An intelligent agent collaborative management and control task planning high-quality data set construction method
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- XIAN AISHENG TECH GRP
- Filing Date
- 2026-03-30
- Publication Date
- 2026-07-21
AI Technical Summary
In the field of intelligent agent collaborative management and control task planning, there are problems such as scarcity of high-quality data, low efficiency and broken decision chains. Existing data construction methods are difficult to meet the needs of deep learning models for massive high-quality samples and lack the ability to model complex constraints.
A rule-based digitization engine based on job rule compilation is adopted to build a digitization system including environmental constraints, target constraints, resource constraints and electromagnetic rule modules. Combined with a five-dimensional task template architecture and a two-stage generation mechanism, industry standards are enforced to generate a full-link job analysis report.
It has achieved high-fidelity, strongly constrained, and full-chain automated production of control and management task data, improved data construction efficiency, met the needs of intelligent decision-making models for massive amounts of high-quality samples, and solved the problem of decision chain breakage.
Smart Images
Figure FT_1 
Figure FT_2 
Figure SMS_1
Abstract
Description
Technical Field
[0001] This invention belongs to the field of artificial intelligence training data engineering, specifically relating to a method for constructing high-quality datasets for collaborative management and control task planning of intelligent agents. Background Technology
[0002] The development of current intelligent decision-making systems faces severe data bottlenecks, particularly in the field of autonomous task planning for intelligent agents. Traditional training data construction methods have fundamental limitations: manual construction relies on the experience of industry experts, resulting in significantly insufficient generation efficiency, which falls far short of meeting the needs of deep learning models for massive amounts of high-quality samples; while automated generation technology can improve output speed, it generally lacks the ability to model complex constraints in the work area, leading to serious errors in the work logic.
[0003] The existing technology system suffers from three structural defects: First, the data scale is severely mismatched with the model requirements, and the long cycle and high cost of manual construction restrict the system's iteration speed. Second, general generation methods are difficult to adapt to the special rules of professional fields, often ignoring environmental dynamics (such as the constraints of meteorological conditions on sensor performance), equipment characteristics (such as the endurance limit of intelligent agents), and operational principles (such as the handling specifications for high-value targets), resulting in a large number of invalid samples that violate operational rules. Third, mainstream datasets often exhibit the problem of broken decision chains, that is, they focus on static question answering and lack the complete deduction process of "situational awareness - resource coordination - dynamic response", which leads to the trained intelligent agents exhibiting defects such as mechanical response and insufficient multi-constraint coordination ability in real adversarial environments.
[0004] Related patented technologies, such as path planning data generation solutions, are mostly limited to simple single-machine scenarios and cannot support the complexity of multi-machine collaborative operations. Although industry standards have standardized the data element framework, the implementation process still heavily relies on individual experience, resulting in problems such as weak scalability and low standardization. Commercial natural language generation tools, lacking domain knowledge embedding, often produce operational solutions with serious flaws such as misuse of terminology and logical contradictions. These limitations have substantially hindered the practical application of intelligent command and decision-making systems, necessitating the establishment of a new data construction paradigm that combines economies of scale with operational fidelity. Summary of the Invention
[0005] The technical problem to be solved by this invention is the scarcity of high-quality data in the training of intelligent decision-making models in the field of intelligent agent collaborative management and control task planning. Existing data construction methods suffer from technical bottlenecks such as low efficiency, lack of complex constraint modeling capabilities, and broken decision chains.
[0006] To address the aforementioned technical problems, this invention provides a method for constructing a high-quality dataset for intelligent agent collaborative management and control task planning, comprising: A rule digitization engine is built based on operational rules to transform industry standards into an executable digital rule system. The rule digitization engine includes an environmental constraint module, a target constraint module, a resource constraint module, and an electromagnetic rule module. The rule digitization engine is used to enforce industry standard constraints during the data generation process. A global meteorological parameter model is generated based on the injected dynamic parameters, target system features are deployed, and the meteorological parameter model and target system features are input into the rule digitization engine. Based on the constraints of the rule-based digital engine, real-time situational awareness, resource strategy scheduling, phased task execution, emergency response, and closed-loop decision optimization are performed, generating a full-link operation analysis report and updating the rule base. In particular, a three-stage structure based on reconnaissance, response, and assessment is forcibly constructed during phased task execution.
