A fire-fighting unmanned equipment cluster collaborative task allocation and control system based on a dynamic auction algorithm

Abstract

This application relates to a collaborative task allocation and control system for firefighting unmanned equipment clusters based on a dynamic auction algorithm. It includes a task dynamic publishing module, an equipment intelligent agent module, a dynamic auction coordination module, and a flexible cluster communication and reconfiguration module. Real-time situational information is obtained through a digital twin model of the fire scene, dynamically generating and updating a set of firefighting tasks with spatiotemporal constraints, resource constraints, and coupling relationships. Each firefighting unmanned equipment constructs a dynamic performance model based on its own state and participates in task bidding. The auction coordination module uses a multi-round improved auction algorithm to achieve globally optimal task allocation. The system switches between centralized auction, distributed auction, and autonomous collaborative modes based on communication status and supports task reallocation and cluster reconstruction. This application improves the task allocation efficiency, collaborative flexibility, and overall operational reliability of firefighting unmanned equipment clusters in complex fire scene environments, and is applicable to the intelligent collaborative operation of various types of firefighting unmanned equipment.

Description

Technical Field

[0001] This application relates to the field of collaborative control of intelligent fire protection and unmanned systems, and in particular to a collaborative task allocation and control system for fire protection unmanned equipment clusters based on a dynamic auction algorithm. Background Technology

[0002] Currently, in high-rise / super high-rise buildings, large chemical industrial parks, and forest fire scenarios, the application of unmanned firefighting equipment (drones, ground robots) is evolving from "single equipment demonstrations" to "multi-equipment collaboration." However, existing collaborative control systems face four core challenges: 1. Task dynamism and uncertainty: The fire situation changes rapidly, and sudden events such as new fire ignition points, the location of trapped personnel, and building collapses require the task allocation plan to be adjusted in real time.

[0003] 2. Equipment heterogeneity and capability differences: Reconnaissance drones, firefighting drones, demolition robots, transport robots, and other equipment have different functions, payloads, and endurance.

[0004] 3. Communication constraints and distributed decision-making: Complex fire environments often lead to communication interruptions or bandwidth limitations, and centralized control poses a risk of single point of failure.

[0005] 4. Balancing real-time performance and optimality: Traditional optimization algorithms (such as integer programming) are computationally time-consuming and cannot meet the needs of second-level decision-making; while simple rule allocation (such as nearest distance) is inefficient.

[0006] Furthermore, existing unmanned firefighting systems mostly adopt pre-programmed formations or manual remote control from a central command console, which cannot achieve intelligent cluster operations with autonomous coordination of multiple equipment, dynamic bidding for tasks, and optimal overall benefits.

[0007] Currently, existing technologies have technical defects in the following aspects: 1. A typical solution for centralized optimization is task allocation based on mixed integer linear programming (MILP). Its core drawbacks are high computational complexity, slow response, reliance on global information, high communication pressure, and unsuitability for large-scale centralized management and dynamically changing fire scenes. 2. Typical solutions for market auction algorithms are Contract Network (CN) and multi-round bidding, which are time-consuming. The core drawback is that they do not consider the degradation of equipment capabilities and the coupling relationship between tasks, and are not suitable for long-term collaborative operations. 3. Typical solutions for swarm intelligence employ bee colony and ant colony biomimetic algorithms, but their core drawbacks are unstable convergence speed, difficulty in parameter tuning, and poor interpretability. 4. The typical scheme of pre-set formation adopts master-slave and hierarchical formation. Its core drawback is the lack of flexibility. The failure of any equipment will cause the formation to fail. It is not suitable for complex and ever-changing disaster sites. Summary of the Invention

[0008] To address the aforementioned technical deficiencies, this application provides a collaborative task allocation and control system for firefighting unmanned equipment clusters based on a dynamic auction algorithm.

