Process and hybrid system

Abstract

The invention relates to a method and a hybrid system for configuring an on-board communication network of a vehicle, as well as a use.

Description

[0001] The invention relates to a method and a hybrid system for configuring an on-board communication network of a vehicle and to a use thereof.

[0002] Modern vehicles, especially motor vehicles, today feature an ever-growing number of components, especially control units, and systems that must communicate with each other via an on-board communication network to fulfill their functions. Prominent examples of this include entertainment systems, the increasingly used driver assistance systems, and the emerging control systems for partially or highly automated driving, including autonomous driving. The increase in electrical components in vehicles has led to a very high level of complexity in the entire data processing system of a motor vehicle.

[0003] The well-known on-board communication networks for vehicles include, in particular, so-called vehicle buses, such as the well-known CAN buses and LIN buses or an Ethernet-based communication network, in which a plurality of network nodes, usually in the form of control units, are connected to a common transmission medium, the signal processing network, for the transmission of messages.

[0004] Such signal processing networks, also called on-board networks, with e.g. sensors and processing units in a vehicle, have complex dependencies.

[0005] For example, on-board networks can be implemented in a bus topology controlled by Telematics Control Units (TCUs). The control units communicate with each other and with the connected sensors / actuators via fieldbuses. The various fieldbus systems, such as CAN bus, LIN bus, or FlexRay, communicate with each other via gateways. The fieldbus clusters, in turn, consist of several Electronic Control Units (ECUs) connected to one or more fieldbuses.

[0006] In an on-board network, the control units functionally correspond to the nodes of a classic network, which fulfills certain functionalities that can be accessed by slaves.

[0007] When adding a component, corresponding network nodes must also be added to the on-board communication network.

[0008] Such procedures are known.

[0009] DE 10 2016212755 B3 discloses a method for registering a vehicle component with an Ethernet vehicle network.

[0010] Likewise, when replacing a control unit, the on-board communication networks must be checked for functionality.

[0011] Previously, it was necessary to manually test whether such a component could be safely replaced.

[0012] It is therefore an object of the invention to provide an improved method for configuring an on-board communication network of a vehicle as well as a hybrid system and a use.

[0013] The object is achieved by a method having the features of claim 1 and a hybrid system having the features of claim 5 as well as a use having the features of claim 10.

[0014] The problem is solved by a method for configuring an on-board communication network of a vehicle, comprising the steps: - Providing a digital computer, - Providing the on-board communication network with a set of network nodes in the digital computer, - Providing in the digital computer a complete set of all data to be transmitted, which are transmitted via the on-board communication network to the respective network nodes, - Providing a matrix in the digital computer, which represents a directed graph connecting the network nodes, with matrix elements representing the transmitted data for a corresponding graph section in the graph from a first network node to a second network node, - Providing an adiabatic quantum system, wherein the quantum system is associated with a corresponding energy landscape, wherein the quantum system is coupled to the digital computer, - Generating an energy function based on the matrix and thus the individual graph sections by applying a min-cut (minimal cut) function and generating an initial initial state for the quantum system based on the energy function and initializing the energy landscape based on the initial state, - Changing the initial state using the energy function in the quantum computer into an energetic ground state, where the energetic ground state encodes a global minimal cut on the graph formed as a matrix, - Measurement of the basic energetic state by the digital computer after adiabatic change and encoding of the measurement, as well as clustering as configuration of the network nodes of the on-board communication network by the digital computer based on the encoding.

[0015] A cut is a set of graph segments that divides the connected graph of the on-board communication network into at least two subsets that are as disjoint or as unrelated as possible. A cut divides the network node set of the graph into two non-empty clusters. The initial state of a quantum computer typically corresponds to a solvable quantum system in which the qubits are initialized to the ground state. This initial state is then transferred to the matrix of the problem, which ideally also brings the qubits to the ground state of the problem. In the adiabatic quantum computer, the matrix is ​​translated into a coupling strength between the qubits.

[0016] The following as the on-board communication network and as the set of network nodes, as well as the amount of all data to be transmitted with designated.

[0017] Furthermore, WI as a matrix with matrix elements wi,j⊂D, with i,j∈I which transmits the data for the graph section from a first network node i∈I to a second network node j∈I represent.

[0018] The operating principle of a quantum computer has been known as a theoretical concept for a long time and has been implemented in practice for some time.