[0007] Furthermore, the operation rules are compiled based on a five-dimensional task template architecture, which includes a dynamic monitoring template, a channel blocking template, a rapid suppression template, a radar tracking template, and a maneuver interception template.
[0008] Furthermore, the environmental constraint module is configured to automatically identify the impact of weather conditions on equipment performance and disable specific sensors in severe weather; the target constraint module is configured to forcibly match corresponding reconnaissance and handling procedures based on the target's mobility and concealment characteristics.
[0009] Furthermore, the resource constraint module is configured to dynamically optimize the combination configuration of intelligent agents through cost-benefit analysis; the electromagnetic rule module is configured to monitor changes in the spectrum environment in real time and intelligently switch tracking and countermeasure strategies.
[0010] Furthermore, before compiling and building the rule digitization engine based on job rules, the method further includes: Load a standardized 3D operation area, randomly generate multi-base geographic coordinates, dynamically allocate intelligent agent resource pools, and bind core performance parameters of equipment.
[0011] Furthermore, the execution of the stage tasks includes: During the positioning phase, collaborative reconnaissance routes are planned, and high-frequency coordinate acquisition and real-time transmission are carried out. During the disposal phase, target attributes and disposal unit types are matched, and precise disposal instructions are executed according to categories. During the verification phase, a multimodal assessment of the damage effects was initiated, and targets that did not meet the standards were marked to generate a supplementary treatment list.
[0012] Furthermore, the resource strategy scheduling means selecting a resource combination mode according to the operation strategy. The resource combination mode includes an economical scheme, an efficiency scheme, and a reliability scheme. In the economical scheme, basic machine types are prioritized; in the efficiency scheme, the nearest available unit is activated; and in the reliability scheme, redundant processing forces are deployed.
[0013] Furthermore, the emergency response includes: Monitor signal interruption events and initiate a countdown protocol; The detection target moves out of the monitoring range, triggering an offset alarm. Perform emergency mode switching and emergency resource allocation.
[0014] Furthermore, the closed-loop decision optimization includes: Quantitatively evaluate core operational metrics, including target exposure rate and response latency. Tracing back to the root causes of scenarios that fail to meet standards; dynamically adjusting resource matching parameters and handling strategies.
[0015] Furthermore, the end-to-end operation analysis report includes: Event timeline reconstruction, resource consumption statistics, and calculation of work target achievement rates; After the task is terminated, recyclable work units are released to the standby pool, and archived environment parameters and rule base updates are performed.
[0016] The beneficial effects of this invention are as follows: This invention pioneers a dual-drive architecture of "rule constraints + generative AI." By constructing a rule-based digital engine encompassing four modules—environment, target, resources, and electromagnetic fields—abstract industry standards are transformed into an executable digital rule system. This ensures that the generated data strictly adheres to industry standards, addressing the lack of modeling capabilities for complex constraints in operational areas by traditional automated generation methods and significantly improving data fidelity. This invention enforces a three-stage structure of "reconnaissance-disposal-assessment" during phased task execution, filling the gap in the comprehensive "situational awareness-resource coordination-dynamic response" extrapolation commonly lacking in mainstream datasets. This solves the problem of broken decision chains and provides training samples with causal reasoning paradigms for intelligent agents. Through a five-dimensional task template architecture and a dual-stage generation mechanism, this invention achieves high-fidelity, strongly constrained, and full-chain automated production of control operation task data, significantly improving data construction efficiency and meeting the needs of large-scale intelligent decision-making models for massive amounts of high-quality samples. Attached Figure Description
[0017] Figure 1 This is a flowchart illustrating a method for constructing a high-quality dataset for collaborative management and control task planning of intelligent agents according to the present invention. Figure 2This is a schematic diagram of the logical architecture of the rule digitization engine of the present invention, which is a method for constructing a high-quality dataset for collaborative management and control task planning of intelligent agents. Detailed Implementation
[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with the embodiments of this invention. Obviously, the described embodiments are merely some embodiments of this invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.
[0019] Terminology Explanation: In this invention, the "rule digitization engine" refers to a logical framework that transforms abstract industry standards into an executable digital rule system. By deeply analyzing work specifications, work area constraints, and work principles, it systematically extracts key rules in dimensions such as environmental perception, target characteristics, resource scheduling, and electromagnetic adaptation, and encodes them into machine-readable decision instructions.
