] In addition to the problem of functions no longer being available when partially outsourcing from the vehicle, there is also the problem that if the application or sub-processes are implemented separately within the vehicle or on the server system, there is a limitation of flexibility, since corresponding software and / or hardware resources have to be developed specifically for the vehicle and / or server.

[0007] The invention is therefore based on the objective of providing a calculation system that offers more flexibility and, in particular, improves the availability of functions within the motor vehicle.

[0008] To solve this problem, a calculation system according to claim 1 is required, wherein at least one application calculating output data from input data is executable at least partially through the vehicle-side computing device and at least partially in the cloud.

[0009] According to the invention, a control unit of the computing system is used, which is preferably at least partially located on the vehicle side, to dynamically decide to what extent the application should take place in the vehicle itself or from the cloud communicating with the vehicle. It is particularly preferred if both the vehicle-side computing unit and the cloud, or at least one computing unit of the cloud, can each fully execute the application if circumstances so require, as will be discussed in more detail below.

[0010] According to the invention, it is therefore proposed to provide distributed application computing between a motor vehicle and a cloud infrastructure, wherein preferably both the vehicle-side computing unit and the at least one cloud-side computing unit are also configured for the complete execution of the application. The computing system thus represents a distributed system in which computing components provided on the communicating computing units communicate with each other and coordinate their computational actions by exchanging messages. The computing components interact with each other to pursue a specific goal, in this case, determining the application's output data.The application is therefore implemented as a so-called distributed computer program, whereby the application can be divided into various subprocesses that can be assigned to different computing units, particularly depending on at least one assignment criterion that is checked in the control unit. The control unit can be implemented via a software platform that provides the flexibility to run the application, or a part of the application, either on the vehicle's computing unit or in the cloud. This, in turn, allows the computing load to be shifted between the vehicle's computing unit and the cloud, thus providing the basis for flexible, especially vehicle-related, functions, such as driver assistance functions and / or infotainment functions, which were not specifically developed for the vehicle's computing unit or the cloud.In this context, the control unit may include software platform means that ensure the compatibility and context availability of the application components.

[0011] This creates an architecture that allows functions, particularly those used by motor vehicles and represented by an application, to be dynamically and flexibly distributed between the vehicle and a cloud communicating with the vehicle. Advantageously, applications, including their context and subprocesses, can be distributed in such a way that the greatest benefit can be derived from the available computing power and input data on both sides. The approach described here, according to the invention, can therefore be advantageously used to shift or outsource computing load from the vehicle to the cloud and / or from the cloud to vehicles, offering the flexibility to execute at least one application either in the cloud or the vehicle, or to choose a distributed execution.

[0012] Particularly with regard to enabling full executability both on the vehicle side and in the cloud, it is advantageous to provide sufficient input data at both locations in the same way, specifically using the same interface. Thus, an advantageous embodiment of the present invention provides that at least one database, accessible by the application and providing input data, is provided on both the vehicle and cloud sides. This database has the same structure for both the vehicle and the cloud. Specifically, the databases have identical interfaces and / or the same data formats / structures. The contents of the databases may, however, differ.In many cases, the cloud-based database will contain a larger amount of information than the vehicle-side database and / or more up-to-date information. Due to the identical structure and / or compatibility of the databases, the distributed computing approach allows for the flexibility of running vehicle-side applications with a larger or more up-to-date dataset in the cloud. In addition to the availability of a larger and / or more up-to-date input dataset, it remains true that computational load can still be shifted from the vehicle-side computing system to the cloud.

[0013] As already mentioned, the computing system expediently includes a control unit for distributing subprocesses of the application to the computing devices, wherein the vehicle-side computing device includes at least one component of the control unit and / or at least one client unit that communicates with the control unit. Preferably, in a vehicle-centric approach, where the main focus is on implementing functions used during the operation of the vehicle and / or supporting its user, the control unit can be implemented entirely within the vehicle, at least with regard to the workload distribution of a vehicle-side computing task. This allows, for example, the application to still run even without a communication connection.