[0019] While in conventional digital computers, information is represented in bits, which are essentially switches in an on or off position, the information in quantum computers, which is also essentially binary, is represented by quantum mechanical states. These can typically be the spin of an electron, energy levels of atoms, or the current direction, charge, or magnetic flux in a superconductor. Regardless of the choice of physical implementation, such a quantum mechanical two-state system is called a qubit.

[0020] A quantum annealer, such as a D-Wave machine, can be used as a quantum computer.

[0021] At the beginning of the calculation, the coupling of the quantum annealer qubits is initialized with the initial state and changed to the lowest energy ground state.

[0022] Qubits are a quantum mechanical system with two states (quantum superposition), whereby a qubit can be in a superposition of these two states and can also be entangled with other qubits. A change in the coupling of the qubits thus simultaneously affects the state of all other entangled qubits.

[0023] Entangled qubits thus behave over any distance as if they were connected to each other. For example, a measurement of the spin of one electron determines the spin state of the other. Thus, if one qubit is changed, the entangled qubit adopts a corresponding change without any time delay and regardless of the distance.

[0024] According to the invention, the problem is approximated in such a way that it can be described by a maximally binary energy function. According to the invention, a sufficiently accurate approximation must be found for the problem to be input into the quantum annealer, enabling adiabatic optimization.

[0025] To do this, the task must be formulated as a QUBO (Quadratic Unconstrained Binary Optimization) task with binary variables over a quadratic objective function without constraints.

[0026] According to the invention, it was recognized that signal processing networks, such as sensors and processing units in a car, have complex dependencies, which can be eliminated by modularizing these networks, ie by creating clusters in the network, whereby these clusters may have no or only a few connections and can operate largely independently of each other.

[0027] This requires analyzing the connections and data transfer between the individual network nodes. Due to the multitude of transfer options, which increases with the number of network nodes, such a highly complex analysis was previously impossible and therefore not performed in the current state of the art.

[0028] According to the invention, a digital computer and an adiabatic quantum system (quantum annealer) are provided for this purpose, wherein the quantum annealer is associated with a corresponding energy landscape, and wherein the quantum annealer is coupled to the digital computer.

[0029] According to the invention, an energy function is now generated based on the matrix and thus the individual graph sections by applying a min-cut (minimal cut) function and an initial state for the quantum system is generated based on the energy function and the energy landscape is initialized based on the initial state.

[0030] Furthermore, the energy function is varied into an energetic ground state using the min-cut function in the quantum computer. The energetic ground state encodes an energy function. The energetic ground state represents a global minimal cut on the graph, which is formed as a matrix. The global cut allows for a clustering of the on-board communication network into different subnetworks that have as few connections to each other as possible. The change (cut) thus corresponds to a cleanup of the graph of redundant data streams, as far as possible.

[0031] By measuring the ground energy state after an adiabatic change and encoding the measurement, the digital computer can then cluster the network nodes of the on-board communications network based on the encoding. By connecting a quantum computer and the digital computer in conjunction with the method according to the invention, such cleanup and separation into clusters can also be performed with a large number of data streams / data and network nodes.

[0032] The method according to the invention, which is based on the specific structure of signal processing networks, which can consist of software, hardware or a mixture of both, allows network nodes to be clustered as clusters.

[0033] Clustering creates subnetworks that operate independently of one another, making it possible to tailor the signal processing network to the customer.

[0034] This makes it easier to modify networks by exchanging clusters, which has a positive impact on efficiency, speed, reliability and emissions.

[0035] Furthermore, by reducing the number of network nodes in subnetworks, the security and reliability analysis can be simplified. In further training, each network node i∈I a binary variable x i which indicates whether the network node i∈I a first cluster CO or a second cluster C1. This assignment creates a maximal quadratic energy function, allowing a quantum annealer to be used.

[0036] In further development, the energy function H is given by: H=min∑i,i≠j∈I(−xi−xj+2xixj)∀j,i ∈I where is the set of network nodes.

[0037] By applying the min-cut function method to the matrix graph, a minimal cut is found, allowing the network nodes to be clustered so that a minimal number of graph segments exist between the clusters. A cut refers to a partition of the node set of a graph. The basic principle of the min-cut function is based on the recursive division of the total node set into two or more segments.