[0020] The "Five-Dimensional Mission Template Architecture" refers to a standardized mission logic framework that includes dynamic monitoring templates, channel blocking templates, rapid suppression templates, radar tracking templates, and maneuver interception templates, used to guide operational procedures in different scenarios.
[0021] Prior to this invention, existing methods for constructing intelligent agent task planning data were mainly divided into two categories: one is the manual construction mode, which relies on the experience of industry experts and has significantly insufficient generation efficiency, making it difficult to meet the needs of deep learning models for massive amounts of high-quality samples; the other is automated generation technology, which, although improving the output speed, generally lacks the ability to model complex constraints of the operating area (such as weather conditions and equipment characteristics), resulting in generated samples that often violate industry standards and have distorted tactical logic.
[0022] In this application, the intelligent agent can be a drone.
[0023] To address the aforementioned issues, this invention proposes an innovative solution based on a task template engine and a rule-based digital kernel to overcome the data generation bottleneck in complex decision-making scenarios such as dynamic control, electromagnetic adaptation, and continuous monitoring in intelligent agent collaborative management tasks. Through a five-dimensional standardized template architecture and a two-stage generation mechanism, it achieves high-fidelity, strongly constrained, and end-to-end automated production of management task data. The specific technical solution is as follows: Example 1: This embodiment provides a method for constructing a high-quality dataset for planning collaborative management and control tasks of intelligent agents. For example... Figure 1 As shown, the method includes the following steps: S101. Based on the operation rules, a rule digitization engine is built to transform industry standards into an executable digital rule system. The rule digitization engine includes an environmental constraint module, a target constraint module, a resource constraint module, and an electromagnetic rule module. The rule digitization engine is used to enforce industry standard constraints during the data generation process.
[0024] Among them, the rule digitization engine constructs the underlying logical framework for intelligent decision-making, ensuring that the data generation process is systematic and predictable by transforming industry standards into machine-readable decision instructions.
[0025] Specifically, the engine deeply analyzes operational specifications, operational area constraints, and operational principles, and systematically extracts key rules in dimensions such as environmental perception, target characteristics, resource scheduling, and electromagnetic adaptation, forming an interconnected decision network.
[0026] This enables the digital mapping of industry standards, ensuring that the generated data strictly follows the operational logic of the professional field, and solving the problem that traditional methods lack the ability to model complex constraints.
[0027] In one embodiment of this application, the operation rule compilation is based on a five-dimensional task template architecture, which includes a dynamic monitoring template, a channel blocking template, a rapid suppression template, a radar tracking template, and a maneuvering interception template.
[0028] Among them, the dynamic monitoring template adopts a dual-machine collaborative positioning and all-time monitoring mechanism; the channel blocking template dynamically constructs a control network based on a key node coordinate library; the rapid suppression template establishes an emergency response mechanism; the radar tracking template automatically triggers collaborative positioning through an electromagnetic sensing system; and the mobile interception template analyzes target motion characteristics in real time. Clearly, by executing these steps, diverse operational scenarios can be covered, providing a standardized operational logic framework for intelligent agent collaborative management and control, and improving the scenario adaptability and completeness of the dataset.
[0029] In one embodiment of this application, the environmental constraint module is configured to automatically identify the impact of weather conditions on equipment performance and disable specific sensors under severe weather conditions; the target constraint module is configured to forcibly match the corresponding reconnaissance and handling procedures based on the target's mobility and concealment characteristics.
[0030] The environmental constraint module analyzes the impact of weather on sensors, such as disabling photoelectric sensors in rainy weather; the target constraint module enforces matching procedures based on target characteristics, such as requiring cooperative positioning for moving targets. This ensures that environmental and target characteristics impose mandatory constraints on the decision-making logic, avoiding the generation of invalid samples that violate physical laws and principles, and significantly improving data fidelity.
[0031] In one embodiment of this application, the resource constraint module is configured to dynamically optimize the combination configuration of intelligent agents through cost-benefit analysis; the electromagnetic rule module is configured to monitor changes in the spectrum environment in real time and intelligently switch tracking and countermeasure strategies.
[0032] Among them, the resource constraint module optimizes the configuration through cost-benefit analysis, and the electromagnetic rule module monitors spectrum changes and intelligently switches strategies to form an interconnected decision network, retaining the flexibility to deal with emergencies. This realizes the optimal solution for resource scheduling and the dynamic response to electromagnetic countermeasures, and enhances the value of data samples in complex confrontation environments.