[0009] The technical solution provided in this application for a collaborative task allocation and control system for firefighting unmanned equipment clusters based on a dynamic auction algorithm is as follows: A collaborative task allocation and control system for firefighting unmanned equipment clusters based on a dynamic auction algorithm includes: The task dynamic release module is used to generate and dynamically update a set of fire-fighting tasks with spatiotemporal constraints, resource constraints and coupling relationships based on real-time situation information obtained from the fire scene digital twin model. The equipment intelligent agent module is deployed on each fire-fighting unmanned equipment to build a dynamic performance model based on the current position, remaining load, remaining energy and operation capability of the fire-fighting unmanned equipment, and to calculate the bid value for the task set or task package accordingly. The dynamic auction coordination module is used to execute multi-round improved auction algorithms, receive bidding information from various equipment intelligent agent modules, resolve conflicting bids, and complete task allocation based on the principle of maximizing global utility. The flexible cluster communication and reorganization module is used to switch between different collaborative modes according to changes in communication status, and to maintain the collaborative operation capability of the fire-fighting unmanned equipment cluster in the event of communication restrictions or interruptions.

[0010] By adopting the above technical solutions, the system can perceive the development of fire, environmental changes, and resource status in real time based on the fire scene digital twin model through the task dynamic release module, dynamically generate and update fire-fighting tasks with spatiotemporal constraints and coupling relationships, and improve the real-time and targeted nature of task planning. The equipment intelligent agent module enables each fire-fighting unmanned equipment to have autonomous perception and decision-making capabilities. It can combine its own position, load, energy, and operational capabilities to evaluate the task execution efficiency in real time and make reasonable bids, realizing distributed intelligent decision-making for task allocation. The dynamic auction coordination module uses a multi-round improved auction algorithm to globally coordinate and resolve conflicts in bidding, maximizing the overall utility of the cluster while taking into account efficiency and fairness, and avoiding resource conflicts and duplicate operations. The flexible cluster communication and reorganization module supports flexible switching of collaborative modes under different conditions such as normal, limited, or interrupted communication, ensuring the continuous operation capability and system robustness of the cluster in complex communication environments.

[0011] Optionally, the dynamic performance model in the equipment intelligent agent module is represented as follows: Where EbaseiE_{base_i}Ebasei represents the basic operational capability of the equipment, Li(t) represents the load, Bi(t) represents the remaining energy, and D(task,posi(t) represents the distance or path cost from the current position of the equipment to the mission position.

[0012] By adopting the above technical solutions, the model provides a unified and comparable decision-making basis for the bidding value of equipment intelligent agents to calculate tasks, which is conducive to achieving accurate matching of equipment capabilities and resource status during task allocation, avoiding unreasonable scheduling of equipment with insufficient energy or excessive load, thereby improving the rationality of cluster task allocation, execution efficiency and overall operational safety.

[0013] Optionally, the weighting coefficients α, β, γ, and δ are adaptively adjusted by the dynamic auction coordination module according to the urgency of the task, the risk level, and changes in the fire situation, and broadcast to each equipment intelligent agent module through the cluster communication module.

[0014] By adopting the above technical solutions, the dynamic performance model can flexibly emphasize different decision-making factors according to the urgency of the task, the risk level, and changes in the fire situation. In urgent or high-risk tasks, it prioritizes response speed and operational capabilities, while in routine tasks, it takes into account energy consumption and load balancing. Furthermore, it ensures consistent evaluation standards for all equipment agents through unified broadcasting via the cluster communication module.

[0015] Optionally, the method for the dynamic auction coordination module to handle tasks with coupled relationships includes: Identify multiple tasks that have temporal or cooperative dependencies; The multiple tasks are combined to form a task package; Individual firefighting drones or virtual alliances of multiple firefighting drones are allowed to bid on the mission package as a whole.

[0016] By adopting the above technical solutions, tasks with temporal or collaborative dependencies can be identified and packaged, enabling multiple tasks that originally affected each other to be allocated as a whole under a unified framework. This avoids task fragmentation and assignment that could lead to execution conflicts or reduced efficiency. At the same time, it allows individual equipment or virtual alliances to bid on task packages as a whole, which is conducive to giving full play to the collaborative advantages of multiple equipment, improving the success rate of complex tasks and overall operational efficiency, and enhancing the cluster's collaborative combat capabilities in complex fire environments.

