A multi-unmanned ship distributed coordination control method and system
By employing a multi-agent architecture and a hierarchical distributed coordination control method, the problems of incomplete information and heavy computational burden in multi-unmanned vessel collaborative operations are solved, achieving efficient task allocation and path planning, and improving the task execution efficiency and environmental adaptability of the unmanned vessel system.
Patent Information
- Application Number
- CN202610216109.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-02-14
- Publication Date
- 2026-08-25
AI Technical Summary
Existing multi-unmanned vessel collaborative operation strategies suffer from incomplete information acquisition, limited communication, and excessive computational burden in complex aquatic environments, making it difficult to meet the needs of marine exploration missions in large-scale unknown environments.
Employing a multi-agent architecture and a hierarchical approach, distributed coordinated control is achieved through unmanned surface vessel pose data fusion and a task scoring model. This optimizes task allocation and path planning, and utilizes the Jacobian matrix and the A+ algorithm for collaborative operations.
It improves the mission execution efficiency and environmental adaptability of multi-unmanned surface vessel systems, reduces the computational burden, adapts to dynamic environmental changes, and improves the efficiency of task allocation and path planning.
Smart Images
Figure CN122632675A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of unmanned vessel coordination and control technology, specifically to a distributed coordination and control method and system for multiple unmanned vessels. Background Technology
[0002] Unmanned surface vessels (USVs), as an innovative marine operational tool and an important branch of marine robotics, possess a high degree of autonomy and convenience. They typically integrate core subsystems such as autonomous navigation, communication and navigation, environmental perception, propulsion, and mission payloads. By flexibly configuring different functional modules, USVs can be applied to various military and civilian fields. Currently, they have been successfully applied in marine environmental monitoring, maritime security patrols, and maritime emergency rescue, greatly expanding humanity's boundaries and capabilities in exploring the ocean. However, due to the complex and variable aquatic environment and diverse operational requirements, single-vessel operation modes have significant limitations, mainly reflected in: limited operational range, insufficient payload compatibility, and weak environmental adaptability. Especially in large-scale, unknown marine exploration missions, they often fail to meet the complex requirements of actual tasks, potentially leading to mission failure.
[0003] Currently, there are two main strategies for handling multi-unmanned surface vessel (USV) collaborative operations: real-time local environment detection and global information area segmentation and route planning. The real-time local environment detection strategy requires USVs to rely on their own real-time detection of the local environment and information exchange with other USVs to obtain some dynamic information about the surrounding environment, thereby autonomously and flexibly planning navigation routes to avoid obstacles and other vessels. The global information area segmentation and route planning strategy, after clarifying the distribution of obstacles, divides the search area into multiple sub-regions, and the command center assigns search tasks to different USVs to achieve collaborative operations. This strategy can optimize the allocation of search areas as a whole, improve search efficiency, and plan relatively safe navigation paths when obstacle information is known. Compared with real-time local environment detection, the global information strategy is more mature, with a more systematic and comprehensive task allocation and route planning, better able to handle search tasks in known environments, ensuring reasonable resource allocation and efficient task execution, reducing redundant work, and improving the overall performance of the multi-USV system. However, both strategies have their shortcomings: real-time local environment detection suffers from incomplete information acquisition and limited communication; the global information strategy places a heavy computational burden on the central control.
[0004] Therefore, there is an urgent need to conduct in-depth research on multi-unmanned vessel cooperative operation algorithms. Summary of the Invention
[0005] 1. The technical problem to be solved:
[0006] To address the aforementioned technical problems, this invention provides a distributed coordinated control method and system for multiple unmanned vessels.
