Digital simulation platform based on message bus
By introducing a message bus-based architecture into the digital simulation platform, the stability, security and scalability problems of traditional platforms when dealing with large-scale high-concurrency scenarios are solved, and efficient and stable simulation task processing capabilities are achieved.
Patent Information
- Application Number
- CN202510166429.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-14
- Publication Date
- 2025-06-24
AI Technical Summary
When traditional digital simulation platforms deal with simulation tasks in large-scale, high concurrency and complex scenarios, they have problems such as high network stability requirements, weak communication security, and poor scalability of communication nodes, which cannot effectively support the current complex simulation service needs.
The digital simulation platform architecture based on the message bus is adopted, including the view layer, the service layer and the data access layer. Through the message bus, high concurrent communication and interfaces are realized and fast correspondence between algorithm modules, algorithms and simulation environments, and algorithms and equipment models are ensured, ensuring clock consistency and state accuracy during transmission, and providing security measures such as adaptive interface management and data encryption.
It realizes the efficient and stable operation of the digital simulation platform in large-scale, high-concurrency and complex scenarios, improves network stability, communication security and node scalability, and can support the massive rapid deduction training and testing verification of multiple intelligent game confrontation algorithms.
Smart Images
Figure CN120197251A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a digital simulation platform based on a message bus, belonging to the field of computer science. Background Art
[0002] With the rapid development of intelligent technologies and the large-scale application of intelligent equipment, the application tasks of digital simulation platforms have increasingly shown characteristics such as the expansion, complexity, and high precision of scene scales, and the diversification, refinement, and componentization of model and intelligent algorithm types. The demand for data communication and interaction in digital simulation platforms has shown a geometric growth trend. Traditional digital simulation platforms are designed with a "point-to-point" communication method as the basis. They have significant bottlenecks such as high requirements for network stability, weak communication security, and poor scalability of communication nodes, and cannot effectively support the increasing demand for massive data interaction in the platform, and can no longer meet the current increasingly complex simulation service requirements. Summary of the Invention
[0003] The technical problem to be solved by the present invention is: to overcome the deficiencies of the prior art, and provide a digital simulation platform based on a message bus to achieve the simulation task requirements of high-efficiency and stable operation of the digital simulation platform for processing large-scale high-concurrency complex scenarios.
[0004] The technical solution of the present invention is: a digital simulation platform based on a message bus, and the architecture of the digital simulation platform includes: a view layer, a service layer, and a data access layer, where:
[0005] The view layer is responsible for front-end display, receives interactive input from users, configures corresponding training environments and methods, receives simulation situation data, and displays the algorithm training process in real time; it can perform data interaction with the service layer;
[0006] The data access layer includes an algorithm module and an equipment model, which are used to generate and store simulation data and situation information, and provide the stored data to the service layer;
[0007] The service layer is used for data transmission, interface management, and calculation, and includes a message bus and service nodes; where: the message bus performs high-concurrency data communication and transmission with the view layer and the data access layer, realizes the rapid correspondence of interfaces and protocol field conversion between algorithm modules, between algorithm modules and simulation environments, and between algorithm modules and equipment models, ensures consistent clocks and accurate states during the transmission process, and at the same time provides adaptive interface management; the service nodes obtain various data through the message bus and provide various computing platforms and simulation environments.
[0008] Preferably, the service nodes include: a simulation engine service node, an equipment calculation service node, an AI algorithm service node, a database service node, and a simulation scheduling and control service node. Specifically:
[0009] Simulation Engine Service Node: Provides time synchronization services for simulation deduction, uniformly integrates and schedules equipment services, AI algorithm services, and database services, distributes situation information to equipment computing service nodes as needed, and performs global adjudication calculations for simulation deduction;
[0010] Equipment Computing Service Node: Responsible for equipment simulation propulsion calculations, status updates, equipment detection, communication capabilities, and damage capabilities calculations. At the same time, it manages the sending and receiving of equipment situation packetization;
[0011] AI Algorithm Service Node: Responsible for starting algorithm modules, accessing the corresponding simulation environment for algorithms, and responding to AI instructions;
[0012] Database Service Node: Responsible for simulation data management, scenario data management, and equipment model data management, providing data support for building the simulation environment;
[0013] Simulation Management and Service Node: Cooperates with the visual layer to establish a complex simulation service environment for users and manages and controls the processes of multi-scenario simulations.