[0017] Optionally, when adjudicating bids for a task package, the dynamic auction coordination module comprehensively evaluates the overall estimated completion time, resource consumption, risk exposure, and contribution to the overall task completion of the task package, and calculates the corresponding bid utility value. The dynamic auction coordination module employs an improved auction algorithm with multiple rounds of iteration. In each round of auction, it dynamically adjusts task priorities and bidding constraints based on the allocation results of the previous round to improve overall allocation stability and convergence speed.

[0018] By adopting the above technical solutions, the task package adjudication process comprehensively considers the expected completion time, resource consumption, risk exposure, and contribution to the overall task completion, achieving a multi-dimensional and global evaluation of the bidding value of complex tasks. This avoids efficiency or security risks caused by allocating based on a single indicator. Furthermore, through multiple rounds of iterative improvement of the auction algorithm, the task priority and bidding constraints are dynamically adjusted, allowing the task allocation process to gradually converge to a stable solution, thereby improving the rationality, robustness, and overall execution efficiency of the allocation results.

[0019] Optionally, the elastic cluster communication and reassembly module supports at least the following three collaboration modes: Centralized auction model with full communication coverage; Grouped distributed auction mode under communication-constrained conditions; An autonomous collaborative model based on commitment mechanisms and local negotiation under conditions of communication interruption; In the fully centralized auction mode, the dynamic auction coordination module centrally collects bidding information for all equipment and uniformly completes task allocation decisions.

[0020] By adopting the above technical solutions, the collaborative strategy can be flexibly switched according to changes in the fire scene environment and communication conditions, ensuring that the task allocation mechanism can operate stably under different communication states; in particular, the centralized collection of bids and unified decision-making under the full communication centralized auction mode is conducive to achieving the optimal task allocation globally, improving the overall collaborative efficiency and operational reliability of the fire-fighting unmanned equipment cluster.

[0021] Optionally, in the grouped distributed auction mode under the communication-restricted conditions, the firefighting unmanned equipment is divided into several sub-clusters according to geographical location or task relevance, and each sub-cluster independently performs local auctions and completes task allocation.

[0022] By adopting the above technical solutions, the dependence on global communication bandwidth and latency is reduced, and the system's adaptability and robustness in complex fire environments are improved; it enables rapid response and efficient coordination of tasks in local areas, and avoids the failure of overall task allocation due to communication limitations.

[0023] Optionally, in the autonomous collaborative mode under the condition of communication interruption, each fire-fighting unmanned equipment autonomously decides to execute tasks based on pre-set commitment rules, task priorities and local situation information, and performs state synchronization after communication is restored.

[0024] By adopting the above technical solutions, each firefighting unmanned equipment can make independent decisions and execute tasks based on preset commitment rules, task priorities, and local situation information, thereby achieving continuous response to fire scene tasks. After communication is restored, status synchronization is performed to update task completion status and cluster distribution in a timely manner, ensuring the continuity and consistency of global task coordination and resource utilization, and improving the operational reliability and flexibility of firefighting unmanned equipment clusters in extreme communication-limited environments.

[0025] Optionally, the task dynamic release module can cancel, split, or reorganize the assigned tasks according to changes in the fire situation and trigger a new round of dynamic auction process. The firefighting unmanned equipment includes firefighting drones, firefighting unmanned vehicles, firefighting robots, or combinations thereof, and different types of firefighting unmanned equipment correspond to different capability parameters in the dynamic performance model.

[0026] By adopting the above technical solutions, task allocation can fully consider the characteristics and capabilities of equipment, thereby achieving high efficiency, flexibility and targetedness in cluster operations and improving the overall task completion effect in complex fire environments.

[0027] Optionally, the system further includes a task execution monitoring module, which monitors the task execution status of each fire-fighting unmanned equipment in real time and sends the execution feedback information to the dynamic auction coordination module to support task reallocation and cluster reconstruction.

[0028] By adopting the above technical solutions, dynamic perception of task execution can be achieved. The mechanism supports timely task reassignment and cluster reconstruction when tasks are delayed, equipment malfunctions or the fire situation changes, thereby improving the adaptability and operational continuity of equipment clusters and ensuring that the overall firefighting mission is efficient and reliable.

