Unmanned ship automatic control system and method based on plug-in architecture

Abstract

The invention discloses an unmanned ship automatic control system based on a plug-in architecture, and belongs to the technical field of unmanned ship control. The system comprises a front-end display layer, an API gateway layer, an AI decision-making layer, a business service layer, a data management layer, a hardware abstraction layer and an edge computing layer. The front-end display layer is used for receiving a task request to generate a structured control request, and the API gateway layer is used for receiving the structured control request for verification and combination to generate a unified request object; the AI decision-making layer is used for receiving the unified request object and the environment perception data and generating a decision-making scheme based on a multi-model fusion decision-making; the business service layer is used for receiving the decision scheme to generate a task execution plan; the data management layer is used for receiving task execution plan generation state enhancement data; the hardware abstraction layer is used for receiving the state enhancement data for mapping and generating an equipment control instruction; and the edge calculation layer is used for receiving the equipment control instruction and generating an execution result in combination with the sensor real-time stream. The system improves the automation level of the unmanned ship system.

Description

Technical Field

[0001] This application belongs to the field of unmanned vessel control technology, and in particular relates to an unmanned vessel automatic control system and method based on a plug-in architecture. Background Technology

[0002] With the development of technologies such as artificial intelligence, big data, and edge computing, unmanned surface vessels (USVs) are increasingly being used in scientific research and surveying, water patrol, environmental management, and emergency rescue. The core of an USV system is its control system, which must simultaneously possess capabilities such as navigation control, mission management, data communication, and safety assurance to ensure the stable operation of the USV in complex aquatic environments. However, most existing USV control systems adopt closed or tightly coupled architectures, with strong dependencies between functional modules, making it difficult to achieve rapid expansion and flexible configuration in different application scenarios. When adding new functions or adapting to different vessel types is required, significant modifications to the core code are often necessary, resulting in high maintenance and upgrade costs, and insufficient compatibility and scalability.

[0003] Furthermore, existing systems have significant shortcomings in terms of automation level and operational mechanisms. Unmanned surface vessels (USVs) typically rely on pre-set paths or human intervention to execute tasks, lacking global situational awareness and autonomous decision-making capabilities, making it difficult to meet the demands of complex operating environments. Simultaneously, the lack of a unified communication mechanism between modules results in insufficient flexibility in functional expansion and information exchange; inconsistent data models affect interoperability; and limited task scheduling and resource management capabilities easily lead to delays or task conflicts, reducing operational efficiency. More importantly, existing systems remain inadequate in communication security, navigation boundary violation protection, and emergency response mechanisms, lacking multi-level response strategies, making it difficult to ensure the safety and stability of USVs in emergency situations.

[0004] Most existing unmanned vessel control systems adopt a closed or tightly coupled architecture, with strong dependencies between functional modules, making it difficult to flexibly expand across different mission scenarios and vessel types, resulting in high system upgrade and maintenance costs. Summary of the Invention

[0005] This application aims to address at least one of the technical problems existing in the prior art. To this end, this application proposes an unmanned surface vessel (USV) automatic control system based on a plug-in architecture, which improves the automation level of the USV system.

[0006] In the first aspect, this application provides an unmanned vessel automatic control system based on a plug-in architecture, the system including a front-end presentation layer, an API gateway layer, an AI decision-making layer, a business service layer, a data management layer, a hardware abstraction layer, and an edge computing layer; The front-end presentation layer is used to receive task requests and generate structured control requests. The task requests include operation instructions, sensor signals, task planning requests, or environmental information. The API gateway layer is used to receive structured control requests, verify and merge them, and generate a unified request object after identity authentication. The AI ​​decision-making layer is used to receive unified request objects and environmental perception data, and generate decision schemes based on multi-model fusion decision-making. The business service layer is used to receive decision-making schemes, refine the decision-making schemes into executable tasks based on the current system status, and perform scheduling and plug-in assembly to generate task execution plans. The data management layer is used to receive task execution plans, process them in combination with historical execution data and real-time monitoring data, and generate status enhancement data. The hardware abstraction layer is used to receive state enhancement data, map it, and generate device control commands. The edge computing layer is used to receive device control commands and generate execution results by combining real-time sensor streams. The front-end presentation layer, API gateway layer, AI decision layer, business service layer, data management layer, hardware abstraction layer, and edge computing layer are independent of each other and are decoupled through an event-driven communication bus.

[0007] According to one embodiment of this application, the front-end presentation layer includes a command parsing submodule, a multimodal fusion submodule, and a security verification submodule; The command parsing submodule is used to receive and parse task requests, and generate multimodal instructions; The multimodal fusion submodule is used to fuse multimodal instructions to generate structured instructions; The security verification submodule is used to perform permission verification and legality verification on structured instructions. If the permission verification and legality verification pass, a structured control request is generated.

[0008] According to one embodiment of this application, the API gateway layer includes an authentication and authorization submodule, a request aggregation submodule, and a security audit submodule; The authentication and authorization submodule is used to receive structured control requests, perform identity authentication and permission verification, and generate an authentication context; The request aggregation submodule is used to parse the authentication context and match rules to generate an aggregated request. The security audit submodule is used to collect and format logs of aggregated requests to generate a unified request object.

[0009] According to one embodiment of this application, the AI ​​decision-making layer includes a data fusion submodule, a strategy reasoning submodule, and a risk assessment submodule; The data fusion submodule is used to receive unified request objects and environmental perception data, perform fusion based on multiple models, and generate a fusion context. The policy reasoning submodule is used to perform policy reasoning on the fusion context and generate policy proposals; The risk assessment submodule is used to assess the risks of the proposed strategies and generate a decision-making plan.

[0010] According to one embodiment of this application, the business service layer includes a task scheduling submodule, a plug-in management submodule, and a task monitoring submodule; The task scheduling submodule is used to receive decision schemes and current system status, perform task decomposition and resource scheduling, and generate task scheduling plans. The plugin management submodule is used to receive task scheduling plans, perform plugin matching and assembly, and generate plugin assembly schemes. The task monitoring submodule is used to receive the plug-in assembly scheme, monitor its execution status and handle fault tolerance, and generate a task execution plan.

[0011] According to one embodiment of this application, the data management layer includes a data acquisition submodule, a data cleaning submodule, and a state fusion submodule; The data acquisition submodule is used to receive the task execution plan, perform multi-source data acquisition, and generate raw environmental data. The data cleaning submodule is used to clean and preprocess the raw environmental data to generate standard format data; The state fusion submodule is used to fuse standard format data, historical execution data, and real-time monitoring data to generate state-enhanced data.

[0012] According to one embodiment of this application, the hardware abstraction layer includes a protocol conversion submodule, a device monitoring submodule, and a security rollback submodule; The protocol conversion submodule is used to receive state enhancement data, perform protocol mapping and instruction conversion, and generate standard device instructions; The equipment monitoring submodule is used to perform equipment status verification and resource checks on standard equipment instructions, and generate executable equipment instructions. The security rollback submodule is used to perform security boundary verification and anomaly contingency matching on executable device instructions, and generate device control instructions.

[0013] According to one embodiment of this application, the edge computing layer includes a local decision-making submodule, a real-time control submodule, and an anomaly detection submodule; The local decision-making submodule is used to receive device control commands, perform real-time decision optimization, and generate optimized control commands. The real-time control submodule is used to perform signal conversion and timing control on the optimized control commands, and generate drive signals; The anomaly detection submodule is used to detect anomalies and provide status feedback on the drive signals and real-time sensor streams, and generate execution results.

[0014] According to one embodiment of this application, the unmanned vessel automatic control system based on a plug-in architecture supports a plug-in mechanism, allowing each sub-module of the front-end presentation layer, API gateway layer, AI decision layer, business service layer, data management layer, hardware abstraction layer and edge computing layer to be dynamically loaded or unloaded as an independent plug-in, realizing hot-swappable functionality without restarting the system.

