A middleware data forwarding method and system for a ship control system

Abstract

This invention provides a middleware data forwarding method and system for a ship control system, belonging to the field of communication technology. The method includes: acquiring a multi-source heterogeneous dataset in the operating environment of the ship control system; inputting the multi-source heterogeneous dataset into a generative large model to obtain a coarse classification result, the coarse classification result indicating the type of forwarding sub-task to be performed for each original message data; the generative large model, based on the coarse classification result, dynamically calling at least one specialized small model corresponding to the type of forwarding sub-task, and sending the original message data to the called target small model, while simultaneously performing fine-grained transformation processing on the original message data based on the called target small model to generate intermediate transformed data that conforms to the target protocol and target format; and forwarding the intermediate transformed data as the final output data to the data user end that made the data request through the generative large model, completing the data forwarding and achieving accurate matching between the output data and the user end requirements.

Description

Technical Field

[0001] This invention relates to the field of network communication technology, and in particular to a middleware data forwarding method and system for a ship control system. Background Technology

[0002] Ship control systems are used for the operation, control, and management of ship-related equipment. Existing ship control systems face significant flexibility and adaptability issues when performing data conversion. Specifically, ship control systems manage multiple sub-application systems from different manufacturers and sources. These systems have significantly different data formats and communication protocols, leading to complex and cumbersome data interaction processes. Traditional solutions typically involve pre-writing point-to-point data conversion algorithms. This approach works when the system configuration is fixed, but becomes rigid when the ship control system needs flexible reconfiguration, equipment additions, removals, or updates, requiring substantial manpower for modification and adjustment. While some advanced control systems employ traditional middleware technology, their flexibility remains insufficient, requiring strict pre-definition of data formats and conversion rules. This makes them unable to effectively address the data conversion needs arising from dynamic system changes, resulting in high maintenance costs and low efficiency. Summary of the Invention

[0003] This invention provides a middleware data forwarding method and system for ship control systems to overcome the shortcomings of existing technologies. By introducing generative large model technology into middleware data conversion, it can complete complex conversions of different data types between multiple application systems and has good scalability and flexibility.

[0004] In a first aspect, the present invention provides a middleware data forwarding method for a ship control system, comprising:

[0005] Acquire multi-source heterogeneous datasets from the ship control system operating environment; the multi-source heterogeneous datasets contain raw message data of different protocols, formats and lengths output by different sub-application systems;

[0006] The multi-source heterogeneous dataset is input into the generative large model to obtain a coarse classification result. The coarse classification result indicates the type of forwarding subtask to be performed for each original message data. Based on the coarse classification result, the generative large model dynamically calls at least one specialized small model corresponding to the type of forwarding subtask and sends the original message data to the called target small model.

[0007] Based on the invoked target mini-model, the original message data is subjected to fine-grained transformation processing to generate intermediate transformed data that conforms to the target protocol and target format;

[0008] The intermediate transformed data is used as the final output data by the generative large model and forwarded to the data user who made the data request, thus completing the data forwarding.

[0009] According to the middleware data forwarding method of the ship control system, the generative large model is used to perform task instruction understanding and coarse data category classification on the original message data to obtain coarse classification results.

[0010] According to the middleware data forwarding method for the ship control system, the step of obtaining the multi-source heterogeneous dataset in the ship control system operating environment includes:

[0011] During the operation of the actual ship control system or bench test control system, the raw message data is collected in real time via bypass mirroring or bus monitoring; and / or

[0012] Run the task simulation software to generate the original message data in batches according to the set time sequence, message length and protocol type.

[0013] According to the middleware data forwarding method of the ship control system, the forwarding sub-task types include: data format conversion, protocol conversion, data serialization, and data security analysis; the specialized small models are correspondingly set as data format conversion small models, protocol conversion small models, data serialization small models, and data security analysis small models; the generative large model calls the specialized small models once or in cascades according to the task tags in the coarse classification results to complete multi-level conversion.

[0014] According to the middleware data forwarding method of the ship control system, the step before inputting the multi-source heterogeneous dataset into the generative large model to obtain the coarse classification result includes:

[0015] The multi-source heterogeneous dataset is randomly divided into training set, validation set and test set according to a preset ratio;

[0016] The generative large model and the specialized small model are jointly fine-tuned using the divided training data, and hyperparameters are optimized using validation data. The overall forwarding accuracy is evaluated using test data until a preset accuracy threshold is met.