[0029] In summary, this application includes at least one of the following beneficial technical effects: Enhance the real-time and intelligent level of task planning and allocation: Based on the digital twin model of the fire scene, the task dynamic release module generates and updates fire-fighting tasks in real time, realizes dynamic perception of the fire situation, environmental changes and resource status, and makes task allocation more timely, accurate and targeted. Enhance equipment's autonomous decision-making and performance matching capabilities: The equipment intelligent agent module evaluates task bid values ​​based on a dynamic performance model. It can make distributed intelligent decisions by combining the equipment's current location, load, remaining energy, and operational capabilities, thereby achieving precise matching between tasks and equipment capabilities, avoiding resource waste or overload, and improving the efficiency and safety of cluster operations. Improve the collaborative execution capability of complex tasks: The dynamic auction coordination module packages tasks with coupling relationships and conducts multi-dimensional bidding evaluation based on estimated completion time, resource consumption, risk exposure and global task contribution to ensure reasonable overall task allocation and improve the collaboration efficiency and task completion rate of the cluster in complex fire environment. Achieving global optimization and conflict resolution: The multi-round improved auction algorithm dynamically adjusts task priorities and bidding constraints, so that task allocation gradually converges to the global optimal solution, reducing conflicts and duplicate operations, and improving the stability and reliability of overall task allocation; Enhance the robustness of the system under various communication environments: The elastic cluster communication and reorganization module supports centralized auctions for all communication, distributed auctions for communication-constrained groups, and autonomous collaborative modes under communication interruption, so as to achieve continuity and reliability of task allocation and execution under different communication states. Enhance the adaptability and flexibility of equipment clusters: The task execution monitoring module provides real-time feedback on the task execution status, providing a basis for dynamic reallocation and cluster reconstruction, enabling the cluster to quickly adjust its strategy in the event of equipment failure, task delays or changes in the fire situation, and ensuring continuous task execution. Supports collaborative operation of multiple types of firefighting unmanned equipment: The system is applicable to firefighting drones, unmanned vehicles, robots or combinations thereof, and performs dynamic performance calculations based on different equipment capability parameters to improve the overall operational efficiency of the cluster and its adaptability to complex fire scenes; Improve fire response efficiency and safety: By adopting an adaptive decision-making mechanism that comprehensively considers the urgency of the task, the risk level and changes in the fire situation, the equipment cluster can prioritize the response to high-risk tasks, thereby improving the speed of fire rescue and the safety of personnel and equipment. Reduce the risk of task allocation failure under communication constraints: Grouped distributed auction and autonomous collaborative modes can maintain the continuity of local or global operations under communication constraints or interruptions, and improve the system's fault tolerance to extreme environments; Supporting global resource optimization and collaborative decision-making: The dynamic auction coordination mechanism combines multi-dimensional evaluation, global utility maximization, and virtual alliance strategies to achieve reasonable allocation and collaborative scheduling of equipment resources, thereby improving the overall combat effectiveness of firefighting unmanned equipment clusters in complex fire environments. Attached Figure Description

[0030] Figure 1 This is a schematic diagram of the overall structure of the fire-fighting unmanned equipment cluster collaborative task allocation and control system according to an embodiment of this application.

[0031] Figure 2 This is a flowchart of the task dynamic publishing module.

[0032] Figure 3 This is a schematic diagram of the dynamic performance model calculation process for the intelligent agent module.

[0033] Figure 4 This is a schematic diagram of the task allocation process for the dynamic auction coordination module.

[0034] Figure 5 This is a schematic diagram illustrating the collaborative mode switching between the elastic cluster communication and reassembly module.

[0035] Figure 6 This is a flowchart illustrating the workflow of the task execution monitoring module.

[0036] Figure 7 This is a multi-round iterative auction process according to an embodiment of this application.

[0037] Figure 8 This is a system architecture diagram of an embodiment of this application. Detailed Implementation

[0038] The following is in conjunction with the appendix Figures 1-8 This application will be described in further detail.

[0039] This application discloses a collaborative task allocation and control system for fire-fighting unmanned equipment clusters based on a dynamic auction algorithm.