[0033] For example, the dynamic surveillance template is configured to employ a dual-machine collaborative positioning and all-time monitoring mechanism, providing sufficient reconnaissance units for moving targets. Emergency strikes are activated only when the target exceeds the monitoring range, ensuring continuous target exposure. Reconnaissance resources can be cyclically utilized, and processing units are activated on demand, achieving efficient cost control.
[0034] The channel blocking template is configured to dynamically construct a control network based on a key node coordinate library, deploying a multi-level monitoring system: reconnaissance units patrol and perceive the entire area, response units hold key positions, and reserve forces provide flexible support. A minimum control effectiveness threshold is set, and reinforcement procedures are automatically initiated if the threshold is not met.
[0035] The rapid suppression template is configured to establish an emergency response mechanism, prioritizing the dispatch of the nearest operational units. A high-confidence handling mode is enforced for moving targets, while resource-intensive handling strategies are simultaneously disabled. A reconnaissance unit reuse mechanism balances operational efficiency and economy.
[0036] The radar tracking module is configured to automatically trigger cooperative positioning based on an electromagnetic sensing system, ensuring target locking accuracy. A rapid switching scheme is designed to address signal interruptions, maintaining continuous monitoring capabilities. A spectrum discrimination module is included to effectively identify and eliminate interference from false targets.
[0037] The mobile interception template is configured to analyze target movement characteristics in real time and dynamically construct path blocking schemes. Pre-deployed response units form area control, and a collaborative mechanism combining damage assessment and secondary response ensures the effectiveness of the interception.
[0038] refer to Figure 2 The rule-based digitization engine is configured as such to transform abstract industry standards into an executable digital rule system, building the underlying logical framework for intelligent decision-making. This engine deeply analyzes operational specifications, operational area constraints, and operational principles, systematically extracting key rules across dimensions such as environmental perception, target characteristics, resource scheduling, and electromagnetic adaptation, and encoding them into machine-readable decision instructions.
[0039] The environmental constraint module is configured to automatically identify the impact of weather conditions on equipment performance, such as disabling specific sensors in severe weather; the target constraint module forces the matching of corresponding reconnaissance and disposal procedures based on the characteristics of target mobility and concealment; the resource constraint module dynamically optimizes the combination configuration of intelligent agents through cost-benefit analysis; and the electromagnetic rules module monitors changes in the spectrum environment in real time and intelligently switches tracking and countermeasure strategies.
[0040] These rules are not simple conditional judgments, but rather form an interconnected decision-making network: when radar signal characteristics are abnormal, the decoy identification process is automatically triggered; when a moving target leaves the monitoring range, the emergency response protocol is immediately activated; when resource consumption exceeds a threshold, an economical strategy is autonomously switched. Through this deep coupling of multi-dimensional rules, it ensures that every action strictly adheres to industry standards, while retaining flexibility to respond to unforeseen circumstances.
[0041] In one embodiment of this application, before compiling and building a rule digitization engine based on job rules, the method further includes: Load a standardized 3D operation area, randomly generate multi-base geographic coordinates, dynamically allocate intelligent agent resource pools, and bind core performance parameters of equipment.
[0042] This involves loading a standardized 3D operational area and randomly generating multi-base geographic coordinates; dynamically allocating the intelligent agent resource pool, configuring reconnaissance, response, and special mission aircraft according to a preset ratio, and binding core performance parameters such as flight speed, endurance, and payload capacity. This constructed basic spatial and resource environment provides the necessary data foundation and physical carrier for subsequent rule compilation and deduction.
[0043] For example, the spatial framework parameters are: a standardized operating area of 400km × 400km, with 5 bases randomly distributed on the west side. Parameter injection includes: meteorological parameters: temperature (15-45℃), humidity (20-90%), wind speed (0-15m / s), and wind direction (tailwind / crosswind). Targets include: a mixture of mobile targets (aircraft, engineering vehicles) and fixed facilities (factory buildings, underground storage). Resources include: the type and quantity of intelligent agents are dynamically allocated according to the operating specifications.
[0044] The processor can then perform the following steps to accomplish the following: Environmental perception: Analyzing the impact of weather on sensors (such as photoelectric failure due to rain).
[0045] Resource matching: Selecting the optimal combination based on strategy.
[0046] Timing planning: Mandate a three-stage structure of "reconnaissance-disposal-assessment".
[0047] Effectiveness verification: Output quantitative indicators such as situational maintenance rate and exposure duration.
