Engineering vehicle task autonomous planning system and method based on multi-agent cooperation

Abstract

The invention discloses an engineering vehicle task autonomous planning system and method based on multi-agent cooperation, and relates to the technical field of engineering vehicle task autonomous planning, and the method comprises the steps: obtaining a building information model BIM of an engineering project, carrying out the voxelization segmentation of a construction region, and generating a plurality of voxel units with attribute marks; calculating a space-time coupling degree (STCD) between tasks based on a space distance and a process dependency relationship between the initial task units corresponding to each voxel unit, and clustering the task units according to the space-time coupling degree to form a plurality of collaborative operation clusters; constructing an equipment capability portrait of each engineering vehicle in the engineering vehicle team, combining the matched engineering vehicles into equipment groups in combination with task combination features of the collaborative operation clusters, and allocating a corresponding equipment group to each collaborative operation cluster; and deploying a multi-agent system on each engineering vehicle, establishing distributed communication connection through an inter-vehicle ad hoc network, and distributing a scheduling strategy of the collaborative operation cluster to a corresponding device group.

Description

Technical Field

[0001] This invention relates to the field of autonomous mission planning technology for engineering vehicles, and in particular to an autonomous mission planning system and method for engineering vehicles based on multi-agent collaboration. Background Technology

[0002] Autonomous task planning technology for engineering vehicles refers to a technological system that automatically allocates tasks, plans routes, and coordinates work sequences for engineering vehicle fleets through intelligent algorithms and systems, with minimal or no human intervention, and dynamically adjusts strategies based on environmental changes during execution. Therefore, how to utilize advanced technologies to improve the intelligence level and safety of autonomous task planning for engineering vehicles has become one of the most pressing issues to be addressed.

[0003] In the field of autonomous planning of engineering vehicle tasks, existing technologies usually divide tasks based on human experience or simple regional divisions without considering the spatial proximity and procedural dependencies between tasks. This results in scattered task units, low collaboration efficiency, and traditional methods only focus on spatial distance, ignoring construction procedural dependencies such as excavation before transportation and filling before compaction, which can easily lead to chaotic task order. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an autonomous planning method for engineering vehicle tasks based on multi-agent collaboration. This solves the problem that existing technologies usually divide tasks based on human experience or simple regional divisions, without considering the spatial proximity and process dependencies between tasks. This results in scattered task units, low collaboration efficiency, and traditional methods only focus on spatial distance, ignoring construction process dependencies such as excavation before transportation and filling before compaction, which can easily lead to chaotic task order.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides an autonomous planning method for engineering vehicle tasks based on multi-agent cooperation, comprising:

[0008] Obtain the Building Information Model (BIM) of the project and perform voxel segmentation on the construction area to generate multiple voxel units with attribute labels;

[0009] Based on the spatial distance and process dependency between the initial task units corresponding to each voxel unit, the spatiotemporal coupling degree (STCD) between tasks is calculated, and the task units are clustered according to the spatiotemporal coupling degree to form multiple collaborative operation clusters.

[0010] Construct equipment capability profiles for each engineering vehicle in the engineering fleet, combine them with the task combination characteristics of collaborative operation clusters, group the matched engineering vehicles into equipment groups, and assign corresponding equipment groups to each collaborative operation cluster;

[0011] A multi-agent system is deployed on each engineering vehicle, and a distributed communication connection is established through an inter-vehicle self-organizing network to distribute the scheduling strategy of the collaborative operation cluster to the corresponding equipment group. Each engineering vehicle constructs a local belief state based on local observation information.

[0012] During the collaborative decision-making process, the task completion confidence (TCC) of each engineering vehicle is calculated in real time. The task completion confidence comprehensively considers equipment health, availability, energy consumption status and historical task completion rate, and uses it as a dynamic weighting factor to participate in the calculation of the value function of multi-agent joint action, so as to adjust the contribution weight of each engineering vehicle in collaborative decision-making.

[0013] During the task execution, when any engineering vehicle detects an environmental anomaly or a change in state, a distributed replanning mechanism is triggered based on local observation. Through belief state sharing and a lightweight consensus protocol, the local collaborative strategy is dynamically adjusted without the need for intervention from a central node.

[0014] After the task is completed, actual operation data is collected, and the task completion confidence prediction model, equipment capability profile, and spatiotemporal coupling parameters are updated online to form a closed-loop optimization of planning-execution-learning.