[0017] According to the middleware data forwarding method of the ship control system, before forwarding the intermediate transformation data as the final output data through the generative large model to the data user who made the data request, the process includes:

[0018] Receive the intermediate conversion data and the current ship operation context information, perform consistency verification on the interrelated multi-source intermediate conversion data, and generate a credibility score;

[0019] If the credibility score is lower than the set threshold, the generative large model will prioritize selecting redundant data sources or issue an anomaly alert to the data user.

[0020] According to the middleware data forwarding method of the ship control system, the step of receiving the intermediate conversion data and the current ship operating context information, and performing consistency verification on the interrelated multi-source intermediate conversion data, includes:

[0021] The dynamic context state engine establishes a dynamic baseline for the current navigation state based on historical operational data;

[0022] The multi-source consistency checker compares multiple data values ​​related to the same physical quantity in the intermediate transformation data with the dynamic baseline. If the difference exceeds the allowable range, it is marked as an anomaly and the credibility score is reduced, thus obtaining the credibility score and anomaly label.

[0023] In a second aspect, the present invention provides a middleware data forwarding system for a ship control system, the system comprising:

[0024] The data acquisition module is used to acquire multi-source heterogeneous datasets from the operating environment of the ship control system; the multi-source heterogeneous datasets include raw message data of different protocols, formats and lengths output by different sub-application systems;

[0025] The large model processing module is used to input the multi-source heterogeneous dataset into the generative large model to obtain a coarse classification result. The coarse classification result indicates the type of forwarding sub-task to be performed for each original message data. The generative large model dynamically calls at least one specialized small model corresponding to the forwarding sub-task type based on the coarse classification result, and sends the original message data to the called target small model.

[0026] The small model processing module is used to perform fine-grained transformation processing on the original message data based on the called target small model, and generate intermediate transformed data that conforms to the target protocol and target format;

[0027] The data forwarding module is used to forward the intermediate transformed data as the final output data to the data user end that made the data request through the generative large model, thus completing the data forwarding.

[0028] Thirdly, the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the middleware data forwarding method of any of the above-described ship control systems.

[0029] Fourthly, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the middleware data forwarding method of any of the above-described ship control systems.

[0030] The middleware data forwarding method and system for ship control systems provided by this invention have the following advantages compared to the prior art:

[0031] The system acquires a multi-source heterogeneous dataset from the ship's control system operating environment. This dataset is then input into a generative large model to obtain a coarse classification result, indicating the type of forwarding subtask required for each original message data. Based on the coarse classification result, the generative large model dynamically calls at least one specialized small model corresponding to the forwarding subtask type and sends the original message data to the called target small model. Because the generative large model directly performs task understanding and coarse classification of the multi-source heterogeneous original messages and instantly calls specialized small models to complete fine-tuning, the system does not require any pre-written or maintained point-to-point conversion code.

[0032] Simultaneously, the original message data is refined and transformed based on the called target small model to generate intermediate transformed data that conforms to the target protocol and target format. The intermediate transformed data is then used as the final output data by the generative large model and forwarded to the data user end that made the data request, thus completing the data forwarding and achieving accurate matching between the output data and the user end requirements. This avoids the risk of format mismatch or protocol mismatch that is common in traditional middleware, thereby replacing multiple sets of customized conversion software with a single general middleware, significantly improving the flexibility, maintainability and operational reliability of the ship control system. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0034] Figure 1 This is a flowchart illustrating an optional middleware data forwarding method for a ship control system provided in an embodiment of the present invention.

[0035] Figure 2 This is a schematic diagram of the structure of an optional middleware data forwarding system for a ship control system provided in an embodiment of the present invention;

[0036] Figure 3 This is a schematic diagram of the structure of an optional electronic device provided in an embodiment of the present invention. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0038] It should be noted that in the description of the embodiments of the present invention, the terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. Unless otherwise expressly specified and limited, the terms "installed," "connected," and "linked" should be interpreted broadly, for example, as a fixed connection, a detachable connection, or an integral connection; a mechanical connection or an electrical connection; a direct connection or an indirect connection through an intermediate medium; or a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in the present invention according to the specific circumstances.

[0039] The terms "first," "second," etc., used in this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more.

[0040] The following is combined Figures 1-3 This invention describes a middleware data forwarding method and system for a ship control system provided in an embodiment of the invention.