[0040] Reference Figures 1-8 The system includes the following modules: 1. Task Dynamic Publishing Module The task dynamic release module is used to acquire real-time situational information based on the fire scene digital twin model, generate and dynamically update a set of firefighting tasks with spatiotemporal constraints, resource constraints, and coupling relationships. During firefighting operations, fire situation information includes the location of the fire source, the speed of fire spread, temperature changes, smoke concentration, and the distribution of obstacles and personnel. Based on this information, the dynamic task release module can generate new firefighting tasks in real time, such as firefighting, search and rescue, patrol, or material delivery, and dynamically adjust the task allocation strategy according to task priority, time constraints, and equipment resource status to ensure the real-time nature and relevance of task planning. 2. Equip with intelligent agent module The equipment intelligence module is deployed on each piece of unmanned firefighting equipment. It is used to build a dynamic performance model based on the equipment's current location, remaining load, remaining energy, and operational capabilities, and to calculate the bid value for a task set or task package. The dynamic performance model can be expressed as: Where E_base_i represents the basic operational capability of the equipment, L_i(t) represents the load, B_i(t) represents the remaining energy, D(task,pos_i(t)) represents the distance or path cost from the current position of the equipment to the task position, and α, β, γ, and δ are weighting coefficients that can be adaptively adjusted by the dynamic auction coordination module according to the urgency of the task, the risk level, and changes in the fire situation, and then broadcast to each equipment agent module. Through this model, each piece of equipment can assess its own effectiveness in completing different tasks, providing a unified and comparable basis for task bidding, achieving precise matching between tasks and equipment capabilities, and improving task execution efficiency and overall cluster operation safety. 3. Dynamic Auction Coordination Module The dynamic auction coordination module is used to execute a multi-round improved auction algorithm, receive bidding information from various equipment agent modules, resolve conflicting bids, and complete task allocation based on the principle of maximizing global utility. For tasks with coupling relationships, such as firefighting or rescue missions that require the coordinated efforts of multiple pieces of equipment, the dynamic auction coordination module can combine multiple tasks with temporal or collaborative dependencies into a task package. This allows a single piece of equipment or multiple pieces of equipment to form a virtual alliance to bid on the task package as a whole. During the task package bidding adjudication process, the system comprehensively evaluates the expected completion time, resource consumption, risk exposure, and contribution to the overall task completion of the task package, calculates the bidding utility value, and dynamically adjusts the task priority and bidding constraints through a multi-round iterative algorithm to improve allocation stability and convergence speed. 4. Elastic Cluster Communication and Reassembly Module This module switches between different collaborative modes based on changes in communication status to ensure the collaborative operation capability of the firefighting unmanned equipment cluster, including the following three modes: Centralized auction mode with full communication: The dynamic auction coordination module centrally collects bidding information for all equipment, uniformly completes task allocation decisions, and achieves global optimization. Communication-constrained grouped distributed auction mode: Firefighting unmanned equipment is divided into several sub-clusters based on geographical location or task relevance. Each sub-cluster independently executes local auctions and completes task allocation, reducing dependence on global communication and improving system adaptability and robustness. Autonomous Collaborative Mode During Communication Interruption: Each firefighting unmanned equipment autonomously decides to execute tasks based on pre-set commitment rules, task priorities, and local situational information, and synchronizes its status after communication is restored to ensure the continuity of task execution and global consistency. Example 1 In Embodiment 1 of this technical solution, the following modules can be added: 5. Task Execution Monitoring Module The system further includes a task execution monitoring module, which monitors the task execution status of each firefighting unmanned equipment in real time and sends the execution feedback information to the dynamic auction coordination module. When there is a task delay, equipment failure, or change in the fire situation, the system can trigger task reassignment and cluster reconstruction to ensure task continuity and execution efficiency. 6. Types of unmanned firefighting equipment The aforementioned unmanned firefighting equipment includes firefighting drones, unmanned firefighting vehicles, firefighting robots, or combinations thereof. Different types of equipment correspond to different capability parameters in the dynamic performance model, ensuring that task allocation fully considers equipment characteristics and capability differences, thereby improving the efficiency and flexibility of cluster operations. Technical effects of Example 1: Through this example, the system can achieve: dynamic generation and intelligent allocation of tasks in complex fire environments; autonomous evaluation of equipment execution efficiency, realizing distributed intelligent bidding and reasonable scheduling; multi-dimensional global optimization of complex coupled tasks to improve collaborative combat capabilities; maintaining continuous cluster operation capability under different communication states to enhance system robustness and adaptability; supporting collaborative operation of multiple types of equipment to improve fire response speed and safety; dynamically monitoring task execution status to achieve real-time reallocation and cluster reconstruction, improving task completion efficiency and reliability.