[0014] Preferably, when each node runs in the architecture:
[0015] Each computing node establishes a channel connection with other computing nodes through a centralized control and processing high-concurrency message bus, and real-time detects the software and hardware transformations of the corresponding computing node.
[0016] Preferably, the message bus includes a centralized control and processing high-concurrency message bus and an adaptive interface management module, where:
[0017] The centralized control and processing high-concurrency message bus is responsible for high-concurrency data communication and transmission with the view layer and the data access layer, realizes the rapid correspondence of interfaces between algorithm modules, between algorithm modules and the simulation environment, and between algorithm modules and equipment models, and protocol field conversion, ensuring consistent clocks and accurate states during the transmission process;
[0018] The adaptive interface management module is responsible for adaptive interface management during the data transmission process, including: simulation platform and algorithm interface management unit, heterogeneous interface access management unit, and multi-granularity model adaptive access management unit.
[0019] Preferably, the centralized control and processing high-concurrency message bus maps different protocols into a unified interaction framework; encapsulates algorithm interfaces according to the corresponding protocols, automatically schedules and executes appropriate algorithms to process received requests; through protocol conversion algorithms, converts different protocols into formats that the framework can process, and uses interface access protocols for asynchronous transmission.
[0020] Preferably, the simulation platform and algorithm interface management unit encapsulates the data exchanged between each algorithm module and the simulation platform to construct a standardized data interaction interface; constructs an algorithm test scheduling management mechanism, enabling each algorithm module and the simulation platform to interact based on RPC protocol network communication.
[0021] Preferably, the heterogeneous interface access management unit comprehensively uses RESTful-based web service integration technology, RPC remote call protocol, and publish-subscribe-based bus distributed communication method to complete the integration between tools for the transmission and call of different static data and dynamic data; where the static data includes various simulation data, and the dynamic data includes algorithms, situation information, and equipment models.
[0022] Preferably, the multi-granularity model adaptive access management unit accesses each equipment model to the centralized control processing high-concurrency message bus through a plug-in method of distributed loose-coupling asynchronous communication, and each accessed equipment model communicates through the publish-subscribe bus distributed communication method.
[0023] Preferably, when communicating through the publish-subscribe bus distributed communication method:
[0024] The equipment model generates situation information, and the simulation engine service node sends it to the centralized control processing high-concurrency message bus;
[0025] After the equipment computing service node receives the message in the centralized control processing high-concurrency message bus, it updates the situation information of the equipment model and sends the updated situation information to the centralized control processing high-concurrency message bus;
[0026] The simulation engine service node receives the situation information after the update of each equipment model and synchronously updates the situation information.
[0027] The present invention has the following advantages compared with the prior art:
[0028] (1) The design of the message bus of the present invention improves the problem of high requirements for network stability in the traditional digital simulation platform communication method, achieving relatively stable communication through various methods even when there are faults or delays in the network between modules;
[0029] (2) The centralized control processing high-concurrency message bus of the present invention maps different protocols to a unified interaction framework, strengthens the security of network communication in the digital simulation platform, and realizes data encryption, access control, and log monitoring for network communication between modules;
[0030] (3) The cooperative design of the message bus and service nodes of the present invention improves the node scalability of the digital simulation platform, reduces the communication overhead of the system, enhances the response speed, and achieves the effect of supporting the massive, rapid deduction, training, testing, and verification of various intelligent game confrontation algorithms. Description of the Drawings
[0031] Figure 1 It is a composition diagram of the digital simulation platform based on the message bus of the present invention;
[0032] Figure 2 It is a schematic diagram of the centralized control and processing high-concurrency message bus of the present invention;
[0033] Figure 3 It is a schematic diagram of the C / S architecture of the present invention;
[0034] Figure 4 It is a distributed architecture diagram of the simulation platform of the present invention. Detailed Implementation Manner
[0035] The present invention proposes a digital simulation platform based on a message bus that can be applied to large-scale, high-concurrency, and multi-entity complex task scenarios to solve the problems of the prior art.