[0041] Figure 1 This is a flowchart illustrating the middleware data forwarding method for a ship control system provided by the present invention, as shown below. Figure 1 As shown, including but not limited to the following steps:

[0042] S102, acquire the multi-source heterogeneous dataset in the ship control system operating environment; the multi-source heterogeneous dataset contains raw message data of different protocols, formats and lengths output by different sub-application systems;

[0043] S104, input the multi-source heterogeneous dataset into the generative large model to obtain a coarse classification result, the coarse classification result indicating the type of forwarding sub-task to be performed for each original message data; the generative large model, based on the coarse classification result, dynamically calls at least one specialized small model corresponding to the type of forwarding sub-task, and sends the original message data to the called target small model.

[0044] S106, Based on the called target small model, perform fine-grained transformation processing on the original message data to generate intermediate transformed data that conforms to the target protocol and target format;

[0045] S108, the intermediate transformation data is used as the final output data through the generative large model and forwarded to the data user who made the data request, thus completing the data forwarding.

[0046] The present invention discloses a middleware data forwarding method and system for a ship control system, the steps of which can be deployed and run in a ship bench test environment or a real ship network.

[0047] In shipboard integrated bridge systems, such as IBS, the first step is to acquire a multi-source heterogeneous dataset from the ship's control system operating environment. This multi-source heterogeneous dataset contains raw message data of different protocols, formats, and lengths output by different sub-application systems.

[0048] In an optional embodiment, acquiring the multi-source heterogeneous dataset in the ship control system operating environment includes:

[0049] During the operation of the actual ship control system or bench test control system, the original message data is collected in real time through bypass mirroring or bus monitoring; and / or the task simulation software is run to generate the original message data in batches according to the set time sequence, message length and protocol type.

[0050] In this embodiment, NMEA2000 messages from the radar subsystem, CAN2.0B frames from the host control unit, and custom TCP / IP binary streams from the ship security system can be collected through the mirror port to form a multi-source heterogeneous dataset.

[0051] The multi-source heterogeneous dataset is then fed into a pre-trained generative large model. This model performs task instruction understanding and coarse data category classification on the original message data, yielding a coarse classification result. The coarse classification result indicates the type of forwarding sub-task to be performed for each original message. These sub-task types include at least data format conversion, protocol conversion, data serialization, and data security analysis. The specialized small models are correspondingly set as data format conversion, protocol conversion, data serialization, and data security analysis small models. Based on the task labels in the coarse classification result, the generative large model calls the specialized small models individually or in cascading fashion to complete multi-level conversions.

[0052] Optionally, the generative large model and the specialized small model are deployed in the same high-performance GPU computing environment. The generative large model can be a DeepSeek-14B version. The specific configuration of the high-performance GPU computing environment can be 12 NVIDIA GTX4090 graphics cards, connected via NV Link to form a two-way 6-card topology, with a total single-precision computing power of 82 TFLOPS, meeting the DeepSeek-14B requirement of inference latency <200 ms at FP16 precision. The specialized small model can use a distilled 0.3 B parameter lightweight network, registered to the large model scheduler via the gRPC interface, to achieve millisecond-level dynamic invocation. It is worth noting that the specialized small model is a lightweight neural network model with a smaller parameter size than the generative large model, and the generative large model achieves unified scheduling of the specialized small model through an API interface.

[0053] Then, the generative large model performs task instruction understanding and coarse data category classification on the three types of messages without relying on any preset protocol description, and outputs coarse classification results, such as marking NMEA2000 messages as "navigation-speed-format conversion task" and CAN frames as "propulsion-torque-protocol conversion task".

[0054] Next, the large model dynamically calls the corresponding specialized small models based on the task tags. For example, the navigation speed conversion small model completes the NMEA2000→JSON format conversion, and the propulsion torque conversion small model completes the CAN→Modbus-TCP protocol conversion, generates intermediate conversion data, and returns it to the large model.

[0055] Finally, the generative large model forwards the JSON and Modbus-TCP data to the electronic chart display station and the energy consumption monitoring station, respectively. The entire process does not require modification of the configuration files of any sub-application system or user terminal, thus solving the problem of rewriting code and long debugging cycles when adding new radars or changing host models in traditional point-to-point conversion logic.