[0041] Example 2 This embodiment addresses the application of the system in high-risk fire environments such as high-rise building fires or chemical fires, showcasing its use in multi-equipment joint firefighting operations. 1. System Configuration and Task Assignment The firefighting unmanned equipment cluster includes: Firefighting drones are used for high-altitude spraying of fire extinguishing agents, fire scene reconnaissance, and situational awareness. Firefighting unmanned vehicles are used for ground firefighting, transporting firefighting supplies, and supporting drone operations. Firefighting robots are used to enter high-temperature or toxic environments to perform firefighting, rescue, or detection tasks. The task dynamic release module updates task information in real time based on the fire scene digital twin model and generates task sets, including high-altitude spraying, fire extinguishing point control, rescue route planning, etc. Each task has time and space constraints, resource constraints and collaborative dependencies. The equipment intelligence agent module calculates a dynamic performance model based on the current location, load, remaining energy, and basic operational capabilities of each piece of equipment, and generates bid values. The dynamic auction coordination module makes overall bid decisions for complex task packages, taking into account all factors. 1.1 Estimated completion time; 1.2 Resource consumption (such as fire extinguishing dosage, power consumption); 1.3 Risk exposure level; 1.4 Contribution to overall task completion By iteratively improving the auction algorithm through multiple rounds, the global utility under high-risk fire conditions is maximized, thereby improving fire extinguishing efficiency and safety. 2. Communication Modes and Collaboration Strategies In high-risk fire zones, communication may be interfered with or partially interrupted. The resilient trunking communication and reassembly modules can flexibly switch according to the communication status. 2.1 Centralized Auction Mode with Full Communication: In areas with normal communication, the dynamic auction coordination module collects bids and completes task allocation in a unified manner, achieving global optimal scheduling; 2.2 Communication-Constrained Grouped Distributed Mode: The cluster is divided into sub-clusters, and each sub-cluster independently completes local auctions and task execution, ensuring rapid local response capabilities; 2.3 Autonomous Coordination Mode During Communication Interruption: Equipment independently executes tasks based on committed rules, task priorities, and local situational information, and synchronizes task status after communication is restored; The above strategies enable multi-equipment coordinated firefighting, continuous mission execution, and cluster stability in high-risk fire sites.

[0042] The technical effects of Example 2 are: it can improve the response speed and operational safety of high-risk fire scene tasks; it can realize the joint coordination of multiple types of equipment and make full use of the advantages of equipment capabilities; it can ensure the continuity of tasks and the adaptive capability of the cluster in the case of limited or interrupted communication; it can dynamically adjust the priority of task allocation, optimize the utilization of global resources, and improve fire fighting efficiency.

[0043] Example 3 This embodiment addresses fire scene scenarios where communication networks are severely interfered with or interrupted, showcasing the system's autonomous collaborative capabilities under extreme communication conditions. System configuration: The firefighting unmanned equipment cluster includes drones, unmanned vehicles, and robots, distributed across different areas of the fire scene to perform tasks. The task dynamic publishing module still generates task sets, but due to communication limitations, the equipment cannot access the global centralized auction. Autonomous Collaboration Strategy: Under communication interruption conditions, the elastic cluster communication and reassembly module switches to autonomous collaboration mode, and each piece of equipment operates according to: Preset commitment rules (task execution order, task priority); Local situation information (fire, obstacles, status of surrounding equipment); Calculation results of dynamic performance model; The equipment autonomously decides to execute tasks. After completing a task, it records its status in a local cache and synchronizes the cluster status after communication is restored to ensure global task coordination and consistency. The technical effects of Example 3 are: even without global communication, each piece of equipment can still complete its tasks autonomously, ensuring the continuity of fire scene tasks; making full use of local information to achieve rapid response and efficient collaboration; automatically synchronizing task status and resource information after communication is restored, ensuring global task optimization and cluster consistency; and improving the system's robustness and adaptability under extreme conditions to adapt to complex fire scene environments.