[0015] Secondly, this application provides an automatic control method for unmanned vessels based on a plug-in architecture, the method comprising: Obtain the task request, send the task request to the front-end presentation layer, and generate a structured control request; The structured control request is sent to the API gateway layer for verification and merging to generate a unified request object after identity authentication. The unified request object is sent to the AI ​​decision-making layer, and a decision scheme is generated based on multi-model fusion decision-making in combination with environmental perception data. The decision-making plan is sent to the business service layer, which then refines the plan into executable tasks based on the current system status, schedules and assembles plugins to generate a task execution plan. The task execution plan is sent to the data management layer, where it is processed in conjunction with historical execution data and real-time monitoring data to generate status enhancement data. The state augmentation data is sent to the hardware abstraction layer for mapping, generating device control commands. Device control commands are sent to the edge computing layer, and execution results are generated by combining real-time sensor streams.

[0016] Thirdly, this application 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 computer program to implement the unmanned vessel automatic control method based on a plug-in architecture as described in the second aspect above.

[0017] Fourthly, this application provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the unmanned vessel automatic control method based on a plug-in architecture as described in the second aspect above.

[0018] Fifthly, this application provides a chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run programs or instructions to implement the unmanned vessel automatic control method based on a plug-in architecture as described in the second aspect.

[0019] Sixthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the unmanned vessel automatic control method based on a plug-in architecture as described in the second aspect above.

[0020] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application.

[0021] The present invention provides an unmanned surface vessel automatic control system based on a plug-in architecture, which has the following advantages over the prior art: (1) This invention decouples functional modules through a layered architecture, event-driven communication mechanism and unified data model, supports flexible expansion by combining plug-in and hot-swappable mechanisms, and improves the autonomous operation capability and safety of unmanned vessels in complex water environments through automatic decision-making and edge computing.

[0022] (2) By introducing a plug-in architecture and an event-driven communication mechanism, this invention effectively reduces the coupling between modules and minimizes the possibility of fixed functions and difficulties in expansion of the unmanned vessel system. The hot-swappable plug-in approach allows for rapid customization and expansion of functions, significantly shortening the development and deployment cycle of new functions and improving the system's adaptability in different application scenarios. By setting up an AI decision-making layer and combining various algorithms such as deep learning, reinforcement learning, and expert systems, global situational awareness, route optimization, and predictive maintenance are achieved, enhancing the autonomous intelligence level of the unmanned vessel in complex water environments and reducing reliance on human intervention.

[0023] (3) By introducing a hardware abstraction layer, this invention reduces the differences in underlying hardware, enabling the system to be adapted to different models of unmanned ships and their sensors, actuators and communication modules, thereby improving the platform’s versatility and portability. Through real-time computing and AI inference engine of edge computing layer, this invention achieves millisecond-level processing of key task data, ensuring that unmanned ships can continue to operate safely under unstable communication or network outage conditions, improving the modularity, flexibility and intelligence of unmanned ship control system, improving operational efficiency and operational safety, and is applicable to various application scenarios such as patrol boats, scientific research vessels, and cleaning vessels. Attached Figure Description

[0024] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a schematic diagram of the structure of the unmanned vessel automatic control system based on a plug-in architecture provided in the embodiments of this application; Figure 2 This is a flowchart of the event-driven communication mechanism provided in the embodiments of this application; Figure 3This is a flowchart of the security mechanism provided in the embodiments of this application; Figure 4 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0025] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0026] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of 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. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0027] The following description, in conjunction with the accompanying drawings, details the unmanned vessel automatic control system based on a plug-in architecture, the unmanned vessel-oriented task priority scheduling and resource awareness management device, electronic equipment, and readable storage medium provided in this application, through specific embodiments and application scenarios.

[0028] Among them, the unmanned vessel automatic control system based on the plug-in architecture can be applied to the terminal, and can be executed by the hardware or software in the terminal.

[0029] The terminal includes, but is not limited to, portable communication devices such as mobile phones or tablets with touch-sensitive surfaces (e.g., touchscreen displays and / or touchpads). It should also be understood that, in some embodiments, the terminal may not be a portable communication device, but rather a desktop computer with touch-sensitive surfaces (e.g., touchscreen displays and / or touchpads).

[0030] The following embodiments describe a terminal including a display and a touch-sensitive surface. However, it should be understood that the terminal may include one or more other physical user interface devices such as a physical keyboard, mouse, and joystick.

[0031] The unmanned vessel automatic control system based on a plug-in architecture provided in this application embodiment can be executed by an electronic device or a functional module or entity in an electronic device that can implement the unmanned vessel automatic control system based on a plug-in architecture. The electronic devices mentioned in this application embodiment include, but are not limited to, mobile phones, tablets, computers, cameras and wearable devices. The following uses an electronic device as the execution subject to describe the unmanned vessel automatic control system based on a plug-in architecture provided in this application embodiment.

[0032] Figure 1 This is a schematic diagram of the unmanned vessel automatic control system based on a plug-in architecture provided in an embodiment of this application, as shown below. Figure 1 As shown, the unmanned surface vessel (USV) automated control system based on a plug-in architecture includes a front-end presentation layer, an API gateway layer, an AI decision-making layer, a business service layer, a data management layer, a hardware abstraction layer, and an edge computing layer. Each layer is logically relatively independent but interacts through a unified data interface and event-driven mechanism, thereby achieving decoupling and expansion of system functions.

[0033] It's easy to understand that the unmanned surface vessel (USV) automated control system adopts a distributed and hierarchical deployment model. The overall system operating environment comprises three parts: shore-based, ship-based, and cloud-based. The shore-based unit primarily serves as the central node for operator-system interaction, undertaking functions such as task assignment, remote monitoring, and data visualization. The ship-based unit, as the core execution unit, is responsible for real-time control of various sensors and actuators on the USV, and achieves localized automated decision-making and data processing through edge computing units. The cloud-based unit acts as the global scheduling and resource management center, providing large-scale data storage, model training, and strategy optimization support, while maintaining data synchronization and strategy assignment with the shore-based unit. Through the collaboration of these three components, the system can achieve full-link closed-loop control and dynamic optimization of tasks.

[0034] In terms of hardware and software operating platforms, the ship-based system is equipped with high-performance edge computing units, supporting multi-threaded parallel processing and lightweight AI model inference to achieve key functions such as trajectory planning, situational awareness, and predictive maintenance. The shore-based system typically operates on monitoring terminals with visual interfaces, integrating task configuration and status management modules. The cloud server adopts a distributed cluster architecture, capable of storing and processing operational data from multiple unmanned vessels, and providing strategy optimization and safety warnings based on big data and deep learning methods. All parts of the system are interconnected via 4G / 5G cellular communication, satellite links, or dedicated wireless networks, with encryption and redundancy mechanisms employed in the communication links to ensure secure and reliable data transmission.

[0035] In terms of architecture and operation, a plug-in design concept is adopted, with all functional modules loaded as plug-ins, and the system core retaining only a lightweight operating framework. Plug-ins can be dynamically loaded, unloaded, and upgraded during operation, supporting hot-swapping operations and enabling functional expansion or repair updates without restarting the system. Plug-ins can be enabled and disabled through configuration files or a management interface. Different tasks can quickly combine different functional plug-ins according to their needs, thereby achieving flexible customization and continuous iteration of the unmanned vessel control system. This mechanism not only reduces system upgrade and maintenance costs but also greatly improves the scalability and adaptability of the architecture.

[0036] The front-end presentation layer is used to receive task requests and generate structured control requests. The task requests include operation instructions, sensor signals, task planning requests, or environmental information. In some embodiments, the front-end presentation layer includes a command parsing submodule, a multimodal fusion submodule, and a security verification submodule; The command parsing submodule is used to receive and parse task requests, and generate multimodal instructions; The multimodal fusion submodule is used to fuse multimodal instructions to generate structured instructions; The security verification submodule is used to perform permission verification and legality verification on structured instructions. If the permission verification and legality verification pass, a structured control request is generated.

[0037] The front-end presentation layer is used to provide a human-computer interaction interface that is adapted to multiple devices. The interface includes a responsive component library, a personalized interface engine, and supports multimodal interaction methods such as touch, voice, and gesture to meet the usage needs of different roles such as captain, driver, and maintenance personnel. The front-end presentation layer, located at the top, primarily provides the human-machine interface for real-time visualization of task configuration, equipment control, and operational status. Operators can control the unmanned vessel through a graphical interface, virtual joystick, or touch components, and receive feedback from the AI ​​decision-making layer and business service layer, thereby enabling task configuration and execution monitoring.