[0038] In further training, the basic energetic state of the matrix WI, as a modified matrix with the W'I Matrix elements w'i,j⊂ⅅ, with i,j∈1 be given with w'i,j:=|wi,j\∪i≠k∈Iwk,j|

[0039] The matrix elements can w'i,j⊂D, with i,j∈I be converted by the digital computer into a real clustered on-board communication network.

[0040] This creates a redundant data stream wk,j,i≠k∈I prevented.

[0041] Furthermore, the object is achieved by a hybrid system for configuring an on-board communication network of a vehicle, wherein the hybrid system comprises a digital computer with a memory unit, on which an on-board communication network of the vehicle with a set of network nodes is stored, and wherein the hybrid system further comprises a complete set of all data to be transmitted to the respective network nodes via the on-board communication network, and wherein the hybrid system is designed to generate a matrix as a directed graph connecting the network nodes, with matrix elements representing the transmitted data for a corresponding graph section in the graph from a first network node to a second network node, further comprising an adiabatic quantum system, wherein the quantum system is associated with a corresponding energy landscape, wherein the quantum system is coupled to the digital computer, and wherein the digital computer is configured to generate an energy function based on the matrix and thus the individual graph sections, to apply a min-cut (minimal cut) function and to generate an initial initial state for the quantum system based on the energy function, and to initialize the energy landscape based on the initial state, and wherein the adiabatic quantum system is configured to change the initial state into an energetic ground state using the energy function, wherein the energetic ground state encodes a global minimum cut on the graph formed as a matrix, and wherein the digital computer is configured to measure the energetic ground state after adiabatic change and encode the measurement, as well as cluster the configuration of the network nodes of the on-board communication network based on the encoding.

[0042] This means that the energetic ground state encodes a minimal energy function.

[0043] This ensures that the clustered subnetworks have as few connections to each other as possible. This change thus represents a cleanup of redundant data streams from the graph, as far as possible.

[0044] The hybrid system is particularly designed to carry out the method according to the invention.

[0045] Furthermore, the advantages and advantageous embodiments of the method can be transferred to the hybrid system.

[0046] The on-board communication network as well as the network nodes can be designed virtually, so that clustering and the exchange of subnetworks can be tested before the actual prototype construction.

[0047] Due to the ever-increasing number of connection options, both the number of data itself and the number of network nodes in an on-board communications network, it has previously been impossible to meaningfully cluster the on-board communications network. This is now possible for the first time using the method and hybrid system according to the invention.

[0048] In a further development, a display unit is provided, whereby the on-board communication network with the set of network nodes can be displayed. The on-board communication network and the network nodes can be implemented virtually, allowing clustering and modification of a subnetwork to be tested prior to the actual prototype construction. This virtual design can save both time and money.

[0049] In particular in autonomous vehicles, the security of data transmission can be tested in the event of a failure or malfunction of a network node, so that if necessary, for example, a redundancy node connecting the subnetworks can be installed to ensure fail-safe operation of the on-board communication network.

[0050] The network nodes are at least partially designed as a control unit of a vehicle component, so that at least part of the data to be transmitted can be generated by the control unit and / or its vehicle component.

[0051] Furthermore, the object is achieved by using a hybrid system as described above and / or a method as described above for clustering an on-board communication network into different clusters.

[0052] Further features and advantages of the present invention will become apparent from the following description with reference to the accompanying figures, which schematically show: Fig. 1: a method according to the invention schematically, Fig. 2: a hybrid system and a clustered on-board communication network.

[0053] Fig. 1 shows the method according to the invention schematically.

[0054] In a first step S1, an on-board communication network with a set of network nodes The on-board communications network with a set of network nodes can be provided virtually on a display unit.

[0055] Furthermore, a complete set of all data to be transmitted D which is transmitted via the on-board communication network to the respective network nodes It is also known which data wi,j⊂D, with i,j∈I from a first network node i∈I to a second network node j∈I be transferred.

[0056] In a second step S2, the corresponding Data wi,j⊂D, with i,j∈I as matrix elements of a matrix WI which has a directed, the network nodes interconnecting graphs. Thus, the data corresponds wi,j⊂D, with i,j∈I the transmitted data for a corresponding graph section in the graph from a first network node i∈I to a second network node j∈I. The Matrix WI thus represents a weighted, directed graph, which represents the individual network nodes connects each other via data technology.