[0015] As a preferred embodiment of the autonomous planning method for engineering vehicle tasks based on multi-agent collaboration described in this invention, the steps of calculating the spatiotemporal coupling degree (STCD) between tasks based on the spatial distance and process dependency relationship between the initial task units corresponding to each voxel unit, and clustering the task units according to the spatiotemporal coupling degree to form multiple collaborative operation clusters are as follows:

[0016] Acquire multiple initial task units, each task unit corresponding to a voxel unit, including its three-dimensional spatial position and task type;

[0017] For any two task units, calculate the spatial distance between them; the closer the distance, the higher the probability of collaboration.

[0018] Query the construction process knowledge base to determine whether there is a process dependency between the two tasks;

[0019] Taking into account both spatial distance and process dependency, the spatiotemporal coupling degree between the two is calculated using the following formula:

[0020] ;

[0021] in, For spatiotemporal coupling degree, This refers to the spatial distance between two task units. For process-dependent strength, These are weighting coefficients used to adjust the relative importance of space and process. This is the distance attenuation coefficient, which controls the rate at which spatial influence decreases with distance. It is a natural constant;

[0022] All task units are paired together The values ​​form a similarity matrix, and the task units are grouped using a spectral clustering algorithm;

[0023] Each group forms a collaborative work cluster, where tasks within the cluster are spatially adjacent and closely related in terms of construction process.

[0024] As a preferred embodiment of the autonomous planning method for engineering vehicle tasks based on multi-agent collaboration described in this invention, the steps of constructing equipment capability profiles for each engineering vehicle in the engineering vehicle fleet, combining the task combination characteristics of collaborative operation clusters, grouping matching engineering vehicles into equipment groups, and assigning corresponding equipment groups to each collaborative operation cluster are as follows:

[0025] Collect basic information about each engineering vehicle in the fleet, including equipment type, maximum load capacity, current battery or fuel level, equipment health status, and historical task completion status.

[0026] For each collaborative job cluster, analyze the combination of task types it contains;

[0027] Include engineering vehicles with matching equipment types from the fleet into the candidate set;

[0028] From the candidate set, prioritize devices with sufficient current battery power, good health status, and high historical task completion rate;

[0029] The selected devices are grouped into a device group and assigned to the corresponding collaborative task clusters to ensure that each type of task has at least one matching device.

[0030] The equipment group establishes a binding relationship with the collaborative operation cluster.

[0031] As a preferred embodiment of the autonomous planning method for engineering vehicle tasks based on multi-agent collaboration described in this invention, wherein: during the collaborative decision-making process, the task completion confidence level (TCC) of each engineering vehicle is calculated in real time, and the task completion confidence level comprehensively considers equipment health, availability, energy consumption status, and historical task completion rate. The specific steps are as follows:

[0032] For each engineering vehicle in the equipment group, obtain its real-time status data:

[0033] Equipment health is provided by the on-board health monitoring system, reflecting the operating status of key components of the engine and hydraulic system;

[0034] Availability is used to determine whether a device is idle; idle means yes, and currently in operation means no.

[0035] Energy consumption status is the percentage of current battery or fuel level relative to full capacity;

[0036] Historical task completion rate is the percentage of similar tasks successfully completed in the past 30 days.

[0037] The four indicators mentioned above are weighted according to preset weights to obtain the task completion confidence score (CC). The calculation formula is as follows:

[0038] ;

[0039] in, Confidence level for task completion Equipment health status For availability, In energy consumption state, For historical task completion rate;

[0040] The calculated The value serves as the basis for the credibility of the device in collaborative decision-making.

[0041] As a preferred embodiment of the autonomous planning method for engineering vehicle tasks based on multi-agent cooperation described in this invention, the specific steps for using it as a dynamic weighting factor in the calculation of the value function of the joint action of multiple agents to adjust the contribution weight of each engineering vehicle in collaborative decision-making are as follows:

[0042] Each engineering vehicle outputs a local value score of its suggested action in its local multi-agent decision-making module, representing the expected contribution of the action to completing the task.

[0043] Set the role coefficient based on the role of the device in the current task;

[0044] Confidence of task completion Multiply by the role coefficient to obtain the device's dynamic weight in joint decision-making;

[0045] When making collaborative decisions within the equipment group, high The recommended actions of devices with higher value receive higher weight and are adopted first.

[0046] Choose the combination of actions that maximizes overall synergistic benefits.

[0047] As a preferred embodiment of the autonomous planning method for engineering vehicle tasks based on multi-agent collaboration described in this invention, wherein: during task execution, when any engineering vehicle detects an environmental anomaly or state change, a distributed replanning mechanism is triggered based on local observation. Through belief state sharing and a lightweight consensus protocol, dynamic adjustment of the local collaboration strategy is achieved without the intervention of a central node. The specific steps are as follows:

[0048] An engineering vehicle uses sensors to detect obstacles, geological changes, or communication interruptions, determining that the environment is abnormal.