[0038] The front-end display layer serves as the operating interface of the unmanned surface vessel's automatic control system, primarily providing users with task configuration, status monitoring, and human-machine interaction functions. In this embodiment, the front-end display layer adopts a component-based interface framework, which includes a task panel, device control panel, video monitoring window, and system status bar.

[0039] In terms of interaction methods, the front-end display layer supports multiple input modes, including touch operation, voice commands, and gesture control. Through virtual joysticks, sliders, and other controls, users can intuitively control the unmanned vessel's course and speed; the introduction of voice commands and gesture recognition technology further enhances the convenience and automation of operation.

[0040] In addition, the front-end presentation layer supports personalized configuration, allowing users to freely adjust the interface layout and display parameters according to different task requirements. For example, users can choose to display multiple camera feeds or focus on displaying battery status and device health information, thereby achieving a task-oriented customized interactive experience.

[0041] Through the above design, the front-end presentation layer can adapt to the application needs of multiple scenarios, improving the human-computer interaction experience and operational efficiency of the system.

[0042] The API gateway layer is used to receive structured control requests, verify and merge them, and generate a unified request object after identity authentication. The API gateway layer is deployed between the front-end presentation layer and the back-end services, undertaking functions such as request forwarding, authentication and authorization, access control, and load balancing. This layer shields the complexity of the back-end through a unified interface, provides a standardized access point, ensures that front-end commands can be securely and efficiently transmitted to the business service layer and AI decision-making layer, and manages and optimizes data traffic to reduce system latency caused by request conflicts.

[0043] In some embodiments, the API gateway layer includes an authentication and authorization submodule, a request aggregation submodule, and a security audit submodule; The authentication and authorization submodule is used to receive structured control requests, perform identity authentication and permission verification, and generate an authentication context; The request aggregation submodule is used to parse the authentication context and match rules to generate an aggregated request. The security audit submodule is used to collect and format logs of aggregated requests to generate a unified request object.

[0044] The API gateway layer serves as the unified service entry point for the system, responsible for request management and security protection. This layer includes: The authentication and authorization submodule supports multiple authentication methods such as digital certificates, biometrics, dynamic passwords, and hardware keys, enabling fine-grained role-based access control and ensuring system access security.

[0045] The request aggregation submodule can merge multiple independent requests into a batch request, reducing communication overhead and improving response speed through automatic caching and prefetching mechanisms.

[0046] The security audit submodule monitors API call volume, response time, and error rate in real time, and provides a reference for subsequent optimization in conjunction with the performance analysis module.

[0047] Through the above design, the API gateway layer effectively improves the system's security and communication efficiency, providing reliable interface support for subsequent business services and AI decision-making.

[0048] The AI ​​decision-making layer is used to receive unified request objects and environmental perception data, and generate decision schemes based on multi-model fusion decision-making. As the system's automated core, the AI ​​decision-making layer incorporates task decision-making models and path planning algorithms, enabling it to optimize course, assess situational awareness, and perform predictive maintenance based on real-time data. This layer models and analyzes the unmanned surface vessel's (USV) operating environment by subscribing to sensor data and status information from the event bus, and sends the generated decision commands to the business service layer for execution, thereby achieving autonomous operation and dynamic adjustment of the USV.

[0049] In some embodiments, the AI ​​decision-making layer includes a data fusion submodule, a strategy reasoning submodule, and a risk assessment submodule; The data fusion submodule is used to receive unified request objects and environmental perception data, perform fusion based on multiple models, and generate a fusion context. The policy reasoning submodule is used to perform policy reasoning on the fusion context and generate policy proposals; The risk assessment submodule is used to assess the risks of the proposed strategies and generate a decision-making plan.

[0050] The AI ​​decision-making layer is responsible for the autonomous operation and mission optimization of the unmanned vessel. This layer utilizes multi-source algorithms, including deep learning, reinforcement learning, and expert systems, combined with a unified data model and mission context information, to achieve automated decision-making in complex aquatic environments. Specifically, this layer includes: Data fusion submodule: Utilizing multi-source sensor data (including radar, cameras, inertial navigation systems, GNSS, etc.) and combining data fusion algorithms, it enables dynamic modeling of the aquatic environment and can identify obstacles, waterway status, and weather conditions in real time.

[0051] The strategy reasoning submodule, based on reinforcement learning and heuristic search algorithms, dynamically plans and corrects the trajectory of the unmanned vessel in real time, achieving the comprehensive goals of obstacle avoidance, energy saving, and optimal mission completion.

[0052] Risk assessment submodule: By modeling and analyzing the operating data of key components such as engines, propulsion systems, and batteries, machine learning methods are used to identify potential failure trends and trigger maintenance and protection measures in advance.

[0053] Expert system support: Built-in expert rule base for specific tasks (such as water quality sampling, obstacle removal, and inspection) to provide decision-making suggestions in complex or emergency situations, thereby improving the reliability of the system.

[0054] The AI ​​decision-making layer is the core of the automated operation of unmanned surface vessels (USVs), and its main functions are trajectory optimization based on real-time data, environmental perception, and predictive maintenance. In this implementation, the AI ​​decision-making layer incorporates multiple decision-making models, including deep learning models, rule engines, and hybrid optimization algorithms.

[0055] In terms of data processing, the AI ​​decision-making layer fuses and models multi-source data (such as GPS, radar, cameras, water quality sensors, etc.) from the data management layer to generate an environmental situation map, and calculates the optimal trajectory or adjusts the navigation strategy accordingly. Regarding equipment operating status, the AI ​​decision-making layer can use time series prediction models to predict the health of key components (batteries, thrusters, etc.), identify potential risks in advance, and trigger maintenance tasks.

[0056] Ultimately, the AI ​​decision-making layer sends the generated decision instructions to the business service layer, which further transforms them into executable control commands to achieve autonomous operation and dynamic optimization of the unmanned vessel.

[0057] Through the above design, the AI ​​decision-making layer can endow unmanned ships with the ability to operate autonomously, thereby reducing human intervention and improving the accuracy and stability of mission execution.

[0058] The business service layer is used to receive decision-making schemes, refine the decision-making schemes into executable tasks based on the current system status, and perform scheduling and plug-in assembly to generate task execution plans. The business service layer mainly comprises plug-in functional modules, including navigation control, device management, communication scheduling, task management, and security management. Each function exists as an independent plug-in and can be loaded or unloaded as needed for the task. As an intermediary layer connecting the AI ​​decision-making layer with the lower-level data and hardware, this layer is responsible for translating upper-level decisions into specific executable task instructions and providing feedback on the task execution status.

[0059] In this embodiment, by setting up an AI decision-making layer and combining various algorithms such as deep learning, reinforcement learning and expert systems, global situational awareness, route optimization and predictive maintenance are achieved, thereby improving the autonomous intelligence level of unmanned vessels in complex water environments and reducing reliance on human intervention.

[0060] In some embodiments, the business service layer includes a task scheduling submodule, a plug-in management submodule, and a task monitoring submodule; The task scheduling submodule is used to receive decision schemes and current system status, perform task decomposition and resource scheduling, and generate task scheduling plans. The plugin management submodule is used to receive task scheduling plans, perform plugin matching and assembly, and generate plugin assembly schemes. The task monitoring submodule is used to receive the plug-in assembly scheme, monitor its execution status and handle fault tolerance, and generate a task execution plan.

[0061] The business service layer, as the core logic layer for system functionality, is responsible for translating the control requirements of the unmanned vessel into specific business functions and providing modular interfaces to support expansion. This layer includes: Task scheduling submodule: Supports task creation, decomposition, priority setting and real-time scheduling. Combined with dynamic resource allocation mechanism, it realizes reasonable scheduling and conflict reduction in multi-task environment.

[0062] The mission monitoring submodule processes and stores sensor data, track information, and external inputs in real time and in a structured manner, supporting statistical analysis, status monitoring, and historical backtracking.