[0048] S102. Generate a global meteorological parameter model based on the injected dynamic parameters, deploy the target system features, and input the meteorological parameter model and target system features into the rule digitization engine.
[0049] This includes generating a global meteorological parameter model, configuring temperature gradient curves, setting humidity fluctuation ranges, and combining wind speed and direction data; deploying target system characteristics, constructing motion trajectory equations for moving targets, generating coordinate networks for fixed facilities, and simulating electromagnetic signal characteristic spectra.
[0050] Specifically, the generated meteorological parameter model and target system characteristics are used as input variables and passed to the rule digitization engine to trigger subsequent intelligent decision-making inference. This realizes the parameterized input of dynamic environment and target characteristics, provides specific scenario data support for the operation of the rule engine, and ensures the diversity of data samples.
[0051] S103. Based on the constraints of the rule-based digital engine, perform real-time situational awareness, resource strategy scheduling, phased task execution, emergency response, and closed-loop decision optimization, generate a full-link operation analysis report and update the rule base. In particular, during the phased task execution, a three-stage structure based on reconnaissance, response, and assessment is forcibly constructed.
[0052] Among them, it continuously scans the target motion vector and predicts trajectory deviation, dynamically monitors changes in signal characteristics, performs resource scheduling, task execution and emergency response based on rule constraints, quantitatively evaluates core operation indicators, and generates a full-link operation analysis report.
[0053] Specifically, a three-stage structure of "reconnaissance-disposal-assessment" is forcibly constructed during the execution of phased tasks to ensure that each decision-making link conforms to a logical closed loop, realizes intelligent decision-making simulation of the entire process, solves the problem of broken decision chains in mainstream datasets, and ensures the integrity and logical rigor of data samples.
[0054] In one embodiment of this application, the execution of stage tasks includes: During the positioning phase, collaborative reconnaissance routes are planned, and high-frequency coordinate acquisition and real-time transmission are carried out. During the disposal phase, target attributes and disposal unit types are matched, and precise disposal instructions are executed according to categories. During the verification phase, a multimodal assessment of the damage effects was initiated, and targets that did not meet the standards were marked to generate a supplementary treatment list.
[0055] The process includes: positioning phase, planning collaborative reconnaissance routes and collecting coordinates; disposal phase, matching target attributes with disposal unit types; and verification phase, initiating multimodal damage effect assessment and marking targets that do not meet the standards.
[0056] This led to the construction of standardized operational process data, filling the gap in the full-process simulation data of "situational awareness-resource coordination-dynamic response" and improving the standardization of agent training.
[0057] In one embodiment of this application, resource strategy scheduling refers to selecting a resource combination mode according to the operation strategy. The resource combination mode includes an economical scheme, an efficiency scheme, and a reliability scheme. In the economical scheme, basic machine types are prioritized; in the efficiency scheme, the nearest available unit is activated; and in the reliability scheme, redundant processing forces are deployed.
[0058] The economical approach prioritizes the use of basic equipment, the efficiency approach activates the nearest available unit, and the reliability approach deploys redundant resources. The optimal response path is calculated, and the energy consumption cost matrix is evaluated. Specific strategy selection methods are detailed in Table 1.
[0059] Table 1 Strategy Selection Methods
[0060] By providing a diverse range of strategy selection samples, the agent's adaptability to different mapping and resource conditions is enhanced.
[0061] In one embodiment of this application, emergency response includes: Monitor signal interruption events and initiate a countdown protocol; The detection target moves out of the monitoring range, triggering an offset alarm. Perform emergency mode switching and emergency resource allocation.
[0062] Among these measures, the system monitors signal interruption events and initiates a countdown protocol; detects targets leaving the monitoring range and triggers offset alarms; and performs emergency mode switching (backup unit replacement, tracking technology conversion) and emergency resource scheduling.
[0063] Furthermore, it simulates unexpected situations in real-world confrontations, enhancing the agent's adaptability and robustness in uncertain environments.
[0064] In one embodiment of this application, closed-loop decision optimization includes: Quantitatively evaluate core operational metrics, including target exposure rate and response latency. Tracing back to the root causes of scenarios that fail to meet standards; dynamically adjusting resource matching parameters and handling strategies.
[0065] Among them, the core indicators of the operation are quantitatively evaluated (target exposure rate, response delay time); the root causes of non-compliance scenarios are traced back; and the logic of dynamically adjusting resource matching parameters and handling strategies is dynamically adjusted.