[0049] The engineering vehicle generates local replanning proposals, including new driving routes or work sequences;

[0050] The proposal and its current status information are broadcast to other engineering vehicles in the same group via the vehicle-to-vehicle communication network;

[0051] After other engineering vehicles receive the equipment, check for path conflicts or resource contention. If there are no conflicts, return "agree".

[0052] When more than half of the devices agree to the proposal, it is considered that a consensus has been reached, and the new plan will be implemented simultaneously by all devices.

[0053] If no consensus is reached, the device with the highest confidence in completing the task within the group will initiate a new proposal until a consistent strategy is formed.

[0054] As a preferred embodiment of the autonomous planning method for engineering vehicle tasks based on multi-agent collaboration described in this invention, the method includes: after the task is executed, collecting actual operation data, and updating the task completion confidence prediction model, equipment capability profile, and spatiotemporal coupling parameters online to form a closed-loop optimization of planning-execution-learning. The specific steps are as follows:

[0055] For each completed task, collect actual data, including actual time spent, energy consumption, whether it was completed on time, and path deviation;

[0056] Compare the actual completion status with the confidence level of the tasks completed in the planning phase. If the predictions are significantly different across multiple comparisons, adjustments should be made. The weighting coefficients in the calculation formula;

[0057] Update the device's historical task completion rate and synchronize it to the device's capability profile;

[0058] If a certain type of task combination frequently experiences coordination delays during actual execution, the process dependency weight in its spatiotemporal coupling should be appropriately increased. ;

[0059] All updates are completed locally, without relying on a central server, supporting continuous self-optimization of the system.

[0060] Secondly, this invention provides an autonomous task planning system for engineering vehicles based on multi-agent cooperation, comprising:

[0061] The module includes BIM feature segmentation, task clustering, equipment matching, collaborative decision-making, dynamic replanning, closed-loop learning, and communication collaboration.

[0062] The BIM voxel segmentation module is used to acquire the building information model (BIM) of the project, segment the construction area into voxels, and generate multiple voxel units with attribute labels as the basic units for task generation.

[0063] The task clustering module is used to calculate the spatiotemporal coupling degree between tasks based on the task spatial distance and process dependency relationship corresponding to the voxel unit, and to combine highly coupled tasks into collaborative operation clusters through clustering algorithms.

[0064] The equipment matching module is used to construct a profile of the equipment capabilities of engineering vehicles, and combine the task type combination characteristics of the collaborative operation cluster to select engineering vehicles with matching types and good status in the fleet, form equipment groups, and complete task allocation.

[0065] The collaborative decision-making module is used to perform multi-agent joint decision-making within the equipment group, calculate the task completion confidence of each engineering vehicle in real time, and generate dynamic weights by combining role coefficients, which are used to weight and fuse local action values.

[0066] The dynamic replanning module is used to generate a replanning proposal based on local observations when any engineering vehicle detects an environmental anomaly during task execution. The proposal is then broadcast through inter-vehicle communication and a lightweight consensus is reached, enabling dynamic adjustment of local strategies without centralized intervention.

[0067] The closed-loop learning module is used to collect actual operation data after the task is completed, compare the prediction with the actual performance, and dynamically update the task completion confidence model parameters, equipment capability profile and spatiotemporal coupling weight.

[0068] The communication and coordination module is used to establish an inter-vehicle self-organizing network among engineering vehicles, supporting scheduling strategy distribution, local information status sharing, replanning proposal broadcasting and consensus message interaction, and ensuring the real-time performance and reliability of distributed collaborative communication.

[0069] Thirdly, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the autonomous planning method for engineering vehicle tasks based on multi-agent cooperation as described in the first aspect of the present invention.

[0070] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the autonomous planning method for engineering vehicle tasks based on multi-agent cooperation as described in the first aspect of the present invention.