[0063] The plugin management submodule supports loading, uninstalling, upgrading, and configuring plugins, and provides a plugin marketplace interface to facilitate users in expanding and customizing functionality according to their needs.

[0064] As the intermediary layer bridging the system, the business service layer is primarily responsible for converting the output of the AI ​​decision-making layer into executable control commands and providing rich functional extensions through a plug-in mechanism. In this embodiment, the business service layer consists of multiple functional plug-ins, including a navigation service plug-in, a device control plug-in, a communication plug-in, a data acquisition plug-in, a task management plug-in, and a security management plug-in.

[0065] The navigation service plugin receives trajectory optimization results generated by the AI ​​decision-making layer and translates them into specific heading and speed control commands, enabling the unmanned vessel to navigate safely along a predetermined path. The equipment control plugin provides fine-grained management of hardware such as thrusters, servos, cameras, and robotic arms, and can control equipment from different manufacturers through standard interfaces. The communication plugin handles multi-channel communication scheduling, supports automatic switching between 4G / 5G, satellite links, and dedicated wireless networks, and features packet loss retransmission and bandwidth optimization capabilities.

[0066] The data acquisition plugin manages raw data from sensors (such as radar, water quality sensors, and cameras) and pushes the data to the data management layer for further processing. The task management plugin provides task decomposition, progress tracking, and status feedback, allowing operators to visually view task execution status on the front-end display layer. The security management plugin invokes the security mechanism center to implement multi-level security responses, including risk detection, tiered assessment, and emergency response triggering.

[0067] Through a plug-in design, each functional module of the business service layer supports dynamic loading and unloading, allowing users to flexibly configure plug-in combinations according to specific task scenarios. For example, a water quality acquisition and data transmission plug-in can be enabled in a water quality monitoring task, while a video surveillance and target recognition plug-in can be enabled in a patrol task. This mechanism not only ensures the high scalability of the business service layer but also reduces the complexity of system maintenance and function upgrades.

[0068] Through the above design, the business service layer not only ensures the functional integrity of the unmanned vessel during operation, but also provides a technical foundation for the modular expansion and functional upgrade of the system. Based on an event-driven bus, message passing between modules within the unmanned vessel is realized, and multi-vessel collaborative operations are supported, enabling distributed task execution and information sharing.

[0069] The data management layer is used to receive task execution plans, process them in combination with historical execution data and real-time monitoring data, and generate status enhancement data. The data management layer is used to collect, process, and store data during system operation. Internally, it includes a data flow module, a log management module, and a privacy protection module. It can clean, standardize, and persistently store multi-source data from sensors and task execution, while providing a unified data interface for the AI ​​decision-making layer. Through end-to-end data tracking, this layer also supports subsequent analysis and tracing needs.

[0070] In some embodiments, the data management layer includes a data acquisition submodule, a data cleaning submodule, and a state fusion submodule; The data acquisition submodule is used to receive the task execution plan, perform multi-source data acquisition, and generate raw environmental data. The data cleaning submodule is used to clean and preprocess the raw environmental data to generate standard format data; The state fusion submodule is used to fuse standard format data, historical execution data, and real-time monitoring data to generate state-enhanced data.

[0071] As the system's data hub, the data management layer is responsible for collecting, storing, processing, and sharing multi-source data during the unmanned surface vessel's operation, ensuring data consistency, integrity, and traceability. The data management layer establishes cross-module data standards and unified formats through a unified data model, covering sensor data, control commands, trajectory information, and mission parameters, thereby ensuring consistency and interoperability in data interaction between different modules.

[0072] In addition, the data management layer uses a combination of distributed databases and time-series databases to achieve efficient storage of large-scale sensor stream data and task logs, and supports fast retrieval and backtracking.

[0073] An efficient synchronization mechanism is established between the shipboard terminal and the remote server, and data availability is improved in weak network or offline environments through local caching. Encrypted storage, access control, and anti-tampering mechanisms ensure the security and confidentiality of unmanned vessel operational data.

[0074] As the core of data flow and storage in the system, the data management layer's main function is to process, store, and manage the security of multi-source data, providing reliable data support for the AI ​​decision-making layer and business service layer.

[0075] Under the above mechanism, the data management layer not only achieves efficient management of multi-source heterogeneous data, but also provides accurate and complete data support for upper-level AI decision-making and business services, thereby improving the stability and reliability of the unmanned vessel automatic control system.

[0076] The hardware abstraction layer is used to receive state enhancement data, map it, and generate device control commands. The Hardware Abstraction Layer (HAL) sits between the software and physical devices, shielding them from differences. Internally, it contains standardized hardware interfaces and drivers, enabling unified management of various devices such as propulsion systems, servos, cameras, radar, and water quality sensors. This layer not only provides real-time monitoring of hardware operating status but also supports firmware upgrades and secure boot mechanisms, ensuring stable system operation in complex environments.

[0077] In some embodiments, the hardware abstraction layer includes a protocol conversion submodule, a device monitoring submodule, and a security rollback submodule; The protocol conversion submodule is used to receive state enhancement data, perform protocol mapping and instruction conversion, and generate standard device instructions; The equipment monitoring submodule is used to perform equipment status verification and resource checks on standard equipment instructions, and generate executable equipment instructions. The security rollback submodule is used to perform security boundary verification and anomaly contingency matching on executable device instructions, and generate device control instructions.

[0078] The Hardware Abstraction Layer (HAL) shields the underlying hardware differences, providing a unified interface to support the access of different models of unmanned surface vessels (USVs) and their peripherals. The HAL provides standardized drivers for various sensors (such as radar, water quality monitors, and cameras), actuators (such as thrusters and robotic arms), and communication modules, simplifying upper-layer calls. Through a unified interface specification, it achieves compatibility with hardware from different manufacturers and models, reducing the coupling between the system and the hardware.

[0079] Real-time collection of hardware operating parameters, including power supply voltage, device temperature, and power consumption, provides health status assessment and reduces system failures caused by hardware malfunctions.

[0080] The hardware abstraction layer, acting as a bridge between the system and physical devices, is primarily responsible for shielding the differences between various hardware devices and providing a unified calling interface. In this embodiment, the hardware abstraction layer defines a standardized hardware interface protocol, enabling the business service layer and the AI ​​decision-making layer to achieve unified control of devices such as thrusters, servos, cameras, radars, and water quality sensors without needing to know the specific models or driver differences of the underlying hardware.

[0081] This layer integrates a device health monitoring module, which can collect real-time data on the current, voltage, temperature, and operating status parameters of each device. Upon detecting anomalies, it reports to the safety management plugin via the event bus, triggering the corresponding risk response mechanism. Furthermore, the hardware abstraction layer supports a firmware upgrade mechanism, allowing for remote downloading and updating of device firmware during system operation, ensuring that all hardware components maintain optimal performance at all times.

[0082] Through the above design, the hardware abstraction layer decouples the system from the underlying hardware, enabling it to adapt to different models of unmanned vessels and hardware platforms, thus enhancing the system's versatility and scalability.

[0083] In some embodiments, to further enhance system reliability, a secure boot mechanism is introduced to perform integrity verification on firmware and drivers when the system is powered on or restarted, reducing malicious tampering or loading of corrupted files, thereby ensuring the stable operation of the unmanned vessel in complex environments.

[0084] The edge computing layer is used to receive device control commands and generate execution results by combining them with real-time sensor streams.

[0085] The edge computing layer, as the lowest layer of the system, is deployed directly on the deck of the unmanned vessel. This layer integrates lightweight computing nodes and local AI inference modules, enabling preprocessing and preliminary analysis of sensor data within milliseconds, reducing data transmission latency, and maintaining the independent operation of core tasks even when communication is limited or interrupted. Through collaboration with the upper layers, the edge computing layer ensures the autonomy and robustness of the unmanned vessel.

[0086] In some embodiments, the edge computing layer includes a local decision-making submodule, a real-time control submodule, and an anomaly detection submodule; The local decision-making submodule is used to receive device control commands, perform real-time decision optimization, and generate optimized control commands. The real-time control submodule is used to perform signal conversion and timing control on the optimized control commands, and generate drive signals; The anomaly detection submodule is used to detect anomalies and provide status feedback on the drive signals and real-time sensor streams, and generate execution results.