[0007] 2. Technical Solution:
[0008] A distributed coordinated control method for multiple unmanned surface vessels includes:
[0009] Step 1: Treat the coordinated control behavior of multiple unmanned vessels as corresponding intelligent agents; fuse the acquired pose data of multiple unmanned vessels and their corresponding surrounding environment data in the same coordinate system;
[0010] Step 2: Construct an optimal planning and completion model for unmanned surface vessel (USV) tasks. This model aims to complete the tasks and, under the constraints of USV capabilities and resource conditions, outputs a score for each unexecuted task. The model then uses the USV's own pose and real-time environmental data to score the unexecuted tasks, resulting in a list of unexecuted task scores. Each USV then filters its own list of unexecuted task scores to obtain a list of target tasks for each USV.
[0011] Step 3: Each unmanned surface vessel (USV) broadcasts its own list of target tasks to neighboring USVs to achieve distributed information exchange;
[0012] Step 4: Each unmanned surface vessel (USV) scores all received target tasks based on the USV task optimal planning completion model, resulting in a target task score list for each USV. Each USV then broadcasts its target task score list to USVs within a preset range.
[0013] Step 5: Each unmanned vessel retrieves its target task rating list based on each target task in its own target task list, thus obtaining all the ratings for each target task in each unmanned vessel's target task list.
[0014] Step Six: Each unmanned surface vessel (USV) filters all scores for each of its target tasks based on the USV task optimal planning completion model, and assigns the USV with the highest score for each target task as the task allocation object for each target task.
[0015] Furthermore, the data fusion in step one specifically includes: obtaining the original position, heading, and corresponding coordinate system of the unmanned vessel; synchronizing the time of all unmanned vessels; constructing a pose graph with the pose of the unmanned vessel as the node and the position data and observation data as the edge, and then obtaining the pose of each unmanned vessel by constructing the Jacobian matrix of each pose graph; and fusing the poses of all unmanned vessels to obtain the fused pose.
[0016] Furthermore, in step two: the optimal planning and completion model for the unmanned surface vessel (USV) mission includes optimal task selection at the top level and optimal path planning for task completion at the bottom level; the top level learns the optimal task selection by choosing the optimal task from the task pool and passes the task objective to the bottom level; the bottom level learns the path planning strategy, plans the optimal path for the USV mission, and feeds back the task execution status to the top level; the top level then optimizes its task selection strategy based on the feedback and dispatches a new optimal task selection task to the bottom level; the new optimal task selection task is the top level learning the optimal task selection after excluding the already selected optimal task from the task pool.
[0017] Furthermore, the list of unexecuted tasks in the multiple unmanned vessels includes all unexecuted tasks in the system.
[0018] Furthermore, the objective function for optimal task selection at the top level is to minimize the total travel path length from the unmanned vessel to its corresponding unexecuted task, with constraints including unmanned vessel navigation capability constraints, range constraints, type constraints, position constraints, and quantity constraints; the optimal path planning for completing the task at the bottom level adopts the A+ algorithm.
[0019] Furthermore, in step three, the neighboring unmanned vessels are unmanned vessels within a preset distance. If the neighboring unmanned vessels and the unmanned vessel belong to different base stations, information exchange is achieved through communication between the base stations.
[0020] A distributed coordination control system for multiple unmanned vessels is proposed, which utilizes a distributed coordination control method to achieve distributed coordination control of multiple unmanned vessels.
[0021] 3. Beneficial effects:
[0022] This invention provides a distributed coordinated control method and system for multiple unmanned surface vessels (USVs), which can improve the execution efficiency and adaptability of task allocation in multi-USV systems and shorten the USV's range. To adapt to dynamic environmental changes and alleviate the difficulties caused by joint optimization of task allocation and path planning, this solution adopts a multi-agent architecture and a hierarchical approach to model the joint optimization problem. The relationships between the multiple agents are similar to those between multiple robots, involving cooperation and competition, and they can interact with the dynamic environment in real time. The hierarchical approach decomposes the problem into multiple layers for processing, thereby mitigating the problems encountered in large-scale scenarios. Attached Figure Description
[0023] Figure 1 This is an overall flowchart of the present invention;
[0024] Figure 2 This is the overall architecture diagram of the optimal planning and completion model for unmanned surface vessel missions in this invention; Figure 3 This is a schematic diagram of an embodiment. Detailed Implementation
[0025] The present invention will now be described in detail with reference to the accompanying drawings.