[0066] This enables the self-correction and optimization of the decision-making process, ensures the high quality and high fidelity of the data samples, and provides excellent feedback signals for model training.
[0067] In one embodiment of this application, the end-to-end job analysis report includes: Event timeline reconstruction, resource consumption statistics, and calculation of work target achievement rates; After the task is terminated, recyclable work units are released to the standby pool, and archived environment parameters and rule base updates are performed.
[0068] This includes generating a full-link operation analysis report (event timeline reconstruction, resource consumption statistics, and operation indicator achievement rate calculation); releasing recyclable operation units to the standby pool; and archiving environment parameters and rule base updates.
[0069] This completed the final archiving and resource recovery of the data samples, providing a traceable basis for analysis and data accumulation for subsequent tasks.
[0070] Example 2 This embodiment takes the scenario of emergency collaborative patrol and material delivery for urban flood disasters as an example to explain in detail the specific application of the above method.
[0071] I. Task Template Adaptation: This scenario uses both the dynamic monitoring template and the rapid suppression template for adaptation. The dynamic monitoring template is used for continuous monitoring of flooded areas, traffic congestion points, and the locations of trapped personnel, supporting multi-machine collaborative positioning and continuous target tracking. The rapid suppression template transforms into an emergency response mechanism, prioritizing the dispatch of the nearest intelligent agent to perform material delivery or communication relay tasks, disabling resource-intensive disposal strategies, and emphasizing a balance between reconnaissance unit reuse and response timeliness.
[0072] II. Mission Area Initialization: The spatial framework is set as a 30 km × 30 km urban built-up area. Three temporary take-off and landing points are randomly distributed on the west side of the area, located at a school playground, a gymnasium, and an emergency command center, respectively. Meteorological parameters are dynamically injected: temperature is set to 24-28 degrees Celsius, humidity range is 85%-95%, wind speed fluctuates randomly between 6 and 12 meters per second, wind direction is variable with gusts, weather condition is set as intermittent heavy rain, and visibility is less than 200 meters. The target system adopts a hybrid generation mode. Mobile targets include trapped people and emergency vehicles, while fixed facilities include water accumulation points, communication base stations, emergency material warehouses, and temporary shelters. The intelligent agent resource pool is dynamically allocated according to operational specifications, configuring three types of intelligent agents: reconnaissance intelligent agents, delivery intelligent agents, and relay intelligent agents, each bound to core performance parameters such as flight speed, endurance, and payload capacity.
[0073] III. The rules are digitally compiled, and the environmental constraint module automatically identifies the impact of weather conditions on equipment performance, disabling light delivery agents when wind speeds exceed 10 meters per second and forcibly activating thermal imaging and radar fusion sensing modes when visibility is below 200 meters. The target constraint module forcibly matches handling procedures based on target characteristics; delivery of supplies is only authorized after three consecutive frames of confirmation from a trapped target, and the safety radius around the landing point must be assessed to be no less than 3 meters before delivery. The resource constraint module dynamically optimizes agent combination configuration through cost-benefit analysis, ensuring that delivery agents do not repeatedly deploy more than three times in a single mission, and prioritizing the deployment of communication relay agents at the edge of signal blind spots. The electromagnetic rules module monitors spectrum environment changes in real time, automatically identifying and avoiding interference from civilian communication frequency bands, and automatically hopping frequencies to prevent collisions in multi-machine communication.
[0074] IV. Intelligent Decision-Making and Inference: In the environmental perception phase, the impact of rainfall on the attenuation of photoelectric sensors is analyzed, and the risk of communication interruption caused by building obstruction is identified. In the resource matching phase, the optimal combination is selected based on the strategy, prioritizing the use of reconnaissance and delivery teams to perform tasks at stranded points, with relay agents deployed on the rooftops of high-rise buildings. In the time-series planning phase, a four-stage structure of "location—confirmation—delivery—evaluation" is enforced, and an effect evaluation is completed within 30 seconds after delivery to confirm whether the supplies have been accurately delivered.
[0075] V. Emergency Response: The system continuously monitors for signal interruption events and initiates a countdown protocol. When an agent signal interruption is detected, a relay agent is automatically activated to fill the gap and the system switches to a backup communication link. When a target leaves the monitoring range, an offset alarm is triggered, and the system updates trajectory predictions and replans delivery points. Emergency mode switching is performed, including backup unit replacement and tracking technology conversion. When resources are scarce, the system automatically switches to an economical delivery mode, prioritizing high-priority stranded points.