[0044] System architecture diagram of this application A [Perception and Task Generation Layer] A1 [Fire Scene Digital Twin Potential Map] A2 [Dynamic Task Pool] A3 [Task Value Evaluator] B [Core Control Layer - Dynamic Auction Engine] B1 [Auction Coordinator] B2 [Bidding Manager] B3 [Winning Decision and Conflict Resolution Tool] C[Intelligent Agent Layer - Unmanned Firefighting Equipment] C1 [Unmanned Aerial Vehicle - Reconnaissance Type] C2 [Unmanned Aerial Vehicle - Firefighting Type] C3 [Robot - Demolition Type] C4 [Robot - Transport Type] C5 [Status and Performance Assessment Module] C6 [Local Task Planner] D [Communication and Driver Layer] D1 [Anti-interference Mesh Network] D2 [Cluster Driver Interface] A -- Task list and value --> B B -- Auction Announcement / Winning Bid Instruction --> C C -- Bidding Information / Status Feedback --> B C -- Control command --> D D -- State data --> C The innovative aspects of this technical solution are: Core Innovation Point 1: Intelligent Bidding Mechanism Based on Multi-Dimensional Performance Evaluation Each piece of equipment is no longer a simple "receiver of instructions," but an intelligent agent with the ability to make autonomous bidding decisions.

[0045] 1. Equipment Dynamic Performance Model: The bidding capability of each piece of equipment `i` at time `t` is determined by its dynamic performance value `E_i(t)`: E_i(t) = α * (E_base_i) + β (1 - L_i(t) / L_max_i) + γ (B_i(t) / B_max_i) - δ D(task, pos_i(t)) `E_base_i`: Base capability coefficient (e.g., water-based payload capacity of firefighting drones).

[0046] `L_i(t) / L_max_i`: Load consumption ratio (e.g., remaining extinguishing agent ratio).

[0047] `B_i(t) / B_max_i`: Remaining energy ratio.

[0048] `D(task, pos_i(t))`: Normalized distance cost to reach the task point.

[0049] `α, β, γ, δ`: Dynamic weights, adjusted by the coordinator based on the urgency of the task.

[0050] 2. Intelligent Bidding Algorithm (Equipment-side Pseudocode): class FirefightingAgent: defbid_for_task(self,task_broadcast,current_pos,self_status): (1). Self-check of task feasibility ifnotself.can_perform_task(task_broadcast['type']): return None (2). Calculate the multidimensional performance value efficiency=self.calculate_efficiency(task_broadcast, current_pos) (3). Introduce "expected completion rate" and "opportunity cost" expected_completion self.estimate_completion_rate(task_broadcast, self_status) opportunity_cost = self.calculate_opportunity_cost(task_broadcast) If the original task is abandoned... (4). Generate the comprehensive bid value bid_value = efficiency * expected_completion - opportunity_cost (5). Additional tender information: estimated start time, completion time, and required support. bid_package = { 'agent_id': self.id, 'task_id': task_broadcast['id'], 'bid_value': bid_value, 'est_start': current_time + travel_time, 'est_finish': ..., 'required_support': [...] such as requiring reconnaissance aircraft guidance for firefighting. } return bid_package Innovation Point Two: Multi-round Improved Auction Considering Task Coupling and Dependencies Traditional auctions treat tasks as independent commodities, while this invention innovatively addresses the spatiotemporal coupling between tasks.

[0051] 1. Task relationship diagram modeling: Sequence dependency: The rescue robot cannot enter until the demolition task is completed.

[0052] Synergy dependency: Reconnaissance provides target guidance for firefighting, and the two must be synchronized.

[0053] Mutual exclusion: Two tasks compete for the same resource.