[0071] The beneficial effects of this invention are as follows: By constructing a task clustering mechanism based on spatiotemporal coupling, equipment capability profiling, and a dynamic confidence evaluation model, efficient collaborative planning and autonomous decision-making for engineering vehicle tasks are achieved. By adopting a distributed multi-agent architecture combined with local observation and a lightweight consensus protocol, real-time replanning is supported in environments with incomplete information and limited communication, significantly improving system robustness and responsiveness. Through task-equipment collaborative matching, dynamic weight decision-making, and online learning closed loop, resource utilization and construction efficiency are effectively optimized, empty running rate and collaborative conflicts are reduced, and intelligent, autonomous, and continuously evolving task planning for engineering vehicle groups in complex construction scenarios is realized. Attached Figure Description

[0072] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0073] Figure 1 This is a flowchart of the autonomous planning method for engineering vehicle tasks based on multi-agent collaboration in Example 1; Figure 2 This is a schematic diagram of the autonomous planning system for engineering vehicle tasks based on multi-agent collaboration in Example 1. Detailed Implementation

[0074] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0075] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0076] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.

[0077] Example, refer to Figure 1 and Figure 2 This embodiment of the invention provides an autonomous planning method for engineering vehicle tasks based on multi-agent cooperation, comprising the following steps:

[0078] S1. Obtain the Building Information Model (BIM) of the project and perform voxel segmentation on the construction area to generate multiple voxel units with attribute labels.

[0079] It should be noted that voxel segmentation discretizes the continuous three-dimensional construction space into regular cubic units. Each voxel unit carries attribute information such as location coordinates, terrain elevation, foundation type, and task category. As the basic unit for task generation and spatial reasoning, it is conducive to realizing refined task division and spatial conflict prediction, while providing a structured data foundation for subsequent task clustering.

[0080] S2. Based on the spatial distance and process dependency between the initial task units corresponding to each voxel unit, calculate the spatiotemporal coupling degree (STCD) between tasks, and cluster the task units according to the spatiotemporal coupling degree to form multiple collaborative operation clusters.

[0081] Furthermore, multiple initial task units are obtained, each task unit corresponding to a voxel unit, including its three-dimensional spatial position and task type;

[0082] For any two task units, calculate the spatial distance between them; the closer the distance, the higher the probability of collaboration.

[0083] Query the construction process knowledge base to determine whether there is a process dependency between the two tasks;

[0084] Taking into account both spatial distance and process dependency, the spatiotemporal coupling degree between the two is calculated using the following formula:

[0085] ;

[0086] in, For spatiotemporal coupling degree, This refers to the spatial distance between two task units. For process-dependent strength, These are weighting coefficients used to adjust the relative importance of space and process. This is the distance attenuation coefficient, which controls the rate at which spatial influence decreases with distance. It is a natural constant;

[0087] All task units are paired together The values ​​form a similarity matrix, and the task units are grouped using a spectral clustering algorithm;

[0088] Each group forms a collaborative work cluster, where tasks within the cluster are spatially adjacent and closely related in terms of construction process;

[0089] It should be noted that by introducing the spatiotemporal coupling index, spatial proximity and construction process logic are quantified in a unified manner, avoiding the limitations of relying solely on human experience or a single distance factor in traditional task division. The spectral clustering algorithm can effectively identify non-convex task groups, ensuring that tasks within the collaborative work cluster are both spatially concentrated and process-coherent, significantly improving the organizational efficiency of multi-equipment collaborative operations.

[0090] S3. Construct equipment capability profiles for each engineering vehicle in the engineering fleet, combine them with the task combination characteristics of collaborative operation clusters, combine the matched engineering vehicles into equipment groups, and assign corresponding equipment groups to each collaborative operation cluster.

[0091] Furthermore, collect basic information about each engineering vehicle in the fleet, including equipment type, maximum load capacity, current battery or fuel level, equipment health status, and historical task completion status.

[0092] For each collaborative job cluster, analyze the combination of task types it contains;

[0093] Include engineering vehicles with matching equipment types from the fleet into the candidate set;

[0094] From the candidate set, prioritize devices with sufficient current battery power, good health status, and high historical task completion rate;

[0095] The selected devices are grouped into a device group and assigned to the corresponding collaborative task clusters to ensure that each type of task has at least one matching device.

[0096] The equipment group establishes a binding relationship with the collaborative operation cluster;

[0097] It should be noted that the equipment capability profile enables a standardized description of heterogeneous engineering vehicles, which is then matched with task combination characteristics to ensure that the equipment group is highly adapted to the task requirements in terms of type composition, load capacity and operation rhythm, avoiding resource mismatch or capability redundancy, and improving the resource utilization and task response capability of the overall construction system.

[0098] S4. Deploy a multi-agent system on each engineering vehicle, establish a distributed communication connection through the inter-vehicle self-organizing network, distribute the scheduling strategy of the collaborative operation cluster to the corresponding equipment group, and each engineering vehicle constructs a local belief state based on local observation information.