[0026] As attached Figure 1 To be continued Figure 2 As shown, a distributed coordinated control method for multiple unmanned vessels includes:
[0027] Step 1: Treat the coordinated control behavior of multiple unmanned vessels as corresponding intelligent agents; fuse the acquired pose data of multiple unmanned vessels and their corresponding surrounding environment data in the same coordinate system;
[0028] Step 2: Construct an optimal planning and completion model for unmanned surface vessel (USV) tasks. This model aims to complete the tasks and, under the constraints of USV capabilities and resource conditions, outputs a score for each unexecuted task. The model then uses the USV's own pose and real-time environmental data to score the unexecuted tasks, resulting in a list of unexecuted task scores. Each USV then filters its own list of unexecuted task scores to obtain a list of target tasks for each USV.
[0029] Step 3: Each unmanned surface vessel (USV) broadcasts its own list of target tasks to neighboring USVs to achieve distributed information exchange;
[0030] Step 4: Each unmanned surface vessel (USV) scores all received target tasks based on the USV task optimal planning completion model, resulting in a target task score list for each USV. Each USV then broadcasts its target task score list to USVs within a preset range.
[0031] Step 5: Each unmanned vessel retrieves its target task rating list based on each target task in its own target task list, thus obtaining all the ratings for each target task in each unmanned vessel's target task list.
[0032] Step Six: Each unmanned vessel (UV) filters all scores for each of its target tasks based on the UV task optimal planning completion model, and assigns the UV with the highest score for each target task as the task allocation object for each target task.
[0033] Furthermore, the data fusion in step one specifically includes: obtaining the original position, heading, and corresponding coordinate system of the unmanned vessel; synchronizing the time of all unmanned vessels; constructing a pose graph with the pose of the unmanned vessel as the node and the position data and observation data as the edge, and then obtaining the pose of each unmanned vessel by constructing the Jacobian matrix of each pose graph; and fusing the poses of all unmanned vessels to obtain the fused pose.
[0034] Furthermore, in step two: the optimal planning and completion model for the unmanned surface vessel (USV) mission includes optimal task selection at the top level and optimal path planning for task completion at the bottom level; the top level learns the optimal task selection by choosing the optimal task from the task pool and passes the task objective to the bottom level; the bottom level learns the path planning strategy, plans the optimal path for the USV mission, and feeds back the task execution status to the top level; the top level then optimizes its task selection strategy based on the feedback and dispatches a new optimal task selection task to the bottom level; the new optimal task selection task is the top level learning the optimal task selection after excluding the already selected optimal task from the task pool.
[0035] Furthermore, the list of unexecuted tasks in the multiple unmanned vessels includes all unexecuted tasks in the system.
[0036] Furthermore, the objective function for optimal task selection at the top level is to minimize the total travel path length from the unmanned vessel to its corresponding unexecuted task, with constraints including unmanned vessel navigation capability constraints, range constraints, type constraints, position constraints, and quantity constraints; the optimal path planning for completing the task at the bottom level adopts the A+ algorithm.
[0037] Furthermore, in step three, the neighboring unmanned vessels are unmanned vessels within a preset distance. If the neighboring unmanned vessels and the unmanned vessel belong to different base stations, information exchange is achieved through communication between the base stations.
[0038] A distributed coordination control system for multiple unmanned vessels is proposed, which utilizes a distributed coordination control method to achieve distributed coordination control of multiple unmanned vessels.
[0039] Example:
[0040] As attached Figure 3 As shown, the background of this embodiment is an open area of 300km × 300km, with 20 irregularly shaped task areas randomly distributed in the area. Applying this scheme effectively alleviates the explosion of computational load. Therefore, the algorithm can still obtain a high-quality local optimum solution in large-scale scenarios. Furthermore, the superiority and stability of this scheme are verified through simulation experiments.