[0057] In a third step, S3 with S(j)={i∈I|wij≠∅} the set of network nodes i∈I, which data to the network node j∈I sends, be segmented.

[0058] Furthermore, in the third step S3, the set of network nodes which receive data from the pre-selected selection network node i∈I received, as a segment set ℝ(i)={j∈I|wij≠∅} segmented.

[0059] Also, the amount of data that a selected selection network node i∈I as a sending set, Di:=Uj∈ℝ(i)wij⊂D segmented.

[0060] In a fourth step S4, an adiabatic quantum system is provided, in particular a quantum annealer 3, wherein the quantum annealer 3 is associated with a corresponding energy landscape. The quantum annealer 3 is coupled to the digital computer 2.

[0061] In a fifth step S5, an energy function H is calculated based on the matrix WI generated, where is the set of network nodes.

[0062] Each network node i∈I a binary variable x i which indicates whether the network node i∈I a first cluster CO or a second cluster C1. This assignment creates a maximal quadratic energy function, allowing a quantum annealer to be used. H = min ∑i,i≠j∈I(−xi−xj+2xixj) ∀j,i ∈I

[0063] The energy function H represents a min-cut (minimal cut) function.

[0064] In a sixth step (S6), the energy landscape of quantum annealer 3 is initialized using an initial state. The initial state of the quantum computer typically corresponds to a solvable quantum system in which the qubits are initialized to the ground state. This initial state is then transferred to the matrix of the problem, which ideally also brings the qubits to the ground state of the problem. In the adiabatic quantum computer, the matrix is ​​translated into a coupling strength between the qubits.

[0065] In a seventh step S7, the initial state is changed into an energetic ground state using the energy function, where the energetic ground state has a global minimal cut 4 on the matrix WI trained graphs encoded.

[0066] Based on the minimal cut 4, a clustering of the on-board communication network can be carried out in an eighth step S8 into at least different clusters C0 and C1 as subnetworks. Clusters C0 and C1 have as few connections to each other as possible. This change thus represents a cleanup of redundant data streams from the graph, as far as possible.

[0067] The energetic ground state of the matrix W′I which has a directed, the network nodes interconnecting graphs, describes matrix elements w'i,j⊂ D, with i,j ∈ I have, with w′ i,j:=|wi,j / ∪i≠k∈Iwk,j|

[0068] The matrix elements can w′ i,j⊂D, with i,j∈I from the digital computer 2 into a real clustered on-board communication network This creates a redundant data stream wk,j,i≠k∈I prevented.

[0069] Fig. 2 shows the hybrid system 1 with a digital computer 2 and a quantum annealer 3. The hybrid system 1 is designed to carry out the above method and to create an on-board communication network based on the above method into clusters, here into a first cluster C0 and a second cluster C1.

[0070] By means of the hybrid system 1 according to the invention and the method according to the invention, a metric is established which is tailored to a signal processing network and which is useful for grouping in a signal processing network.

[0071] A min-cut function is used, which generates a minimal cut 4. Such a cut 4 divides the network node set of the graph into at least two non-empty clusters C0,C1. For this purpose, a matrix WI created and based on the matrix WI an energy function H is generated.

[0072] The method according to the invention and the hybrid system 1 according to the invention enable a modularization of the on-board communication network be achieved, ie the creation of clusters C0,C1 that are little connected to each other and can function largely independently of each other.

[0073] This enables the on-board communication network for customers. This allows an on-board communication network be changed more easily by making the CO,C1 clusters easier to replace. This also has a positive impact on the efficiency, speed, reliability, and emissions of the on-board communication network. out of.