[0099] Furthermore, for each engineering vehicle in the equipment group, obtain its real-time status data:

[0100] Equipment health is provided by the on-board health monitoring system, reflecting the operating status of key components of the engine and hydraulic system;

[0101] Availability is used to determine whether a device is idle; idle means yes, and currently in operation means no.

[0102] Energy consumption status is the percentage of current battery or fuel level relative to full capacity;

[0103] Historical task completion rate is the percentage of similar tasks successfully completed in the past 30 days.

[0104] The four indicators mentioned above are weighted according to preset weights to obtain the task completion confidence score (CC). The calculation formula is as follows:

[0105] ;

[0106] in, Confidence level for task completion Equipment health status For availability, In energy consumption state, For historical task completion rate;

[0107] The calculated The value serves as the basis for the credibility of the device in collaborative decision-making;

[0108] It should be noted that the local belief state integrates dynamic information such as the vehicle's own state, the location of neighboring vehicles, environmental perception, and task progress, enabling the engineering vehicle to perform autonomous reasoning and decision-making under conditions of limited communication or partial information loss; the combination of multi-agent architecture and self-organizing network eliminates the dependence on central control node and enhances the robustness and scalability of the system in complex construction environments.

[0109] S5. During the collaborative decision-making process, the task completion confidence (TCC) of each engineering vehicle is calculated in real time. The task completion confidence comprehensively considers the equipment health, availability, energy consumption status and historical task completion rate, and uses it as a dynamic weighting factor to participate in the calculation of the value function of the joint action of the multi-agent, so as to adjust the contribution weight of each engineering vehicle in the collaborative decision-making.

[0110] Furthermore, each engineering vehicle outputs a local value score of its suggested action in its local multi-agent decision-making module, representing the expected contribution of the action to completing the task.

[0111] Set the role coefficient based on the role of the device in the current task;

[0112] Confidence of task completion Multiply by the role coefficient to obtain the device's dynamic weight in joint decision-making;

[0113] When making collaborative decisions within the equipment group, high The recommended actions of devices with higher value receive higher weight and are adopted first.

[0114] Choose the combination of actions that maximizes overall synergistic benefits;

[0115] It should be noted that the Task Completion Confidence (TCC) is a dynamic weighting factor that enables the system to adaptively adjust the device's voice in collaborative decision-making based on the device's real-time status. Devices with high confidence gain higher decision weight in critical path tasks, thereby improving the overall reliability and stability of task execution and realizing an intelligent collaborative mechanism where the capable take on more responsibilities and the trustworthy lead the way.

[0116] S6. During the task execution process, when any engineering vehicle detects an environmental anomaly or a change in state, a distributed replanning mechanism is triggered based on local observation. Through belief state sharing and a lightweight consensus protocol, the local coordination strategy is dynamically adjusted without the need for intervention from the central node.

[0117] Furthermore, a certain engineering vehicle uses sensors to detect obstacles, geological changes, or communication interruptions, and determines that the environment is abnormal;

[0118] The engineering vehicle generates local replanning proposals, including new driving routes or work sequences;

[0119] The proposal and its current status information are broadcast to other engineering vehicles in the same group via the vehicle-to-vehicle communication network;

[0120] After other engineering vehicles receive the equipment, check for path conflicts or resource contention. If there are no conflicts, return "agree".

[0121] When more than half of the devices agree to the proposal, it is considered that a consensus has been reached, and the new plan will be implemented simultaneously by all devices.

[0122] If no consensus is reached, the device with the highest confidence in completing the task within the group will initiate a new proposal until a consistent strategy is formed.

[0123] It should be noted that this mechanism can achieve rapid response without relying on a central scheduling server. It achieves local consistency through a lightweight consensus protocol (improved Raft), effectively reducing communication overhead and decision-making delay. It only re-plans the affected areas, avoiding the waste of resources caused by global rescheduling, and ensuring the continuity of construction and the real-time performance of the system.

[0124] S7. After the task is completed, collect actual operation data and update the task completion confidence prediction model, equipment capability profile and spatiotemporal coupling parameters online to form a closed-loop optimization of planning-execution-learning.

[0125] Furthermore, for each completed task, actual data is collected, including actual time consumed, energy consumption, whether it was completed on time, and path deviation.