[0087] The edge computing layer is deployed locally on the unmanned surface vessel (USV) and is responsible for low-latency data processing and automated inference to enhance the system's autonomous operation capabilities in weak network environments or when the network is down. It can process mission-critical data (such as obstacle avoidance information and attitude parameters) at the millisecond level, ensuring the USV's real-time response capabilities.

[0088] The edge computing layer runs deep learning models and rule-based reasoning systems locally, enabling real-time computation for trajectory planning, target recognition, and environmental prediction. It dynamically allocates CPU, GPU, and memory resources to ensure computational efficiency and system stability under multi-tasking conditions. In the event of a communication interruption, the edge computing layer can take over task execution, maintaining the unmanned vessel's safe navigation according to a pre-defined strategy until communication is restored or the mission is completed.

[0089] The edge computing layer is deployed directly on the unmanned vessel's deck, serving as the system's underlying execution environment. It is primarily responsible for local data preprocessing and rapid response. Equipped with lightweight, high-performance computing units, the edge computing layer supports multi-threaded parallel processing and AI model inference, enabling real-time analysis of sensor data within milliseconds. For example, when radar detects a sudden obstacle, the edge computing layer can immediately generate obstacle avoidance commands and send them to the business service layer for execution, ensuring the unmanned vessel's basic safety even in situations of communication latency or cloud unavailability.

[0090] In terms of data processing, the edge computing layer can compress and extract features from video streams and raw sensor data, uploading only necessary information to the cloud, thereby reducing communication bandwidth consumption and improving overall system efficiency. Simultaneously, this layer has offline operation capabilities, allowing it to continue executing preset operations based on locally cached tasks even when communication is lost, and synchronizing task data to the cloud after communication is restored. Through collaboration with upper layers, the edge computing layer not only enhances the autonomy and real-time performance of unmanned surface vessels but also provides reliable support for task execution in complex environments. Through the above design, the edge computing layer not only improves the real-time processing capability and environmental adaptability of unmanned vessels, but also provides technical support for their continuous and stable operation in complex waters and weak network conditions.

[0091] In this embodiment, the unmanned surface vessel (USV) intelligent control system achieves functional decoupling and flexible expansion through a layered design. Each layer provides plug-in interfaces, supporting the hot-swappable integration of new functional modules, enabling upgrades or replacements without a restart. Simultaneously, the system incorporates a plug-in store and configuration file management mechanism, allowing for the rapid activation or deactivation of different plug-ins based on specific task scenarios, forming differentiated functional combinations and thus constructing a complete USV application ecosystem.

[0092] In some embodiments, the unmanned vessel automatic control system based on a plug-in architecture supports a plug-in mechanism, allowing each sub-module of the front-end presentation layer, API gateway layer, AI decision layer, business service layer, data management layer, hardware abstraction layer, and edge computing layer to be dynamically loaded or unloaded as an independent plug-in, achieving hot-swappable functionality without restarting the system.

[0093] It's easy to understand that the unmanned surface vessel (USV) automated control system achieves the ability to assemble and hot-update system functions through a unified plug-in interface specification and plug-in lifecycle management. Specifically, this includes: Plugin Store and Configuration Loading: The system provides a directory and manifest file of functional plugins. Once a plugin meets the unified interface requirements, it can be enabled in the management interface. When a new function (such as camera, actuator control, or water pump control) needs to be added, simply place the plugin in the specified directory and check the enable option in the configuration. The system will then automatically load and integrate it without needing to change the core code or restart the system.

[0094] Hot-swap and online upgrade: Allows dynamic loading / unloading and version upgrades of plugins during operation; takes effect immediately upon activation, shortening debugging and deployment time and reducing downtime costs.

[0095] Plugin security governance: Combine security management services to perform signature verification, permission isolation and access control on plugins, reduce the installation of malicious plugins and unauthorized access, and ensure system stability and data security.

[0096] In some embodiments, to reduce module coupling and improve expansion flexibility, publish / subscribe communication supported by an event bus is adopted: Event bus and topic division: The functional modules do not directly call each other, but interact through the event bus broadcast and subscribed topics. New modules only need to subscribe to existing topics to reuse the information flow without changing the publisher.

[0097] Typical process example: When the GPS module generates a new location, it broadcasts a location update event; the map module, navigation module, and log module respectively perform drawing, path recalculation, and persistence operations based on the event, thereby achieving data-driven decoupled collaboration.

[0098] Figure 2This application provides a flowchart illustrating an event-driven communication mechanism, which enables decoupled interaction between different functional modules. A unified event bus is set up internally, with each functional module interacting as an event publisher and event subscriber. When a module generates a state change or task information, it encapsulates it as a standardized event and broadcasts it through the event bus; other modules that have subscribed to this type of event execute corresponding operations upon receiving the message. For example, the GPS module, acting as an event publisher, generates a location update event and broadcasts it to subscriber modules, including the map module, navigation module, and log module, through the event bus. This achieves decoupled interaction between different functional modules and supports the dynamic integration of new plug-in modules. After detecting a location update, the GPS module broadcasts the location event to the event bus. The navigation module, map module, and log module, acting as subscribers, receive the event and respectively execute path updates, map refreshes, and log recording, thereby achieving loosely coupled collaboration between different modules.

[0099] In terms of data modeling, event content has been standardized to improve consistency and scalability in cross-module communication. Specifically, location-based events use a unified geographic coordinate data format, supporting latitude, longitude, elevation, and timestamp information; sensor-based events use structured encapsulation of data such as temperature, flow rate, and water quality parameters, along with data source identifiers and sampling frequencies; and command stream events are standardized into a task command format, including fields such as command type, priority, execution target, and effective time period. Through standardized modeling, the system can support efficient transmission of multiple event types on a unified channel and facilitates the expansion and integration of subsequent plug-in modules.

[0100] To further improve the efficiency and reliability of the system during multi-task parallel operation, a data hierarchical transmission strategy is introduced into the event-driven communication mechanism. Specifically, events are divided into four priorities based on their importance and real-time nature: P0 events are the highest priority, typically including critical safety events such as emergency obstacle avoidance and communication interruptions; P1 events are high priority, mainly including task-critical events such as navigation path updates and device status alarms; P2 events are medium priority, involving routine task execution and status reporting; and P3 events are low priority, used for non-critical tasks such as log recording and background statistics. During scheduling, the system prioritizes the real-time transmission and processing of high-priority events, while low-priority events are executed when resources are idle, thereby achieving dynamic protection of task execution and system security.

[0101] In some embodiments, the combination of event-driven communication mechanisms and a plug-in architecture further enhances the system's flexibility and scalability. Plugins, as independent functional units, can declare the types of events they need to subscribe to or declare themselves as publishers broadcasting specific events to the system via a registration interface during runtime. When a new plugin is loaded, the system can automatically recognize and integrate it without modifying the core code, thus enabling rapid functional expansion. For example, when a new target recognition plugin is introduced, it only needs to be registered as a subscriber to camera video stream events; the plugin can then receive video frames in real time and perform recognition tasks without altering the camera driver or other business logic.

[0102] Furthermore, the event-driven mechanism provides a natural decoupling guarantee for plugins. Different plugins interact through an event bus rather than directly calling each other's interfaces, thus reducing the strong coupling between modules. Even if a plugin is uninstalled or fails, other plugins can still receive standardized events from the system through the event bus without affecting the overall functionality. This loosely coupled design not only reduces system maintenance costs but also provides greater flexibility and fault tolerance for subsequent feature expansion.

[0103] Furthermore, by combining event priority scheduling with a plugin management mechanism, rational allocation and dynamic scheduling of resources are achieved. When the system operates in a resource-constrained environment, the event bus prioritizes the transmission of high-priority events, while the plugin manager can temporarily suspend or unload low-priority plugins according to task requirements, thereby maintaining overall stability while ensuring the operation of core system functions. This mechanism is particularly suitable for unmanned surface vessels (USVs) performing tasks in complex waters or long-term independent operation scenarios, effectively improving system robustness and task completion rate.