[0074] This also simplifies safety and reliability analysis. List of reference symbols 1 hybrid system 2 digital computers 3 quantum annealers Onboard communications network Network nodes C0,C1 clusters matrix QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] DE 10 2016212755 B3

[0009]

Claims

[1] Method for configuring an on-board communication network (B) of a vehicle, comprising the steps: - Providing a digital computer (2), - Providing the on-board communication network (B) with a set of network nodes (I) in the digital computer (2), - Providing in the digital computer (2) a complete set of all data to be transmitted (D) which are transmitted via the on-board communication network (B) to the respective network nodes (I) be transferred, - Providing a matrix (WI) in the digital computer (2), which has a directed, the network nodes (I) interconnecting graphs, with matrix elements (wi,j⊂D, with i,j∈I), which transmits the transmitted data for a corresponding graph section in the graph from a first network node (i∈I) to a second network node (j∈I) ∈ represent, - providing an adiabatic quantum system, wherein the quantum system is associated with a corresponding energy landscape, wherein the quantum system is coupled to the digital computer (2), - Generate an energy function (H) based on the matrix (WI) and thus the individual graph sections by applying a min-cut (minimal cut) function and generating an initial initial state for the quantum system based on the energy function (H) and initializing the energy landscape based on the initial state, - Changing the initial state using the energy function (H) in the quantum computer into an energetic ground state, where the energetic ground state encodes a global minimal cut on the graph formed as a matrix, - Measuring the energetic ground state by the digital computer (2) after adiabatic change and encoding the measurement, as well as clusters as configuration of the network nodes (i∈I) the on-board communication network (B) by the digital computer (2) based on the encoding. [2] Method according to claim 1, characterized by that each network node (i∈I) a binary variable (x i ) which indicates whether the network node (i∈I) is assigned to a first cluster (C0) or a second cluster (C1). [3] Method according to claim 2, characterized by that the energy function (H) is given by H=min∑i,i≠j∈I(−xi−xj+2xixj)∀j,i ∈I where (I) is the set of network nodes. [4] Method according to one of the preceding claims, characterized by that the energetic ground state of the matrix (WI), as a modified matrix (W'I) with the matrix elements (wi,j⊂D, with i,j∈I) be given by w'i,j:=|wi,j\∪i≠k∈Iwk,j| where (I) is the set of network nodes. [5] Hybrid system (1) for configuring an on-board communication network (B) of a vehicle, where the hybrid system (1) comprises a digital computer (2) with a memory unit, on which an on-board communication network (B) of the vehicle with a set of network nodes (I) is stored, and where the hybrid system (1) further comprises a complete set of all data to be transmitted (D) which is transmitted via the on-board communication network (B) to the respective network nodes (I) to be transferred characterized by , that the hybrid system (1) is designed to form a matrix (WI) as a directed, the network nodes (I) interconnecting graphs, with matrix elements (wi,j⊂D,with i,j∈I), which transmits the transmitted data for a corresponding graph section in the graph from a first network node (i∈I) to a second network node (j∈I) represent, generate, further comprising an adiabatic quantum system, wherein the quantum system is associated with a corresponding energy landscape, wherein the quantum system is coupled to the digital computer (2), and wherein the digital computer (2) is configured to generate an energy function (H) based on the matrix (WI) and thus the individual graph sections to apply a min-cut (minimal cut) function and to generate an initial initial state for the quantum system based on the energy function (H), as well as to initialise the energy landscape based on the initial state, and wherein the adiabatic quantum system is designed to effect a change of the initial state into an energetic ground state using the energy function (H), wherein the energetic ground state encodes a global minimal cut on the graph formed as a matrix, and wherein the digital computer (2) is designed to measure the energetic ground state after adiabatic change and to encode the measurement, as well as to cluster as a configuration of the network nodes (i∈I) the on-board communication network (B) based on the encoding. [6] Hybrid system (1) according to claim 5, characterized by that a display unit is provided and the on-board communication network (B) with the set of network nodes (I) can be displayed. [7] Hybrid system (1) according to claim 5 or 6, characterized by that the network nodes (I) are at least partially designed as a control unit of a vehicle component, so that at least part of the data to be transmitted (D) can be generated by the control unit and / or its vehicle components. [8] Hybrid system (1) according to one of the preceding claims 5 to 7, characterized by that the digital computer (2) is designed to provide each network node (i∈I) a binary variable (x i ) which indicates whether the network node (i∈I) is assigned to a first cluster (C0) or a second cluster (C1). [9] Hybrid system (1) according to claim 8, characterized by that the energy function (H) is given by H=min ∑i,i≠j∈I(−xi−xj+2xixj)∀j,i∈I where (I) is the set of network nodes. [10] Use of a hybrid system (1) according to one of the preceding claims 5 to 9 and / or a method according to one of the preceding claims 1 to 4, for clustering an on-board communication network (B) into different clusters (C0,C1).

Images