[0126] Compare the actual completion status with the confidence level of the tasks completed in the planning phase. If the predictions are significantly different across multiple comparisons, adjustments should be made. The weighting coefficients in the calculation formula;

[0127] Update the device's historical task completion rate and synchronize it to the device's capability profile;

[0128] If a certain type of task combination frequently experiences coordination delays during actual execution, the process dependency weight in its spatiotemporal coupling should be appropriately increased. ;

[0129] All updates are completed locally, without relying on a central server, supporting continuous self-optimization of the system;

[0130] It should be noted that the online update mechanism enables the system to continuously learn and self-optimize. By analyzing feedback from historical execution data, the system dynamically adjusts model parameters and evaluation criteria, gradually improving the accuracy of task prediction and resource matching, thus realizing the evolution from static planning to adaptive intelligent scheduling and enhancing the system's long-term adaptability in diverse construction scenarios.

[0131] This embodiment also provides an autonomous task planning system for engineering vehicles based on multi-agent cooperation, including:

[0132] The module includes BIM feature segmentation, task clustering, equipment matching, collaborative decision-making, dynamic replanning, closed-loop learning, and communication collaboration.

[0133] The BIM voxel segmentation module is used to acquire the Building Information Model (BIM) of an engineering project, segment the construction area into voxels, and generate multiple voxel units with attribute labels, which serve as the basic units for task generation.

[0134] The task clustering module is used to calculate the spatiotemporal coupling degree between tasks based on the task spatial distance and process dependency relationship corresponding to the voxel unit, and to combine highly coupled tasks into collaborative operation clusters through clustering algorithms;

[0135] The equipment matching module is used to build a profile of the equipment capabilities of engineering vehicles, and combine the task type combination characteristics of collaborative operation clusters to select engineering vehicles with matching types and good status in the fleet, form equipment groups and complete task allocation;

[0136] The collaborative decision-making module is used to perform multi-agent joint decision-making within the equipment group, calculate the task completion confidence of each engineering vehicle in real time, and generate dynamic weights by combining role coefficients for weighted fusion of local action values.

[0137] The dynamic replanning module is used to generate a replanning proposal based on local observations when any engineering vehicle detects an environmental anomaly during task execution. It then broadcasts the proposal through inter-vehicle communication and achieves a lightweight consensus, enabling dynamic adjustment of local strategies without centralized intervention.

[0138] The closed-loop learning module is used to collect actual operation data after the task is completed, compare the prediction with the actual performance, and dynamically update the task completion confidence model parameters, equipment capability profile, and spatiotemporal coupling weights.

[0139] The communication and coordination module is used to establish an inter-vehicle self-organizing network among engineering vehicles, supporting the distribution of scheduling strategies, sharing of local information status, broadcasting of replanning proposals and interaction of consensus messages, and ensuring the real-time performance and reliability of distributed collaborative communication.

[0140] This embodiment also provides a computer device applicable to the autonomous planning method for engineering vehicle tasks based on multi-agent cooperation, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the autonomous planning method for engineering vehicle tasks based on multi-agent cooperation as proposed in the above embodiment.

[0141] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0142] This embodiment also provides a storage medium on which a computer program is stored. When executed by a processor, the program implements the autonomous planning method for engineering vehicle tasks based on multi-agent cooperation as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0143] In summary, this invention achieves efficient collaborative planning and autonomous decision-making for engineering vehicle tasks by constructing a task clustering mechanism based on spatiotemporal coupling, equipment capability profiling, and a dynamic confidence assessment model. It employs a distributed multi-agent architecture combined with local observation and a lightweight consensus protocol to support real-time replanning in environments with incomplete information and limited communication, significantly improving system robustness and responsiveness. Through task-equipment collaborative matching, dynamic weight decision-making, and online learning loops, it effectively optimizes resource utilization and construction efficiency, reduces empty running rates and collaborative conflicts, and realizes intelligent, autonomous, and continuously evolving task planning for engineering vehicle groups in complex construction scenarios.