[0104] In some embodiments, the automatic connection management module enables adaptive connection and dynamic switching of the unmanned vessel in multi-network environments, ensuring stable and secure communication during mission execution. This module supports multi-channel access from 4G / 5G cellular networks, satellite communication links, and dedicated wireless networks, and evaluates network quality based on real-time bandwidth, latency, and packet loss rate metrics.

[0105] When a degradation or interruption in the current link performance is detected, the automatic connection management module can automatically trigger a link switching mechanism, prioritizing links with lower latency and higher bandwidth to maintain communication quality. Simultaneously, this module integrates encryption and authentication mechanisms; all data transmission is encrypted through a secure tunnel, reducing the risk of unauthorized theft and tampering. In mission execution scenarios, such as remote video monitoring or high-frequency data backhaul, the automatic connection management module can perform traffic scheduling based on task priority, ensuring that critical mission data is transmitted first when network resources are scarce, thereby improving the mission continuity and safety of the unmanned surface vessel.

[0106] The device control interface module is primarily used to provide intuitive operation and status monitoring of various hardware devices on the unmanned surface vessel (USV). In this embodiment, the device control interface is implemented based on the front-end presentation layer and includes interactive components such as a virtual joystick, push lever slider, and status panel. Users can control the USV's heading and speed in real time using the virtual joystick, and adjust the thruster speed, robotic arm extension and retraction, and camera angle using the push lever slider.

[0107] The interface also integrates equipment operation status visualization, including real-time display of information such as battery level, servo angle, thruster speed, sensor status, and communication quality. When an equipment malfunction is detected, the interface will display a notification via icon highlighting or a pop-up window, and the notification information will be simultaneously reported to the safety management plugin to trigger appropriate risk response measures. Furthermore, this module supports preset operating modes, such as one-click return to home and one-click hovering, allowing users to trigger complex actions with a single command in emergencies, thereby improving the system's ease of operation and emergency response capabilities.

[0108] In some embodiments, a camera optimization mechanism module is also included, primarily used to improve the video acquisition and image transmission quality of the unmanned surface vessel in complex aquatic environments. This module supports multiple video inputs and provides split-screen display and picture-in-picture functionality on the front-end display layer, enabling operators to simultaneously monitor images from cameras in different directions or of different types.

[0109] In terms of video processing, the camera optimization mechanism introduces an adaptive bitrate adjustment function, which can dynamically adjust the video resolution, frame rate, and compression rate according to the current network bandwidth, thereby reducing data transmission load while ensuring real-time performance. For example, when the system detects insufficient network bandwidth, it automatically reduces video clarity and prioritizes preserving details in key areas to improve the accuracy of target recognition and environmental perception.

[0110] In addition, the module incorporates image enhancement algorithms, including defogging, contrast enhancement, and low-light enhancement, enabling the unmanned surface vessel to maintain clear imaging in complex environments such as fog, water surface reflections, or nighttime. For specific tasks, such as water surface debris identification or vessel tracking, this module also supports loading pluggable target detection algorithms to mark and identify targets in the video stream in real time, and push the results to the AI ​​decision-making layer for subsequent task planning and handling.

[0111] In some embodiments, a map and coordinate system module is also included to provide an environmental modeling and path reference framework for the unmanned surface vessel. This module supports the fusion of multi-source map data, including GPS positioning data, radar scan data, and cloud-based map resources. Through data fusion algorithms, the system can generate environmental maps with high accuracy and dynamic updating capabilities, providing a reliable foundation for path planning and mission execution.

[0112] In terms of coordinate system design, this module adopts a unified coordinate system management mechanism, supporting the mutual conversion between latitude and longitude coordinates, local UTM coordinates, and task-defined coordinates, so that path data under different task scenarios can be seamlessly integrated. Through coordinate alignment and time synchronization mechanisms, the system can uniformly map radar point clouds, camera-identified targets, and GPS locations to the same coordinate system, realizing the spatial fusion of multi-source sensing data.

[0113] Furthermore, a boundary crossing detection function has been introduced into the map and coordinate system module. When the unmanned surface vessel's position coordinates deviate from the preset channel or enter a restricted area, the system will immediately trigger a level-two warning event, and the safety management plugin will issue a prompt or execute a return-to-base operation. This mechanism effectively ensures the navigation safety and mission compliance of the unmanned surface vessel.

[0114] In some embodiments, a path planning system module is also included to provide the unmanned surface vessel (USV) with safe and efficient trajectory generation and dynamic adjustment capabilities. The path planning system employs a combination of global and local planning. Global planning generates an initial optimal route based on environmental data provided by a map and coordinate system, combined with the mission objective point and navigation constraints. Local planning responds in real time to environmental changes during the USV's navigation, such as immediately calculating avoidance paths and adjusting the course when encountering temporary obstacles or water flow interference.

[0115] In terms of algorithm implementation, this module supports plug-in integration of various path planning algorithms, including classic algorithms based on A* and Dijkstra, as well as automatic optimization algorithms based on deep reinforcement learning. During operation, the system can automatically select the appropriate planning algorithm based on the task scenario and computing resources to balance path optimality and real-time computation.

[0116] Furthermore, the path planning system supports multi-objective task scheduling. When the unmanned surface vessel (USV) needs to perform tasks between multiple observation points or operational areas, the system can calculate the optimal access order based on the traveling salesman problem (TSP) optimization method, thereby improving task efficiency. In terms of safety, this module introduces a risk weight assessment mechanism, transforming environmental parameters such as wind speed, water flow, and obstacle density into risk factors, and automatically avoiding high-risk areas during path calculation, thus ensuring the USV's navigation safety and mission success rate.

[0117] In some embodiments, a human-computer interaction and AI assistant module is also included to improve the efficiency and automation of operator-system collaboration. This module integrates natural language processing and speech recognition technologies, allowing operators to issue tasks directly via voice commands, such as "plan to the target point" or "start video surveillance." After receiving the voice command, the system performs semantic parsing through the AI ​​assistant and converts it into a corresponding task instruction, which is then further executed by the business service layer.

[0118] In terms of the user interface, the AI ​​assistant can proactively push notifications based on task progress and system status. For example, when the unmanned vessel's battery is low, the AI ​​assistant will prompt "Current battery is low, do you want to return?"; when a deviation from the route is detected, the AI ​​assistant will suggest to the operator whether to replan the route. This module also supports multimodal interaction, meaning that the operator can freely switch between voice commands and the graphical interface, ensuring efficient operation in different usage environments.

[0119] Furthermore, the AI ​​assistant module possesses learning and adaptive capabilities. By analyzing the operator's historical operating habits and task configuration preferences, the system can gradually optimize task recommendations and interaction logic. For example, when an operator repeatedly sets observation points in a specific area, the system will automatically recommend that area as a priority target point in subsequent tasks, thereby improving task configuration efficiency. In this way, the human-computer interaction and AI assistant module not only lower the operational threshold but also enhance the automation level and user experience of the unmanned surface vessel system.

[0120] In this embodiment, functional modules are decoupled through a layered architecture, event-driven communication mechanism and unified data model. Flexible expansion is supported by plug-in and hot-swappable mechanisms. The autonomous operation capability and safety of unmanned vessels in complex water environments are improved through automatic decision-making and edge computing.

[0121] In some embodiments, to improve cross-module interoperability and semantic consistency, a unified data model is established and data flows are organized in layers: A unified data model is used to standardize location information, sensor readings, control commands, timestamps, and precision fields, ensuring semantic and structural consistency across different modules and reducing ambiguity and redundant adaptation.

[0122] Real-time data stream path: sensor → data acquisition unit → data fusion unit → state estimator → controller → actuator, while bypassing and recording to local storage and adding to the upload queue to meet the dual requirements of online control and historical traceability.

[0123] Command control flow chain: host computer interface → API gateway → command verifier → message queue → slave computer receiver → task scheduler → device controller, so that commands are reliably issued after authentication and verification.