[0014] As already explained, the subprocesses are assigned based on at least one allocation criterion. Specifically, the subprocesses are distributed depending on the utilization of the respective computing units. According to the invention, a dynamic distribution will occur depending on the utilization of the respective computing units, so that when a large amount of free computing power is available in the vehicle, a larger proportion can be performed on the vehicle's computing unit than when the vehicle is heavily utilized. Conversely, it is of course also conceivable to offload at least part of a server-initiated application to the vehicle when the vehicle is underutilized, for example, when parked. Another allocation criterion ,The data quality of the input data, which is not part of the invention, can be considered. This means that if the cloud provides more up-to-date, more numerous, and / or more accurate input data, the cloud can be used as the execution location for the corresponding subprocess that requires this input data. Also not part of the invention is the availability and / or quality of the communication link between the cloud and the vehicle. This would allow, for example, subprocesses requiring the transfer of a large amount of data to be preferably carried out by the initiator of the corresponding instance of the application if the communication link is of low quality.This can also generally apply to large amounts of data generated by the vehicle or the cloud. For example, it might be planned that a subprocess evaluating sensor data from the vehicle (i.e., data recorded by the vehicle's sensors) should be preferentially assigned to the vehicle's own computing unit. Unprocessed sensor data can often represent a large volume of data, which can be reduced by processing it in a subprocess to lessen the load on the communication link, especially if the data quality is poor.

[0015] Regarding the communication link between the vehicle and the cloud, it may be provided that this includes a mobile network and / or a local wireless network and / or a communication line of a charging cable in a vehicle connected for charging. The communication link will therefore be wireless in most cases, although in exceptional cases wired communication links may also be available and thus usable, for example, when a charging cable is used to charge a battery installed in the vehicle, which is also designed for high-bandwidth data exchange and can therefore enable a particularly good communication link to the cloud.Regarding wireless communication links, the exchange of information between the cloud and the vehicle, in particular messages during distributed computing, can take place via a mobile network and / or a local wireless network, for example a WLAN network.

[0016] In addition to messages related to distributed computing, at least some of the software components implementing the applications can also be exchanged between the vehicle and the cloud. This makes it possible to perform updates and / or to avoid having to store the software components for implementing the application in both locations, but rather to provide them only when needed for distributed computing. This can potentially further reduce storage requirements.

[0017] In a particularly advantageous embodiment of the present invention, at least one of the at least one application can implement a vehicle function, in particular a navigation function, that is to be used during the operation of the vehicle and / or serves to support the driver of the vehicle. Especially for the complex, modern functions discussed at the outset, which are provided in motor vehicles, including in particular ADAS functions and / or IVI functions, the possibility of flexible, distributed computing of corresponding applications offers a multitude of advantages, particularly with regard to the implementation of more complex functions, but at the same time their improved availability, since computing solely within the vehicle is still possible. Thus, motor vehicles can be significantly improved in this way with regard to both comfort and operation.This becomes clear in the example of navigation functions, where, for instance, partial calculations can be outsourced to the cloud, where more accurate and / or up-to-date input data is available and / or greater computing power is available for faster calculations. Nevertheless, if the communication connection to the cloud is unavailable or of poor quality, the function in the vehicle does not necessarily have to be forgone, as calculations can then be performed within the vehicle itself, and the function is still fundamentally available.

[0018] In addition to calculation tasks (applications) initiated within the vehicle itself, calculation tasks can, of course, also be dynamically outsourced from the cloud to the at least one vehicle in the opposite direction. This is particularly advantageous when multiple vehicles are part of the calculation system. It can be especially beneficial to provide that at least one of the at least one application includes a development application which, particularly for a large number of vehicles, evaluates vehicle operating data as input, specifically including sensor data and / or vehicle status data. A cloud application can thus also be at least partially outsourced to at least one vehicle to perform local calculations, especially based on local input data, such as sensor data.In particular, the cloud application can be run on a variety of computing devices in different motor vehicles in order to utilize the available computing power of motor vehicles where it is not currently needed, and also to use a variety of different input data, for example when it comes to statistical evaluation of sensor data and / or other operating data of the motor vehicle, especially with regard to the development of new functions for motor vehicles, new vehicle systems and / or means that can be used in road traffic.

[0019] It should be noted that, within the scope of the present invention, it is also conceivable that the execution of at least one additional application is limited to the vehicle's computing system and / or the cloud. The distributed computing approach described here therefore does not preclude the possibility of providing separate applications on the vehicle and / or cloud side, for example, by choosing the conventional client-server approach for certain functions, if, for instance, an additional application could not be fully executed within a vehicle due to insufficient computing power, and so on.

[0020] In addition to the calculation system, the present invention also relates to a method for operating a calculation system, according to claim 8.

[0021] Further advantages and details of the present invention will become apparent from the following exemplary embodiments and the drawings. These show: Fig. 1 shows a calculation system according to the invention, and Fig. 2 shows a sketch of the method according to the invention.