[0144] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for autonomous planning of engineering vehicle tasks based on multi-agent cooperation, characterized in that: include: Obtain the Building Information Model (BIM) of the project and perform voxel segmentation on the construction area to generate multiple voxel units with attribute labels; Based on the spatial distance and process dependency between the initial task units corresponding to each voxel unit, the spatiotemporal coupling degree (STCD) between tasks is calculated, and the task units are clustered according to the spatiotemporal coupling degree to form multiple collaborative operation clusters. Construct equipment capability profiles for each engineering vehicle in the engineering fleet, combine them with the task combination characteristics of collaborative operation clusters, group the matched engineering vehicles into equipment groups, and assign corresponding equipment groups to each collaborative operation cluster; A multi-agent system is deployed on each engineering vehicle, and a distributed communication connection is established through an inter-vehicle self-organizing network to distribute the scheduling strategy of the collaborative operation cluster to the corresponding equipment group. Each engineering vehicle constructs a local belief state based on local observation information. During the collaborative decision-making process, the task completion confidence (TCC) of each engineering vehicle is calculated in real time. The task completion confidence comprehensively considers equipment health, availability, energy consumption status and historical task completion rate, and uses it as a dynamic weighting factor to participate in the calculation of the value function of multi-agent joint action, so as to adjust the contribution weight of each engineering vehicle in collaborative decision-making. During the task execution, when any engineering vehicle detects an environmental anomaly or a change in state, a distributed replanning mechanism is triggered based on local observation. Through belief state sharing and a lightweight consensus protocol, the local collaborative strategy is dynamically adjusted without the need for intervention from a central node. After the task is completed, actual operation data is collected, and the task completion confidence prediction model, equipment capability profile, and spatiotemporal coupling parameters are updated online to form a closed-loop optimization of planning-execution-learning.

2. The autonomous planning method for engineering vehicle tasks based on multi-agent cooperation as described in claim 1, characterized in that: The method involves calculating the spatiotemporal coupling degree (STCD) between tasks based on the spatial distance and process dependencies between the initial task units corresponding to each voxel unit, and then clustering the task units according to the STCD to form multiple collaborative task clusters. The specific steps are as follows: Acquire multiple initial task units, each task unit corresponding to a voxel unit, including its three-dimensional spatial position and task type; For any two task units, calculate the spatial distance between them; the closer the distance, the higher the probability of collaboration. Query the construction process knowledge base to determine whether there is a process dependency between the two tasks; Taking into account both spatial distance and process dependency, the spatiotemporal coupling degree between the two is calculated using the following formula: ; in, For spatiotemporal coupling degree, This refers to the spatial distance between two task units. For process-dependent strength, These are weighting coefficients used to adjust the relative importance of space and process. This is the distance attenuation coefficient, which controls the rate at which spatial influence decreases with distance. It is a natural constant; All task units are paired together The values ​​form a similarity matrix, and the task units are grouped using a spectral clustering algorithm; Each group forms a collaborative work cluster, where tasks within the cluster are spatially adjacent and closely related in terms of construction process.

3. The autonomous planning method for engineering vehicle tasks based on multi-agent cooperation as described in claim 2, characterized in that: The process involves creating equipment capability profiles for each engineering vehicle in the construction vehicle fleet, combining them with the task combination characteristics of collaborative work clusters, grouping matched engineering vehicles into equipment groups, and assigning corresponding equipment groups to each collaborative work cluster. The specific steps are as follows: Collect basic information about each engineering vehicle in the fleet, including equipment type, maximum load capacity, current battery or fuel level, equipment health status, and historical task completion status. For each collaborative job cluster, analyze the combination of task types it contains; Include engineering vehicles with matching equipment types from the fleet into the candidate set; From the candidate set, prioritize devices with sufficient current battery power, good health status, and high historical task completion rate; The selected devices are grouped into a device group and assigned to the corresponding collaborative task clusters to ensure that each type of task has at least one matching device. The equipment group establishes a binding relationship with the collaborative operation cluster.

4. The autonomous planning method for engineering vehicle tasks based on multi-agent cooperation as described in claim 3, characterized in that: In the collaborative decision-making process, the task completion confidence level (TCC) of each engineering vehicle is calculated in real time. The task completion confidence level comprehensively considers equipment health, availability, energy consumption status, and historical task completion rate. The specific steps are as follows: For each engineering vehicle in the equipment group, obtain its real-time status data: Equipment health is provided by the on-board health monitoring system, reflecting the operating status of key components of the engine and hydraulic system; Availability is used to determine whether a device is idle; idle means yes, and currently in operation means no. Energy consumption status is the percentage of current battery or fuel level relative to full capacity; Historical task completion rate is the percentage of similar tasks successfully completed in the past 30 days. The four indicators mentioned above are weighted according to preset weights to obtain the task completion confidence score (CC). The calculation formula is as follows: ; in, Confidence level for task completion For equipment health, For availability, In energy consumption state, For historical task completion rate; The calculated The value serves as the basis for the credibility of the device in collaborative decision-making.