[0124] Data synchronization hierarchical strategy: Based on business importance and timeliness requirements, data is divided into P0–P3: P0 (safety-related, such as ship position / collision warning) is transmitted in real time, P1 (control-related) is transmitted in near real time, P2 (monitoring-related, such as video / log) allows for delay, and P3 (history-related) is transmitted offline in batches, balancing bandwidth and reliability.

[0125] Furthermore, to support multi-task parallelism and complex operating conditions, task classification, priority, and adaptive resource management are set up: Task Categories and Thread Pools: System tasks are divided into real-time tasks (such as sensor reading and device status monitoring), scheduled tasks (such as log cleaning and data backup), and computationally intensive tasks (such as path planning and image processing), and are allocated to dedicated thread pools with differentiated scheduling strategies.

[0126] Resource-aware management: The scheduler continuously monitors the usage of resources such as CPU, memory, and network. When resources are scarce, it reduces the frequency of non-critical tasks, clears the cache, or reduces the video bitrate to ensure the timeliness and stability of the core control link.

[0127] Priority system: Establish a task priority sequence from high to low: emergency security tasks > control command tasks > data processing tasks > background maintenance tasks, so that security and control closed loops are responded to in a timely manner before other business.

[0128] Furthermore, to ensure inherent safety in complex and unforeseen circumstances, a three-tiered safety protection system and graded emergency response mechanism are established for communication, navigation, and the system itself. Communication security: Heartbeat monitoring of link health, automatic entry into safe mode upon communication timeout, and support for redundant link switching to reduce the risk of single point of failure.

[0129] Navigation safety: Based on water boundary detection, collision risk assessment and obstacle avoidance strategies, it provides early warning or automatic avoidance when boundary crossing or high-risk situations occur, maintaining track safety.

[0130] System security: Monitors and provides early warnings for battery power, computing / storage resources, and equipment failures to reduce control instability caused by resource depletion or component malfunctions.

[0131] Emergency Response Tier: Level 1 Emergency (Severe): Immediately stop all movement, activate the emergency buoy, and send a distress signal; Level 2 Emergency (Warning): Switch to safe mode, reduce operating speed, and increase monitoring frequency; Level 3 Emergency (Alert): Record the warning message and notify the operator; the system remains under controlled operation. The triggering conditions and actions for the aforementioned tiered handling are jointly determined and executed by the business service layer and the AI ​​decision-making layer.

[0132] Figure 3 This is a flowchart of the security mechanism provided in the embodiments of this application, such as... Figure 3As shown, the implementation of the security mechanism covers three aspects: communication security, navigation security, and system security, forming a multi-level, interconnected security protection system. This mechanism not only ensures the safe execution of missions by unmanned vessels in complex aquatic environments but also triggers emergency measures in a timely manner in case of emergencies, thereby improving the overall robustness and reliability of the system.

[0133] In terms of communication security, multiple encryption and link monitoring mechanisms are adopted. All communication data is encapsulated through encrypted channels during transmission and supplemented by dynamic key updates to reduce the risk of data theft or tampering in the link. The system also has heartbeat detection and link monitoring functions. When a communication interruption or abnormal fluctuation is detected, it immediately reports to the security mechanism center and initiates the appropriate security level response based on the situation.

[0134] In terms of navigation safety, real-time monitoring of the unmanned surface vessel's (USV) trajectory is achieved through boundary detection and obstacle avoidance mechanisms. When the USV deviates from the preset course or enters a restricted area, the system automatically triggers a safety warning and takes measures such as returning to shore, slowing down, or stopping the vessel based on the risk level. Simultaneously, combined with the risk weight assessment of the path planning system, the navigation safety module can proactively avoid high-risk areas, ensuring the USV remains within a controllable range at all times.

[0135] In terms of system security, the stability of hardware and software operation is ensured through device health monitoring and resource scheduling mechanisms. When low battery power, excessive CPU or memory resource usage, or critical hardware failure are detected, the system will immediately trigger corresponding events, which will be graded and handled by the security mechanism center. In this way, the system can maintain the continuity of task execution under adverse environmental conditions and reduce the risk of failure.

[0136] Furthermore, a tiered emergency response mechanism is established, classifying abnormal events into three levels based on their severity: Level 1 Emergency, Level 2 Warning, and Level 3 Alert. Level 1 Emergency corresponds to severe anomalies, such as complete communication interruption, impending collision, or critical hardware damage. In this case, the system will immediately execute an emergency shutdown or return-to-base operation and send a distress signal via buoys. Level 2 Warning corresponds to medium risk, such as weakened signal, deviation from the flight path, or battery level below a threshold. The system will trigger a degraded operating mode, reducing speed and enhancing monitoring. Level 3 Alert corresponds to minor anomalies, such as slight deviation from the flight path, resource criticality, or minor hardware alarms. The system will log and push notifications to the operator while maintaining normal operation.

[0137] Regarding secure communication links, an end-to-end encryption mechanism is employed, with all communication links protected through a combination of symmetric and asymmetric encryption. Specifically, the system uses an asymmetric encryption algorithm for key exchange during mission initialization, followed by real-time encryption and decryption using an efficient symmetric encryption algorithm during data transmission, thereby reducing computational overhead while ensuring security. This link encryption mechanism effectively reduces man-in-the-middle attacks and data tampering, enhancing the confidentiality and integrity of information during remote control and data transmission of the unmanned vessel.

[0138] In terms of plugin management, a plugin verification mechanism is introduced, requiring all newly loaded or upgraded plugins to pass digital signature and integrity verification. Before loading, the plugin manager verifies the plugin's signature; only certified legitimate plugins can enter the system runtime environment, thereby reducing malicious code injection and unauthorized module loading. Simultaneously, the system periodically performs hash value verification on loaded plugins during operation to detect potential tampering, and immediately triggers a security event and initiates a tiered response mechanism upon detecting anomalies.

[0139] Through the above design, the security mechanism not only ensures the data security of the communication link but also controls the trustworthiness of plug-in access from the source, thus forming a full-link, full-lifecycle security protection system. This enables unmanned vessels to maintain a high level of safety and controllability when performing tasks in complex environments.

[0140] It should be noted that technical features such as event-driven mechanisms, plug-in architecture, security classification mechanisms, and path planning methods can be flexibly combined according to different actual application scenarios.

[0141] In this embodiment, by introducing an automatic task scheduling mechanism and resource management strategy, priority management of real-time tasks, scheduled tasks, and computationally intensive tasks is achieved. Combined with dynamic scheduling of CPU, memory, and network resources, efficient operation in a multi-task parallel environment is ensured. By designing a multi-level safety protection and emergency handling mechanism, graded responses are implemented for communication security, navigation security, and system security, enabling the unmanned vessel to operate safely and reliably even in emergencies. By deploying an edge computing layer on the vessel, millisecond-level response, local AI inference, and offline autonomous operation are achieved, enhancing the stability and continuity of the unmanned vessel in weak network or network outage environments.

[0142] This application also provides an automatic control method for unmanned vessels based on a plug-in architecture, which includes the following steps.

[0143] (1) Obtain the task request, send the task request to the front-end presentation layer, and generate a structured control request; (2) Send the structured control request to the API gateway layer for verification and merging to generate a unified request object after identity authentication; (3) Send the unified request object to the AI ​​decision layer, and generate a decision scheme based on multi-model fusion decision in combination with environmental perception data; (4) Send the decision plan to the business service layer, refine the decision plan into executable tasks based on the current system status, schedule and install the plug-in, and generate a task execution plan; (5) Send the task execution plan to the data management layer, process it in conjunction with historical execution data and real-time monitoring data, and generate status enhancement data; (6) Send the state enhancement data to the hardware abstraction layer for mapping and generate device control instructions; (7) Send the device control commands to the edge computing layer and generate execution results by combining the real-time sensor streams. According to the plug-in architecture-based unmanned surface vessel (USV) automatic control method provided in this application embodiment, the complex USV control system is decomposed into functionally independent and interface-standardized modules by adopting a plug-in architecture and hierarchical design. It supports plug-in stores and hot-swappable mechanisms to achieve rapid combination, flexible expansion and dynamic updates of functions. By introducing an event-driven communication mechanism and a unified data model, decoupling and efficient collaboration between different modules are achieved, improving the consistency of data exchange and the stability of system operation. By setting up an AI decision layer and combining various algorithms such as deep learning, reinforcement learning and expert systems, global situational awareness, route optimization and predictive maintenance are achieved, improving the autonomous intelligence level of USVs in complex water environments and reducing reliance on human intervention.