[0022] Fig. 1 Figure 1 shows a schematic diagram of a computing system 1 according to the invention, which comprises at least one computing unit 2, in particular computing units 2 of different motor vehicles 3, and a cloud 4, which in turn comprises at least one, usually several, computing units 5. The motor vehicle 3, or the computing unit 2, and the cloud 4 can communicate via a communication link 6, wherein the communication link 6 can be configured differently, even over time, but in most cases will include at least a wireless segment using a mobile network and / or a local wireless network. If applicable, the communication link 6 can also be established via a charging cable for charging a battery of the motor vehicle 3, thus providing a particularly broadband connection.The computing system 1 also includes a control unit 7, which is only indicated here and can be implemented in a distributed manner, thus having control components in both the computing unit 2 and the computing unit 5.

[0023] The computing system 1, by means of the control unit 7, is now configured to execute applications that calculate output data from input data, distributed across the computing system 1, specifically at least partially on the vehicle-side computing unit 2 and within the cloud 4. Two advantageous features are present. Firstly, the application can be fully executed both in the vehicle 3 and in the cloud 4. This means that program resources for the application are available in both the vehicle 3 and the cloud 4, or at least applicable there, provided the program resources are also exchangeable via the communication link 6. For this purpose, the control unit 7, as part of the vehicle-side computing unit 2, can, for example, include software platform resources 8 that ensure compatibility.Ensure the executability of such software resources; corresponding software platform resources can, of course, also be provided by Cloud 4. The application was therefore not developed solely for Cloud 4 or solely for the vehicle 3, but in such a way that its software resources can be executed in both locations.

[0024] To promote full executability in both vehicle 3 and cloud 4, and in particular to allow the application to run in vehicle 3 even if the communication connection 6 is temporarily unavailable and / or of poor quality, databases 9 and 10 are provided for both vehicle 3 and cloud 4. These databases are structured identically, specifically having the same access interface, data formats, and data structures. It is possible that database 10 in cloud 4 may contain larger amounts of data and / or more accurate and / or up-to-date input data for the application than database 9. Advantageously, with an existing communication connection 6, regular synchronization can be performed, at least with regard to updates.

[0025] Assignment criteria, which determine where which subprocesses of an application to be executed are to be implemented, evaluate the current utilization of the vehicle's computer unit 2 and the cloud 4. Therefore, at least those components of the control unit 7 that are to assign vehicle-initiated applications or their subprocesses are independently functional on the vehicle 3 side. This enables, in particular, the vehicle-initiated applications to be executed solely on the vehicle side, as described above.

[0026] Fig. 2The possibilities offered by flexible distribution are further illustrated in the form of a corresponding diagram. An application 12 is to be executed, which can be divided into several subprocesses 13 and can therefore, in principle, be executed in a distributed manner. The most frequently used option for distributed computing is a division between the cloud 4 and the vehicle-side computing unit 2. This means that parts of the application 12, i.e., at least one subprocess 13, are implemented in the vehicle-side computing unit 2, while the remaining subprocesses 13 are executed in the cloud 4 using its computing units 5. This is symbolized by box 14.Boxes 15 and 16 nevertheless symbolize possible further extreme cases that may arise when evaluating the allocation criteria, namely in the case of box 15 the complete execution by the motor vehicle 3, in the case of box 16 the complete execution by the cloud 4.

[0027] As already explained, application 12 can be a vehicle-initiated application that relates to the operation of the vehicle 3 and / or the support of a user of the vehicle 3. For example, the application can implement a function of an ADAS and / or an IVI system. An example concerning a navigation system, namely route calculation, will be briefly explained in more detail. It is assumed that a navigation map is displayed within the vehicle, showing various points of interest (POIs). The navigation system should then offer the function of immediately informing the user about the spatial and temporal distance to the various POIs.The corresponding route-travel-time calculation, as application 12, can then be distributed between the vehicle 3 and the cloud 4, with the allocation criteria being evaluated according to the current operating state of the computing system 1. For example, the more accurate, up-to-date, and / or comprehensive database 10 makes calculations 13 from the cloud 4 advantageous if the corresponding data is required as input for the subprocess 13; the greater available computing power of the cloud 4 also supports this. However, if computing resources of the computing facility 2 are already available, the calculations of some subprocesses 13 are even based on input data originally generated in the vehicle 3, for example, regarding fuel consumption and / or an environmental model.