[0036] To achieve the above object, the core technical solution of the present invention mainly relies on a digital simulation platform constructed based on the general distributed simulation platform architecture of the message bus technology, as Figure 1 shown, and the specific content is as follows.
[0037] A digital simulation platform based on a message bus is designed with a general distributed simulation platform architecture, and includes a simulation deduction module, a model editing module, a scenario editing module, and a basic function module, wherein:
[0038] The general distributed simulation platform architecture is the core of the digital simulation platform based on the message bus, and is used to construct a basic software framework and define the functions and interaction logics between modules. It adopts centralized control to process high-concurrency message buses and adaptive interface management, and communicates and exchanges data with the simulation deduction module, model editing module, scenario editing module, and data playback module in the platform;
[0039] The following is a detailed description of the components of this platform
[0040] I. General Distributed Simulation Platform Architecture
[0041] (1) Overall Architecture Design
[0042] The overall design of the general distributed simulation platform architecture adopts the C / S architecture design, that is, the client / server architecture mode, as Figure 2 and Figure 3As shown in the figure, the front-end and back-end separation design concept is adopted to decouple the front-end display and simulation data processing. By integrating the distributed architecture and parallel simulation training technology, such as Figure 4 As shown in the figure, the flexibility and efficiency of the distributed architecture are combined with the efficiency of parallel simulation training.
[0043] This architecture abstractly splits the simulation platform into multiple independent computing nodes (front-end display node, simulation engine service node, equipment service computing node, AI algorithm service node, database service node, simulation scheduling and control service node). Among them:
[0044] Front-end display node: Located in the view layer, its main functions include receiving interactive input from users, configuring corresponding training environments and methods, receiving simulation situation data, and displaying the algorithm training process in real time.
[0045] Simulation engine service node: Its main functions include the unified integration and scheduling of simulation deduction timekeeping services, equipment services, AI algorithm services, and database services, the on-demand distribution of situation information, and the global adjudication calculation for simulation deduction of detection, communication, and damage.
[0046] Equipment computing service node: Its main functions include equipment simulation propulsion calculation and status update, equipment detection, communication, damage ability calculation, and equipment situation packet sending and receiving management.
[0047] AI algorithm service node: Its main functions include algorithm startup and situation sending and receiving management, and access to the corresponding simulation environment of the algorithm and AI instruction response.
[0048] Database service node: Its main functions include simulation data management, scenario data management, and equipment model data management, providing data support for the construction of the simulation environment.
[0049] Simulation management and service node: Its main functions include establishing a complex simulation service environment for users and managing and controlling the processes of multi-scenario simulations.
[0050] The processes of the above nodes running in this architecture are as follows:
[0051] After the software is started, each computing node will register with the network layer after startup. At the same time, the network layer will establish a channel connection with the computing node and detect the software and hardware changes of the computing node in real time. The network layer is responsible for managing the online and offline changes of the computing node programs, can listen to relevant topic contents, and can establish a connection with the computing node immediately after it goes online, realizing zero-waiting connection and real-time communication between the computing task nodes of the simulation platform. By adopting the distributed soft bus design, the protocol shelf and the software-hardware cooperation layer shield the communication protocol differences between various communication nodes.
[0052] The architecture enables communication and collaboration among computing nodes via network connections to jointly complete tasks, improving the scalability and reliability of the intelligent algorithm simulation training platform, enhancing simulation efficiency and performance, thus fully utilizing multi-computer computing resources and improving the training efficiency of intelligent algorithms.