5. The autonomous planning method for engineering vehicle tasks based on multi-agent cooperation as described in claim 4, characterized in that: The specific steps for using it as a dynamic weighting factor in the calculation of the value function of multi-agent joint actions to adjust the contribution weight of each engineering vehicle in collaborative decision-making are as follows: Each engineering vehicle outputs a local value score of its suggested action in its local multi-agent decision-making module, representing the expected contribution of the action to completing the task. Set the role coefficient based on the role of the device in the current task; Confidence of task completion Multiply by the role coefficient to obtain the device's dynamic weight in joint decision-making; When making collaborative decisions within the equipment group, high The recommended actions of devices with higher value receive higher weight and are adopted first. Choose the combination of actions that maximizes overall synergistic benefits.

6. The autonomous planning method for engineering vehicle tasks based on multi-agent cooperation as described in claim 5, characterized in that: During task execution, when any engineering vehicle detects an environmental anomaly or state change, a distributed replanning mechanism is triggered based on local observation. Through belief state sharing and a lightweight consensus protocol, dynamic adjustment of the local coordination strategy is achieved without the intervention of a central node. The specific steps are as follows: An engineering vehicle uses sensors to detect obstacles, geological changes, or communication interruptions, determining that the environment is abnormal. The engineering vehicle generates local replanning proposals, including new driving routes or work sequences; The proposal and its current status information are broadcast to other engineering vehicles in the same group via the vehicle-to-vehicle communication network; After other engineering vehicles receive the equipment, check for path conflicts or resource contention. If there are no conflicts, return "agree". When more than half of the devices agree to the proposal, it is considered that a consensus has been reached, and the new plan will be implemented simultaneously by all devices. If no consensus is reached, the device with the highest confidence in completing the task within the group will initiate a new proposal until a consistent strategy is formed.

7. The autonomous planning method for engineering vehicle tasks based on multi-agent cooperation as described in claim 6, characterized in that: After the task is completed, actual operation data is collected, and the task completion confidence prediction model, equipment capability profile, and spatiotemporal coupling parameters are updated online to form a closed-loop optimization of planning-execution-learning. The specific steps are as follows: For each completed task, collect actual data, including actual time spent, energy consumption, whether it was completed on time, and path deviation; Compare the actual completion status with the confidence level of the tasks completed in the planning phase. If the predictions are significantly different across multiple comparisons, adjustments should be made. The weighting coefficients in the calculation formula; Update the device's historical task completion rate and synchronize it to the device's capability profile; If a certain type of task combination frequently experiences coordination delays during actual execution, the process dependency weight in its spatiotemporal coupling should be appropriately increased. ; All updates are completed locally, without relying on a central server, supporting continuous self-optimization of the system.

8. A multi-agent collaborative autonomous planning system for engineering vehicle tasks, based on the multi-agent collaborative autonomous planning method for engineering vehicle tasks as described in any one of claims 1 to 7, characterized in that: include: The module includes BIM feature segmentation, task clustering, equipment matching, collaborative decision-making, dynamic replanning, closed-loop learning, and communication collaboration. The BIM voxel segmentation module is used to acquire the building information model (BIM) of the project, segment the construction area into voxels, and generate multiple voxel units with attribute labels as the basic units for task generation. The task clustering module is used to calculate the spatiotemporal coupling degree between tasks based on the task spatial distance and process dependency relationship corresponding to the voxel unit, and to combine highly coupled tasks into collaborative operation clusters through clustering algorithms. The equipment matching module is used to construct a profile of the equipment capabilities of engineering vehicles, and combine the task type combination characteristics of the collaborative operation cluster to select engineering vehicles with matching types and good status in the fleet, form equipment groups, and complete task allocation. The collaborative decision-making module is used to perform multi-agent joint decision-making within the equipment group, calculate the task completion confidence of each engineering vehicle in real time, and generate dynamic weights by combining role coefficients, which are used to weight and fuse local action values. The dynamic replanning module is used to generate a replanning proposal based on local observations when any engineering vehicle detects an environmental anomaly during task execution. The proposal is then broadcast through inter-vehicle communication and a lightweight consensus is reached, enabling dynamic adjustment of local strategies without centralized intervention. The closed-loop learning module is used to collect actual operation data after the task is completed, compare the prediction with the actual performance, and dynamically update the task completion confidence model parameters, equipment capability profile and spatiotemporal coupling weight. The communication and coordination module is used to establish an inter-vehicle self-organizing network among engineering vehicles, supporting scheduling strategy distribution, local information status sharing, replanning proposal broadcasting and consensus message interaction, and ensuring the real-time performance and reliability of distributed collaborative communication.

9. A computer device, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the autonomous planning method for engineering vehicle tasks based on multi-agent cooperation as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the autonomous planning method for engineering vehicle tasks based on multi-agent cooperation as described in any one of claims 1 to 7.

Images