[0028] Since vehicle 3 also has a database 9, in extreme cases all subprocesses 13 can even be executed within vehicle 3, so that the corresponding function of the navigation system is available in any case. Input data generated by cloud 4 is transmitted back to vehicle 3 accordingly. The input data (route, lateral and spatial distance) for each POI is thus received and can be displayed within vehicle 3.

[0029] Cloud-initiated applications 12, such as development applications, can also be processed in a distributed manner. This is particularly useful when vehicles 3 of the computing system 1 can provide a significant amount of free computing power, for example, when parked, ideally when the vehicle 3 is plugged in to charge, as this ensures sufficient electrical energy. Another area of ​​application, especially for development applications, arises when operational data from vehicles 3 are to be evaluated anyway, such as sensor data and the like, particularly with regard to the development of new vehicle systems / functions and the like.Then, corresponding subprocesses can begin evaluating the operating data of the motor vehicle 3 directly on site in the vehicle-side computing device 2, and the results of a larger number of motor vehicles 3 can be combined in the cloud 4.

Claims

1. A calculation system (1) comprising at least one computing device (2) of at least one motor vehicle (3) and a cloud (4) comprising at least one computing device (5), wherein the computing system (1) comprises a control unit (7) for dynamically distributing subprocesses (13) of at least one application (12), which maps a function to be used on the vehicle side, to the computing devices (2, 5), so that the application, which calculates output data from input data (12) can be executed at least partially by the vehicle-side computing device (2) and at least partially by the cloud, wherein, in addition to messages relating to the distributed computation, software means implementing at least subprocesses (13) of the application (12) can also be exchanged between the motor vehicle (3) and the cloud (4), characterized in that the dynamic distribution takes place depending on the current utilization of the respective computing devices (2, 5), so that when a large amount of free computing power is currently available on the motor vehicle (3), a larger proportion is performed on the vehicle-side computing device (2) than when the vehicle-side computing device is heavily utilized (2)..

2. The calculation system (1) according to claim 1, characterized in that at least one database on the motor vehicle side and at least one database on the cloud side, each of which can be addressed by the application (12), supplies input data, and each of which has the same structure for the motor vehicle (3) and the cloud (4) that can be addressed by the application (12), supplies input data (9, 10), and has the same structure for the motor vehicle (3) and the cloud (4), wherein, in particular, the cloud-side database (10) contains a larger amount of information than the motor vehicle-side database (9).

3. The calculation system (1) according to claim 1 or 2, characterized in that the vehicle-side computing device (2) has at least one subcomponent of the control unit (7) and / or at least one client unit communicating with the control unit (7).

4. The calculation system (1), according to claim 3, characterized in that, in the subprocess (13) evaluating sensor data from the motor vehicle (3), this is preferably assigned to the motor vehicle-side computing device (2).

5. The calculation system (1) according to one of the preceding claims, characterized in that a communication connection (6) between the motor vehicle (3) and the cloud (4) comprises a mobile communications network and / or a local wireless network and / or a communication service of a charging cable in the case of a motor vehicle (3) connected for charging.

6. The calculation system (1) according to one of the preceding claims, characterized in that at least one of the at least one application (12) implements a motor vehicle function, in particular a navigation function, which is to be used during operation of the motor vehicle (3) and / or serves to assist a driver of the motor vehicle (3).

7. The calculation system (1) according to one of the preceding claims, characterized in that at least one of the at least one application (12) comprises a development application which, in particular, evaluates operating data of the motor vehicle (3), in particular comprising sensor data and / or status data of the motor vehicle (3), as input data for a plurality of motor vehicles (3).

8. A method for operating a computing system (1) comprising at least one computing device (2) of at least one motor vehicle (3) and a cloud (4) comprising at least one computing device (5), wherein, in the case of an application (12) to be executed which calculates output data from input data and maps a function to be used on the motor vehicle side, a control unit (7) of the computing system (1) dynamically distributes subprocesses (13) of the application (12) and executes them at least partially on the cloud side, wherein, in addition to messages relating to the distributed calculation, software means implementing at least subprocesses (13) of the application (12) are also exchanged between the motor vehicle (3) and the cloud (4), characterized in that the dynamic distribution takes place depending on the current utilization on the part of the respective computing devices (2, 5), so that when a large amount of free computing power is currently available on the motor vehicle (3), a larger proportion is performed on the motor vehicle-side computing device (2) than when the motor vehicle-side computing device is heavily utilized.