[0053] (2) Detailed Architecture Design
[0054] This architecture is designed to support large-scale high-concurrency distributed digital simulation. Its core components are a centralized control high-concurrency message bus based on message bus technology and adaptive interface management. Specifically:
[0055] 1. Centralized Control and Processing of High-Concurrency Message Bus
[0056] With the centralized control and processing message bus as the core, while ensuring a clear structure, as Figure 2 shown, it realizes rapid correspondence of interfaces between algorithms, between algorithms and the simulation environment, and between algorithms and equipment models, protocol field conversion, supports high-concurrency communication transmission during algorithm simulation operation, and ensures consistent clock and accurate status.
[0057] It realizes unified information interaction of heterogeneous algorithms through methods such as remote procedure scheduling (RPC) based on object-oriented methods and single object access protocol (SOAP) facing messages, and realizes the combination of multiple transport layer or application layer protocols, such as TCP / HTTP / SMTP, etc. The main implementation routes are as follows:
[0058] 1) Encapsulate the algorithm message interface into an RPC protocol or other interface access protocol that supports remote procedure calls in a distributed system, enabling the client to send data status query requests through the network server.
[0059] 2) Conduct asynchronous message passing through message-oriented middleware to enable communication between different algorithm applications / containers via messages.
[0060] 3) Combine multiple transport layer or application layer protocols, design and implement a protocol abstraction layer to handle various different protocols and be able to map them into a unified interaction framework. Through protocol conversion algorithms, convert different protocols into formats that the framework can handle.
[0061] 4) Through the algorithm abstraction layer, handle different algorithms, map algorithm interfaces into a unified framework. Design and implement an algorithm scheduling and execution mechanism to automatically select and execute appropriate algorithms to process received requests.
[0062] Implement the standard resource sharing and interaction services of multiple heterogeneous algorithms in the way of centralized control to process the message bus, and autonomously allocate information routing according to information content and permissions. Support dynamic and flexible access through data routing technology according to business rules. At the same time, the exchange between different data formats based on the centralized control processing message bus platform can support various data communication and transmission methods, and provide security mechanisms for data exchange and functions such as centralized unified deployment, monitoring, tracking, logging, and detection to meet the requirements of distributed deployment and centralized management.
[0063] 2. Adaptive Interface Management
[0064] Adaptive interface management mainly includes a simulation platform and an algorithm interface management unit, a heterogeneous interface access management unit, and a multi-granularity model adaptive access management unit. Among them:
[0065] Simulation platform and algorithm interface management unit: To interact between the simulation platform and algorithms running independently and mounted, it is necessary to achieve decoupling based on cross-platform and cross-programming languages. Interact between the simulation platform and algorithms through RPC network communication to form a data flow loop of controllable instructions; further encapsulate the interaction between the algorithm and the simulation platform, build an algorithm test scheduling management mechanism, facilitate local debugging of algorithm docking, and reduce the difficulty of interface development and operation and maintenance. The algorithm and the simulation platform carry out interactions based on the RPC protocol network communication. After the algorithm module receives the status data of the simulation platform, it generates a feature representation that suits the data format and requirements of artificial intelligence algorithms (mainly deep reinforcement learning) for learning and training. At the same time, the algorithm module encapsulates the data returned by the simulation platform to build a standardized test environment data interaction interface for the algorithm.
[0066] Heterogeneous interface access management unit: The heterogeneous interface access module flexibly adopts corresponding interface protocols according to the characteristics of the transmitted data types, registers various heterogeneous interaction interfaces as standardized interfaces of the evaluation platform, achieves decoupling based on cross-platform and cross-programming languages, and interacts between various tools through network communication to form a data flow loop. Comprehensively use RESTful-based web service integration technology, RPC remote call protocol, and publish-subscribe-based soft bus distributed communication middleware protocol methods to achieve the transmission and communication of static data such as different target data, experimental data, and environmental data, dynamic data such as perception algorithms, decision-making algorithms, and knowledge rules, and dynamic data such as equipment models and digital mock-ups, as well as the invocation of algorithms, and complete the integration between tools.