[0056] In an optional embodiment of the present invention, a middleware data forwarding method for an optional ship control system is provided, wherein before inputting the multi-source heterogeneous dataset into a generative large model to obtain a coarse classification result, the method includes:

[0057] The multi-source heterogeneous dataset is randomly divided into training set, validation set and test set according to a preset ratio;

[0058] The generative large model and the specialized small model are jointly fine-tuned using the divided training data, and hyperparameters are optimized using validation data. The overall forwarding accuracy is evaluated using test data until a preset accuracy threshold is met.

[0059] In this embodiment, the acquired multi-source heterogeneous dataset is divided into training, validation, and test sets according to a certain ratio. One option is to use 70% of the data for training, 20% for validation, and 10% for testing. The multi-source heterogeneous dataset is randomly shuffled according to the ratio given in the first document and then divided as follows: 70% for joint fine-tuning, 20% for early stopping validation, and 10% for final accuracy testing.

[0060] In an optional embodiment of the present invention, an optional middleware data forwarding method for a ship control system is provided, wherein before forwarding the intermediate transformed data as the final output data through the generative large model to the data user terminal that made the data request, the method includes:

[0061] Receive the intermediate conversion data and the current ship operation context information, perform consistency verification on the interrelated multi-source intermediate conversion data, and generate a credibility score;

[0062] If the credibility score is lower than the set threshold, the generative large model will prioritize selecting redundant data sources or issue an anomaly alert to the data user.

[0063] Furthermore, the step of receiving the intermediate conversion data and the current ship operation context information, and performing consistency verification on the interrelated multi-source intermediate conversion data, includes:

[0064] The dynamic context state engine establishes a dynamic baseline for the current navigation state based on historical operational data;

[0065] The multi-source consistency checker compares multiple data values ​​related to the same physical quantity in the intermediate transformation data with the dynamic baseline. If the difference exceeds the allowable range, it is marked as an anomaly and the credibility score is reduced, thus obtaining the credibility score and anomaly label.

[0066] In this embodiment, during actual ship navigation, a context-aware data credibility module is added as a new specialized small model that is called by the large model. For example, when the GPS sends out 15 kn SOG, the Doppler log sends out 5 kn, and the main engine speed sends out 100% of the rated speed, the dynamic context state engine in CDTM establishes a "full speed in open water" baseline based on the historical data of the past 30 minutes. It determines that the difference between the three speed sources exceeds the ±2 kn threshold and outputs a credibility score of 0.42, which is lower than the 0.5 threshold. Based on this, the large model refuses to forward abnormal speed data and pops up a "speed inconsistency" alarm on the bridge display. At the same time, it automatically selects redundant GPS data to continue forwarding, avoiding the crew from making wrong decisions based on incorrect speeds.

[0067] In this embodiment, the multi-source consistency verifier inside the context-aware data credibility module can perform real-time comparison of data such as "rudder angle-heading": if the autopilot command is 10° right rudder, but the rudder angle feedback sensor continuously reports 0° and the heading remains constant, and the difference persists for 5 seconds, the verifier marks the rudder angle data as abnormal, the credibility score drops to 0.3, the generative large model immediately pushes the "rudder angle abnormality" alarm to the IBS alarm bus, and records an abnormal log with a timestamp for post-fault analysis, realizing the "data security analysis" function proposed in the first document.

[0068] In summary, the middleware data forwarding method for a ship control system provided by this invention has the following technical advantages:

[0069] Because this invention directly performs task understanding and coarse classification of multi-source heterogeneous original messages through a generative large model, and instantly calls professional small models to complete fine conversion, the system does not need to pre-write or maintain any point-to-point conversion code. When the protocol and format change due to modification, upgrade or addition of equipment on board, only the new message needs to be connected at the input end. The large model automatically identifies and schedules the corresponding small model, realizing "zero configuration" expansion.

[0070] Meanwhile, because all conversion logic is uniformly orchestrated by the large model, the conversion process is centralized and transparent, the debugging and fault location time is greatly shortened, and the operation and maintenance costs are reduced accordingly.

[0071] Furthermore, the high-performance GPU computing environment ensures that inference between large and small models is completed in milliseconds, meeting the real-time requirements of ship control. The output data is also precisely matched with the user's needs, avoiding the risks of format mismatch or protocol mismatch common in traditional middleware. This allows a single general-purpose middleware to replace multiple sets of customized conversion software, significantly improving the flexibility, maintainability, and operational reliability of the ship control system.