[0076] VI. Closed-loop decision optimization and task termination management: During task execution, core operational indicators are evaluated in real time, including target exposure rate and response latency. The root causes of non-compliance scenarios are traced back, and resource matching parameters and handling strategies are dynamically adjusted. After the task is completed, a full-link operation analysis report is generated, including event timeline reconstruction, resource consumption statistics, and operation indicator achievement rate calculation. Recyclable operation units are released to the standby pool, and archived environment parameters and rule base updates are performed.
[0077] This embodiment generates high-fidelity training data covering the entire process of "situational awareness - resource coordination - dynamic response" through the synergistic effect of structured task templates, rule digitization engine and two-stage generation mechanism, verifying the applicability and scalability of the present invention.
[0078] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.
Claims
1. A method for constructing a high-quality dataset for agent-based collaborative management and control task planning, characterized in that, include: A rule digitization engine is built based on operational rules to transform industry standards into an executable digital rule system. The rule digitization engine includes an environmental constraint module, a target constraint module, a resource constraint module, and an electromagnetic rule module. The rule digitization engine is used to enforce industry standard constraints during the data generation process. A global meteorological parameter model is generated based on the injected dynamic parameters, target system features are deployed, and the meteorological parameter model and target system features are input into the rule digitization engine. Based on the constraints of the rule-based digital engine, real-time situational awareness, resource strategy scheduling, phased task execution, emergency response, and closed-loop decision optimization are performed, generating a full-link operation analysis report and updating the rule base. In particular, a three-stage structure based on reconnaissance, response, and assessment is forcibly constructed during phased task execution.
2. The method according to claim 1, characterized in that, The operation rules are compiled based on a five-dimensional task template architecture, which includes a dynamic monitoring template, a channel blocking template, a rapid suppression template, a radar tracking template, and a maneuver interception template.
3. The method according to claim 1, characterized in that, The environmental constraint module is configured to automatically identify the impact of weather conditions on equipment performance and disable specific sensors in severe weather; the target constraint module is configured to forcibly match the corresponding reconnaissance and handling procedures based on the target's mobility and concealment characteristics.
4. The method according to claim 1, characterized in that, The resource constraint module is configured to dynamically optimize the combination configuration of intelligent agents through cost-benefit analysis; the electromagnetic rule module is configured to monitor changes in the spectrum environment in real time and intelligently switch tracking and countermeasure strategies.
5. The method according to claim 1, characterized in that, Before compiling and building the rule digitization engine based on job rules, the method further includes: Load a standardized 3D operation area, randomly generate multi-base geographic coordinates, dynamically allocate intelligent agent resource pools, and bind core performance parameters of equipment.
6. The method according to claim 1, characterized in that, The execution of the phased tasks includes: During the positioning phase, collaborative reconnaissance routes are planned, and high-frequency coordinate acquisition and real-time transmission are carried out. During the disposal phase, target attributes and disposal unit types are matched, and precise disposal instructions are executed according to categories. During the verification phase, a multimodal assessment of the damage effects was initiated, and targets that did not meet the standards were marked to generate a supplementary treatment list.
7. The method according to claim 1, characterized in that, The resource strategy scheduling refers to selecting a resource combination mode based on the operation strategy. The resource combination mode includes an economical mode, an efficiency mode, and a reliability mode. In the economical mode, basic machines are prioritized; in the efficiency mode, the nearest available unit is activated; and in the reliability mode, redundant processing forces are deployed.
8. The method according to claim 1, characterized in that, The emergency response includes: Monitor signal interruption events and initiate a countdown protocol; The detection target moves out of the monitoring range, triggering an offset alarm. Perform emergency mode switching and emergency resource allocation.
9. The method according to claim 1, characterized in that, The closed-loop decision optimization includes: Quantitatively evaluate core operational metrics, including target exposure rate and response latency. Tracing back to the root causes of scenarios that fail to meet standards; dynamically adjusting resource matching parameters and handling strategies.
10. The method according to claim 1, characterized in that, The end-to-end job analysis report includes: Event timeline reconstruction, resource consumption statistics, and calculation of work target achievement rate; After the task is terminated, recyclable work units are released to the standby pool, and archived environment parameters and rule base updates are performed.