[0067] Multi-granularity Model Adaptive Access Management Unit: It realizes the flexible access of tactical / campaign-level, mission-level, engagement-level, and engineering-level equipment models through a plug-in method of distributed loose-coupling asynchronous communication, and realizes the distributed loose-coupling asynchronous communication between the equipment model and the simulation engine through the publish-subscribe mode of the soft bus. In the publish-subscribe mode of the soft bus, each access equipment model does not directly perform explicit calls, but communicates by subscribing to and publishing messages. The equipment model based on docker containerization can be independently mounted and run on the soft bus. According to the soft bus, a modular design and an equipment model plug-in mechanism can be realized, so that the engine can easily integrate and support the flexible and scalable access of different granularity equipment models.
[0068] The publish-subscribe mode of the soft bus consists of three roles: publisher, subscriber, and soft bus. The above model access management workflow is as follows:
[0069] 1) The simulation platform generates the situation information of the equipment model in the battlefield and sends it to the soft bus.
[0070] 2) After the equipment model service node in the soft bus receives the message as a subscriber, it updates the state of the equipment model. At the same time, the equipment model can also generate signal-level signals and send them to the soft bus. After the equipment calculation service node receives the signal-level signals, it updates the state and sends the updated state to the soft bus.
[0071] 3) The simulation engine service node receives the state information updated by each equipment model and updates the situation information.
[0072] In view of the problems in the actual process of accessing equipment models, such as different equipment models using different devices, systems, and communication protocols, etc., combined with multi-protocol automatic conversion technology, data conversion and adaptation between different communication protocols can be effectively carried out, compatible with different systems and devices, and realizing the interconnection and interoperability between systems.
[0073] In a digital simulation platform based on a message bus, a general distributed simulation platform architecture is used to build a basic software framework, defining the functions and interaction logics between modules. It adopts centralized control to process high-concurrency message buses and adaptive interface management, communicates with and sends and receives data from the simulation deduction module, model editing module, scenario editing module, and data playback module in the platform;
[0074] The general distributed simulation platform architecture is designed as a C / S architecture, that is, a client / server structure. The client part is responsible for completing the interaction tasks with users, and the server side is responsible for the management of interaction data. It adopts a decoupled design in the way of separating the front end and the back end, and the front end and the back end communicate through API interfaces, and adopts centralized heterogeneous interface distributed bus control processing.
[0075] The centralized heterogeneous interface message distributed bus realizes functions such as heterogeneous messages, dynamic conversion, configurable routing, message logging and monitoring, service discovery, SLA / SLO support, and policy-driven security through a network-based distributed bus. At the same time, the model access structure design adopts a plug-in method of distributed loose-coupling asynchronous communication. It uses a soft bus publish-subscribe framework to achieve distributed loose-coupling asynchronous communication between the equipment model and the simulation engine, and then realizes the flexible access of tactical / campaign-level, mission-level, engagement-level, and engineering-level equipment models.
[0076] The present invention proposes a digital simulation platform based on a message bus. Based on the C / S architecture, with the design concept of separating the front and back ends, based on a distributed system, and with a centralized control and processing message bus having a unified data transmission standard and protocol as the core, by means of the parallel computing power of a large-scale hardware cluster, it realizes the rapid correspondence of interfaces between algorithms, between algorithms and the simulation environment, and between algorithms and equipment models, protocol field conversion, and protocol format adaptation, so as to realize the efficient and stable operation of the digital simulation platform to handle the simulation task requirements of large-scale high-concurrency complex scenarios.
[0077] The content not described in detail in the specification of the present invention belongs to the prior art well-known to those skilled in the art.
Claims
1. A digital simulation platform based on a message bus, characterized in that: The architecture of the digital simulation platform includes: a view layer, a service layer and a data access layer, wherein: The view layer is responsible for front-end display, receiving interactive input from users, configuring the corresponding training environment and methods, receiving simulation situation data, and displaying the algorithm training process in real time; it can interact with the service layer for data; The data access layer includes algorithm modules and equipment models, which are used to generate and store simulation data and situation information, and provide the stored data to the service layer; The service layer is used for data transmission, interface management and calculation, including message bus and service nodes; among them: the message bus performs high-concurrency data communication and transmission with the view layer and data access layer, realizes fast interface correspondence and protocol field conversion between algorithm modules, between algorithm modules and simulation environment, and between algorithm modules and equipment models, ensures clock consistency and accurate status during transmission, and provides adaptive interface management; the service node obtains various data through the message bus and provides various computing platforms and simulation environments.