[0072] Figure 2 This is a schematic diagram of the structure of an optional middleware data forwarding system for a ship control system provided in an embodiment of the present invention; as shown. Figure 2 As shown, the present invention also provides a middleware data forwarding system for a ship control system, the system comprising:

[0073] The data acquisition module is used to acquire multi-source heterogeneous datasets from the operating environment of the ship control system; the multi-source heterogeneous datasets include raw message data of different protocols, formats and lengths output by different sub-application systems;

[0074] The large model processing module is used to input the multi-source heterogeneous dataset into the generative large model to obtain a coarse classification result. The coarse classification result indicates the type of forwarding sub-task to be performed for each original message data. The generative large model dynamically calls at least one specialized small model corresponding to the forwarding sub-task type based on the coarse classification result, and sends the original message data to the called target small model.

[0075] The small model processing module is used to perform fine-grained transformation processing on the original message data based on the called target small model, and generate intermediate transformed data that conforms to the target protocol and target format;

[0076] The data forwarding module is used to forward the intermediate transformed data as the final output data to the data user end that made the data request through the generative large model, thus completing the data forwarding.

[0077] It should be noted that the middleware data forwarding system for the ship control system provided in this embodiment of the invention can execute the middleware data forwarding method for the ship control system described in any of the above embodiments during actual operation, which will not be elaborated in this embodiment.

[0078] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention, such as... Figure 3 As shown, the electronic device may include: a processor 310, a communications interface 320, a memory 330, and a communication bus 340, wherein the processor 310, the communications interface 320, and the memory 330 communicate with each other through the communication bus 340. The processor 310 can call logical instructions in the memory 330 to execute a middleware data forwarding method for the ship control system. This method includes: S1, acquiring a multi-source heterogeneous dataset from the ship control system's operating environment; the multi-source heterogeneous dataset contains raw message data of different protocols, formats, and lengths output by different sub-application systems.

[0079] S2, input the multi-source heterogeneous dataset into the generative large model to obtain a coarse classification result, the coarse classification result indicating the type of forwarding sub-task to be performed for each original message data; the generative large model, based on the coarse classification result, dynamically calls at least one specialized small model corresponding to the type of forwarding sub-task, and sends the original message data to the called target small model.

[0080] S3, perform fine-grained transformation processing on the original message data based on the called target small model to generate intermediate transformed data that conforms to the target protocol and target format;

[0081] S4. The intermediate transformation data is used as the final output data through the generative large model and forwarded to the data user who made the data request, thus completing the data forwarding.

[0082] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0083] On the other hand, the present invention also provides a computer program product, the computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions, wherein when the program instructions are executed by a computer, the computer is able to execute the middleware data forwarding method of the ship control system provided in the above embodiments, the method comprising:

[0084] S1, acquire a multi-source heterogeneous dataset from the ship control system operating environment; the multi-source heterogeneous dataset contains raw message data of different protocols, formats and lengths output by different sub-application systems;

[0085] S2, input the multi-source heterogeneous dataset into the generative large model to obtain a coarse classification result, the coarse classification result indicating the type of forwarding sub-task to be performed for each original message data; the generative large model, based on the coarse classification result, dynamically calls at least one specialized small model corresponding to the type of forwarding sub-task, and sends the original message data to the called target small model.

[0086] S3, perform fine-grained transformation processing on the original message data based on the called target small model to generate intermediate transformed data that conforms to the target protocol and target format;

[0087] S4. The intermediate transformation data is used as the final output data through the generative large model and forwarded to the data user who made the data request, thus completing the data forwarding.

[0088] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the middleware data forwarding method of the ship control system provided in the above embodiments, the method comprising:

[0089] S1, acquire a multi-source heterogeneous dataset from the ship control system operating environment; the multi-source heterogeneous dataset contains raw message data of different protocols, formats and lengths output by different sub-application systems;

[0090] S2, input the multi-source heterogeneous dataset into the generative large model to obtain a coarse classification result, the coarse classification result indicating the type of forwarding sub-task to be performed for each original message data; the generative large model, based on the coarse classification result, dynamically calls at least one specialized small model corresponding to the type of forwarding sub-task, and sends the original message data to the called target small model.

[0091] S3, perform fine-grained transformation processing on the original message data based on the called target small model to generate intermediate transformed data that conforms to the target protocol and target format;

[0092] S4. The intermediate transformation data is used as the final output data through the generative large model and forwarded to the data user who made the data request, thus completing the data forwarding.