[0041] Although the present invention has been disclosed above with reference to preferred embodiments, these are not intended to limit the invention. Any person skilled in the art can make various changes or modifications without departing from the spirit and scope of the invention. Therefore, the scope of protection of the present invention should be defined by the scope of the claims of this application.
Claims
1. A distributed coordinated control method for multiple unmanned vessels, characterized in that: include: Step 1: Treat the coordinated control behavior of multiple unmanned vessels as corresponding intelligent agents; The pose data of multiple unmanned vessels and their corresponding surrounding environment data are fused in the same coordinate system; Step 2: Construct an optimal planning and completion model for unmanned surface vessel (USV) tasks. This model aims to complete the tasks and, under the constraints of USV capabilities and resource conditions, outputs a score for each unexecuted task. The model then uses the USV's own pose and real-time environmental data to score the unexecuted tasks, resulting in a list of unexecuted task scores. Each USV then filters its own list of unexecuted task scores to obtain a list of target tasks for each USV. Step 3: Each unmanned surface vessel (USV) broadcasts its own list of target tasks to neighboring USVs to achieve distributed information exchange; Step 4: Each unmanned surface vessel (USV) scores all received target tasks based on the USV task optimal planning completion model, resulting in a target task score list for each USV. Each USV then broadcasts its target task score list to USVs within a preset range. Step 5: Each unmanned vessel retrieves its target task rating list based on each target task in its own target task list, thus obtaining all the ratings for each target task in each unmanned vessel's target task list. Step Six: Each unmanned vessel (UV) filters all scores for each of its target tasks based on the UV task optimal planning completion model, and assigns the UV with the highest score for each target task as the task allocation object for each target task.
2. The distributed coordinated control method for multiple unmanned vessels according to claim 1, characterized in that: The data fusion in step one specifically includes: obtaining the original position, heading, and corresponding coordinate system of the unmanned vessel; synchronizing the time of all unmanned vessels; constructing a pose graph with the pose of the unmanned vessel as the node and the position data and observation data as the edge, and then obtaining the pose of each unmanned vessel by constructing the Jacobian matrix of each pose graph; and fusing the poses of all unmanned vessels to obtain the fused pose.
3. The distributed coordinated control method for multiple unmanned vessels according to claim 1, characterized in that: In step two: the optimal planning and completion model for the unmanned surface vessel (USV) mission includes optimal task selection at the top level and optimal path planning at the bottom level. The top level learns the optimal task selection, selects the optimal task from the task pool, and passes the task objective to the bottom level. The bottom level learns the path planning strategy, plans the optimal path for the USV mission, and feeds back the task execution status to the top level. The top level then optimizes its task selection strategy based on the feedback and dispatches a new optimal task selection task to the bottom level. The new optimal task selection task is the optimal task selection learned by the top level after excluding the already selected optimal task from the task pool.
4. The distributed coordinated control method for multiple unmanned vessels according to claim 1, characterized in that: The list of unexecuted tasks in multiple unmanned vessels includes all unexecuted tasks in the system.
5. The distributed coordinated control method for multiple unmanned vessels according to claim 1, characterized in that: The objective function for optimal task selection at the top level is to minimize the total travel path length from the unmanned vessel to its corresponding unexecuted task. Constraints include unmanned vessel navigation capability, range, type, position, and quantity. The optimal path planning at the bottom level for completing the task uses A... + algorithm.
6. The distributed coordinated control method for multiple unmanned vessels according to claim 1, characterized in that: In step three, the neighboring unmanned vessels are those within a preset distance. If the neighboring unmanned vessels belong to different base stations, information exchange is achieved through communication between the base stations.
7. A distributed coordinated control system for multiple unmanned vessels, characterized in that: Distributed coordinated control of multiple unmanned vessels can be achieved using the methods described in claims 1-6.