2. The digital simulation platform based on the message bus according to claim 1, characterized in that: Service nodes include: simulation engine service node, equipment computing service node, AI algorithm service node, database service node, simulation scheduling and control service node, specifically: Simulation engine service node: provides simulation deduction time system services, integrates and schedules equipment services, AI algorithm services, and database services, distributes situation information to equipment computing service nodes on demand, and performs simulation deduction global decision calculations; Equipment computing service node: responsible for equipment simulation advancement calculation, status update, equipment detection, communication capability, damage capability calculation, and management of equipment situation package sending and receiving; AI algorithm service node: responsible for starting the algorithm module, accessing the algorithm corresponding simulation environment, and responding to AI instructions; Database service node: responsible for simulation data management, scenario data management and equipment model data management, providing data support for simulation environment construction; Simulation management and service node: cooperates with the visual layer to establish a complex simulation service environment for users and perform process management and control of multi-scenario simulation.
3. The digital simulation platform based on the message bus according to claim 2, characterized in that: When each node runs in the architecture: Each computing node establishes a channel connection with other computing nodes through a centralized control processing high-concurrency message bus, and detects the software and hardware changes of the corresponding computing nodes in real time.
4. The digital simulation platform based on the message bus according to claim 1, characterized in that: The message bus includes a centralized control and processing high-concurrency message bus and an adaptive interface management module, where: The centralized control processing high-concurrency message bus is responsible for high-concurrency data communication and transmission with the view layer and the data access layer, realizing fast interface correspondence and protocol field conversion between algorithm modules, between algorithm modules and simulation environment, and between algorithm modules and equipment models, ensuring clock consistency and accurate status during transmission; The adaptive interface management module is responsible for the adaptive interface management during data transmission, including: simulation platform and algorithm interface management unit, heterogeneous interface access management unit, and multi-granularity model adaptive access management unit.
5. The digital simulation platform based on the message bus according to claim 4 is characterized in that: The centralized control processing high-concurrency message bus maps different protocols into a unified interaction framework; encapsulates the algorithm interface according to the corresponding protocol, and automatically schedules and executes the appropriate algorithm to process the received requests; Through the inter-protocol conversion algorithm, different protocols are converted into a format that the framework can handle, and the interface access protocol is used for asynchronous transmission.
6. The digital simulation platform based on the message bus according to claim 4, characterized in that: The simulation platform and algorithm interface management unit encapsulates the data interacting between each algorithm module and the simulation platform, and builds a standardized data interaction interface; it builds an algorithm test scheduling management mechanism to enable each algorithm module and the simulation platform to interact based on RPC protocol network communication.
7. The digital simulation platform based on message bus according to claim 5, characterized in that: The heterogeneous interface access management unit comprehensively uses RESTful-based Web service integration technology, RPC remote call protocol and publish-subscribe-based bus distributed communication method to complete inter-tool integration for the transmission and call of different static data and dynamic data; among them, static data includes various simulation data, and dynamic data includes algorithms, situation information, and equipment models.
8. The digital simulation platform based on message bus according to claim 5, characterized in that: The multi-granularity model adaptive access management unit connects each equipment model to the centralized control and processing high-concurrency message bus through a distributed loosely coupled asynchronous communication plug-in method. Each access equipment model communicates through a publish-subscribe bus distributed communication method.
9. The digital simulation platform based on message bus according to claim 8, characterized in that: When communicating via publish-subscribe bus distributed communication: The equipment model generates situation information, and the simulation engine service node sends it to the centralized control processing high-concurrency message bus; After receiving the message in the centralized control processing high-concurrency message bus, the equipment computing service node updates the situation information of the equipment model and sends the updated situation information to the centralized control processing high-concurrency message bus; The simulation engine service node receives the updated situation information of each equipment model and updates the situation information synchronously.