[0093] The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0094] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0095] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method of middleware data forwarding for a ship control system, characterized by, The method includes: Acquire multi-source heterogeneous datasets from the ship control system operating environment; the multi-source heterogeneous datasets contain raw message data of different protocols, formats and lengths output by different sub-application systems; The multi-source heterogeneous dataset is input into the generative large model to obtain a coarse classification result. The coarse classification result indicates the type of forwarding subtask to be performed for each original message data. Based on the coarse classification result, the generative large model dynamically calls at least one specialized small model corresponding to the type of forwarding subtask and sends the original message data to the called target small model. Based on the invoked target mini-model, the original message data is subjected to fine-grained transformation processing to generate intermediate transformed data that conforms to the target protocol and target format; The intermediate transformed data is used as the final output data by the generative large model and forwarded to the data user who made the data request, thus completing the data forwarding.

2. The middleware data forwarding method of a ship control system according to claim 1, characterized in that, The generative large model is used to understand the task instructions and perform coarse classification of data categories on the original message data to obtain coarse classification results.

3. The middleware data forwarding method for a ship control system according to claim 1, characterized in that, The acquisition of multi-source heterogeneous datasets in the ship control system operating environment includes: During the operation of the actual ship control system or bench test control system, the raw message data is collected in real time via bypass mirroring or bus monitoring; and / or Run the task simulation software to generate the original message data in batches according to the set time sequence, message length and protocol type.

4. The middleware data forwarding method for a ship control system according to any one of claims 1-3, characterized in that, The forwarding subtask types include: data format conversion, protocol conversion, data serialization, and data security analysis; the specialized mini-models are correspondingly set as data format conversion mini-model, protocol conversion mini-model, data serialization mini-model, and data security analysis mini-model; the generative large model calls the specialized mini-models once or in cascades based on the task tags in the coarse classification results to complete multi-level conversion.

5. The middleware data forwarding method for a ship control system according to claim 1, characterized in that, Before inputting the multi-source heterogeneous dataset into the generative large model to obtain the coarse classification result, the method further includes: The multi-source heterogeneous dataset is randomly divided into training set, validation set and test set according to a preset ratio; The generative large model and the specialized small model are jointly fine-tuned using the divided training data, and hyperparameters are optimized using validation data. The overall forwarding accuracy is evaluated using test data until a preset accuracy threshold is met.

6. The middleware data forwarding method for a ship control system according to claim 1, characterized in that, Before forwarding the intermediate transformed data as the final output data to the data user who requested the data through the generative large model, the process further includes: Receive the intermediate conversion data and the current ship operation context information, perform consistency verification on the interrelated multi-source intermediate conversion data, and generate a credibility score; If the credibility score is lower than the set threshold, the generative large model will prioritize selecting redundant data sources or issue an anomaly alert to the data user.

7. The middleware data forwarding method for a ship control system according to claim 6, characterized in that, The process of receiving the intermediate conversion data and the current ship operation context information, and performing consistency verification on the interrelated multi-source intermediate conversion data, includes: The dynamic context state engine establishes a dynamic baseline for the current navigation state based on historical operational data; The multi-source consistency checker compares multiple data values ​​related to the same physical quantity in the intermediate transformation data with the dynamic baseline. If the difference exceeds the allowable range, it is marked as an anomaly and the credibility score is reduced, thus obtaining the credibility score and anomaly label.

8. A middleware data forwarding system for a ship control system, characterized in that, The system includes: The data acquisition module is used to acquire multi-source heterogeneous datasets from the operating environment of the ship control system; the multi-source heterogeneous datasets include raw message data of different protocols, formats and lengths output by different sub-application systems; The large model processing module is used to input the multi-source heterogeneous dataset into the generative large model to obtain a coarse classification result. The coarse classification result indicates the type of forwarding sub-task to be performed for each original message data. The generative large model dynamically calls at least one specialized small model corresponding to the forwarding sub-task type based on the coarse classification result, and sends the original message data to the called target small model. The small model processing module is used to perform fine-grained transformation processing on the original message data based on the called target small model, and generate intermediate transformed data that conforms to the target protocol and target format; The data forwarding module is used to forward the intermediate transformed data as the final output data to the data user end that made the data request through the generative large model, thus completing the data forwarding.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the middleware data forwarding method for the ship control system as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the middleware data forwarding method of the ship control system as described in any one of claims 1 to 7.

Images