[0144] In some embodiments, such as Figure 4 As shown, this application embodiment also provides an electronic device 400, including a processor 401, a memory 402, and a computer program stored on the memory 402 and executable on the processor 401. When the program is executed by the processor 401, it implements the various processes of the above-described embodiment of the unmanned vessel automatic control method based on a plug-in architecture and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0145] It should be noted that the electronic devices in the embodiments of this application include the mobile electronic devices and non-mobile electronic devices described above.

[0146] This application also provides a non-transitory computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the various processes of the above-described embodiment of the unmanned vessel automatic control method based on a plug-in architecture and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0147] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0148] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described unmanned vessel automatic control method based on a plug-in architecture.

[0149] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0150] This application also provides a chip, which includes a processor and a communication interface. The communication interface and the processor are coupled. The processor is used to run programs or instructions to implement the various processes of the above-described embodiment of the unmanned vessel automatic control method based on a plug-in architecture, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0151] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a device-level chip, device chip, chip device, or on-chip device chip, etc.

[0152] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover 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 that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0153] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the unmanned vessel automatic control method based on plug-in architecture of the various embodiments of this application.

[0154] In the description of this application, "first feature" and "second feature" may include one or more of the features.

[0155] In the description of this application, "multiple" means two or more.

[0156] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

[0157] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0158] Although embodiments of this application have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of this application, the scope of which is defined by the claims and their equivalents.

Claims

1. An unmanned ship automatic control system based on a plug-in architecture, characterized by, The system comprises a front-end display layer, an API gateway layer, an AI decision layer, a business service layer, a data management layer, a hardware abstraction layer and an edge computing layer; The front-end display layer is configured to receive a task request, and generate a structured control request, wherein the task request comprises an operation instruction, a sensor signal, a task planning request or environment information; The API gateway layer is configured to receive the structured control request for verification and merging, and generate a unified request object after identity authentication; The AI decision layer is configured to receive the unified request object and environment perception data, and generate a decision scheme based on multi-model fusion decision; The business service layer is configured to receive the decision scheme, refine the decision scheme into executable tasks in combination with a current system state, and perform scheduling and plug-in assembly to generate a task execution plan; The data management layer is configured to receive the task execution plan, process the task execution plan in combination with historical execution data and real-time monitoring data, and generate state enhancement data; The hardware abstraction layer is configured to receive the state enhancement data for mapping, and generate a device control instruction; The edge computing layer is configured to receive the device control instruction, and generate an execution result in combination with a sensor real-time stream. The front-end display layer, the API gateway layer, the AI decision layer, the business service layer, the data management layer, the hardware abstraction layer and the edge computing layer are independent of each other, and are decoupled and communicated through an event-driven communication bus.

2. The unmanned ship automatic control system based on the plug-in architecture according to claim 1, characterized in that, The front-end display layer comprises a command analysis submodule, a multi-modal fusion submodule and a security verification submodule; The command analysis submodule is configured to receive a task request and analyze the task request, and generate a multi-modal instruction; The multi-modal fusion submodule is configured to fuse the multi-modal instruction, and generate a structured instruction; The security verification submodule is configured to perform permission verification and legality verification on the structured instruction, and generate a structured control request if the permission verification and the legality verification are passed.

3. The plugin-based architecture based automatic control system of unmanned ship according to claim 1, characterized in that, The API gateway layer comprises an authentication and authorization submodule, a request aggregation submodule and a security audit submodule; The authentication and authorization submodule is configured to receive a structured control request for identity authentication and permission verification, and generate an authentication context; The request aggregation submodule is configured to perform request analysis and rule matching on the authentication context, and generate an aggregated request; The security audit submodule is configured to collect and format a log of the aggregated request, and generate a unified request object.

4. The plugin-based architecture based automatic control system of unmanned ship according to claim 1, characterized in that, The AI decision layer comprises a data fusion submodule, a policy reasoning submodule and a risk assessment submodule; The data fusion submodule is configured to receive a unified request object and environment perception data, fuse the unified request object and the environment perception data based on a multi-model, and generate a fusion context; The policy reasoning submodule is configured to perform policy reasoning on the fusion context, and generate a policy proposal; The risk assessment submodule is configured to perform risk assessment on the policy proposal, and generate a decision scheme.

5. The plugin-based architecture based automatic control system of unmanned ship according to claim 1, characterized in that, The business service layer comprises a task scheduling submodule, a plug-in management submodule and a task monitoring submodule; The task scheduling submodule is configured to receive a decision scheme and a current system state for task decomposition and resource scheduling, and generate a task scheduling plan; The plug-in management submodule is configured to receive the task scheduling plan for plug-in matching and assembly, and generate a plug-in assembly scheme; The task monitoring submodule is configured to receive the plug-in assembly scheme for task monitoring, and generate a task monitoring result. The task monitoring submodule is configured to receive the plug-in assembly scheme to perform state monitoring and fault handling, and generate a task execution plan.

6. The plugin-based architecture based automatic control system of unmanned ship according to claim 1, wherein, The data management layer comprises a data acquisition submodule, a data cleaning submodule, and a state fusion submodule. The data acquisition submodule is configured to receive the task execution plan to acquire multi-source data, and generate original environment data. The data cleaning submodule is configured to clean and preprocess the original environment data, and generate standard format data. The state fusion submodule is configured to fuse the standard format data, historical execution data, and real-time monitoring data, and generate state enhanced data.

7. The plugin-based architecture based automatic control system of unmanned ship according to claim 1, wherein, The hardware abstraction layer comprises a protocol conversion submodule, a device monitoring submodule, and a security rollback submodule. The protocol conversion submodule is configured to receive the state enhanced data to perform protocol mapping and instruction conversion, and generate standard device instructions. The device monitoring submodule is configured to perform device state verification and resource inspection on the standard device instructions, and generate executable device instructions. The security rollback submodule is configured to perform security boundary verification and exception plan matching on the executable device instructions, and generate device control instructions.

8. The plugin-based architecture based automatic control system of unmanned ship according to claim 1, wherein, The edge computing layer comprises a local decision submodule, a real-time control submodule, and an exception detection submodule. The local decision submodule is configured to receive the device control instructions to perform real-time decision optimization, and generate optimized control instructions. The real-time control submodule is configured to perform signal conversion and timing control on the optimized control instructions, and generate driving signals. The exception detection submodule is configured to perform exception detection and state feedback on the driving signals and sensor real-time stream, and generate execution results.

9. The plugin-based architecture based automatic control system of unmanned ship according to claim 1, wherein, The unmanned ship automatic control system based on the plug-in architecture supports a plug-in mechanism, allows each submodule of the front-end display layer, the API gateway layer, the AI decision layer, the business service layer, the data management layer, the hardware abstraction layer, and the edge computing layer to be dynamically loaded or unloaded in the form of an independent plug-in, and realizes a hot plug function without restarting the system.

10. An unmanned ship automatic control method based on a plug-in architecture, implemented by the unmanned ship automatic control system based on the plug-in architecture according to any one of claims 1 to 9, characterized in that, The method comprises: obtaining a task request, sending the task request to the front-end display layer, and generating a structured control request; sending the structured control request to the API gateway layer for verification and merging, and generating a unified request object after identity authentication; sending the unified request object to the AI decision layer, combining environment perception data, and generating a decision scheme based on multi-model fusion decision; sending the decision scheme to the business service layer, combining the current system state, refining the decision scheme into executable tasks, and scheduling and plug-in assembling to generate a task execution plan; sending the task execution plan to the data management layer, combining historical execution data and real-time monitoring data for processing, and generating state enhanced data; sending the state enhanced data to the hardware abstraction layer for mapping, and generating device control instructions; sending the device control instructions to the edge computing layer, combining sensor real-time stream to generate execution results.

Images