[0054] 2. Multi-round iterative auction process: mermaid flowchart TD A [Coordinator publishes new / updated task] --> B {Is this a coupled task?} B -- No --> C [Single-Task Standard Auction] B -- Yes --> D [Coupled Task Package Auction] C --> E [Each piece of equipment is calculated and submitted for bidding independently] D --> F [Bidding as a "virtual alliance" member after teaming up with equipment] E --> G [Coordinator collects all bids] F --> G G --> H { Trigger rebid conditions?} H -- is If the winning bid equipment suddenly malfunctions --> I[Partial Re-auction] [Notify only affected task stakeholders] H -- No --> J [Declare victory and generate final allocation] I --> J J --> K [Issue winning command and global path planning] 3. Auction coordinator adjudication algorithm (key segment): class AuctionCoordinator: def resolve_bids(self, all_bids, task_graph): allocation = {} Task -> Equipment Allocation (1) Preliminary sorting by bid value sorted_bids = self.sort_bids_by_value(all_bids) (2). Conflict detection and resolution (considering task dependencies) for task in task_graph.tasks: candidate_agents self.get_candidates_for_task(sorted_bids, task) best_agent = None best_global_utility = -inf for agent in candidate_agents: Temporary allocation, calculating global utility (considering the completion time of dependent task chains). temp_alloc = allocation.copy() temp_alloc[task] = agent utility = self.calculate_global_utility(temp_alloc, task_graph) Introduce a "marginal contribution" assessment, rather than simply looking at the bid value. marginal_gain = utility - current_global_utility if marginal_gain > best_global_utility: best_global_utility = marginal_gain best_agent = agent (3). Confirm the allocation and lock the relevant resources. if best_agent: allocation[task] = best_agent self.update_resource_locks(task, best_agent, task_graph) return allocation Core Innovation Point 3: Elastic Cluster Reassembly Protocol Under Communication Constraints To address the extreme situation of unstable communication at fire sites, a tiered and degraded collaborative strategy is designed.

[0055] 1. Three-level communication adaptive mode: Communication status Collaborative mode Decision-making methods Information Interaction good Centralized and optimized auction Coordinator Global Decision Full status, full bidding Restricted Grouped Distributed Auction Intra-group auctions, inter-group coordinator negotiations Compression status. Abstract tender. Interruption Commitment-based autonomous collaboration Final instructions and preset rule actions Local broadcast beacon signal 2. Flexible Reorganization Protocol (Pseudocode Logic): When equipment i detects a communication interruption with the coordinator exceeding the threshold T: (1) Switch to “Autonomous Collaboration Mode”.

[0056] (2) Broadcast its own status and current task within the local communication range.

[0057] (3) Receive broadcasts from nearby equipment to form a temporary local view.

[0058] (4) If you cannot complete your task alone, try to negotiate with nearby equipment to form a temporary team.

[0059] (5) Based on the shared local view, run a simplified version of the auction algorithm and redistribute tasks within the group.

[0060] (6) Once communication is restored, the local decision and status are immediately synchronized to the coordinator, which then performs global calibration.

[0061] Application scenario of this technical solution: Fire rescue in high-rise buildings. Task sequence: 1.T0: Fire confirmed, initial task pool generated: {Detect fire source, scout personnel, shut off gas valve, spray water to cool the surrounding area}.

[0062] 2.T1: First round of auction. Reconnaissance UAV A (high efficiency) won the bid for "Reconnaissance of fire source" and "Reconnaissance personnel" (coupled mission package); Firefighting UAV B won the bid for "Outer perimeter water spray".

[0063] 3.T2: Reconnaissance aircraft A discovers people trapped on the 4th floor, and the fire is spreading to the 5th floor. Dynamically insert new mission: `{Accurate personnel location, contain the fire on the 5th floor, deliver breathing masks}`.

[0064] 4.T3: Second round of auction. The coordinator broadcasts a new mission. Drone B, originally tasked with "peripheral water spraying," calculates that its remaining extinguishing agent is better suited for "intercepting the fire on the 5th floor," and its bid is higher. Therefore, it requests a mission change. The coordinator approves and initiates a re-auction for the "peripheral water spraying" mission, which is won by the newly arrived drone C.

[0065] 5.T4: The demolition robot needs to enter, but the passage is blocked. The coordinator issues a "Demolition Entrance" task, and the fire-fighting drone D and the demolition robot E form a "virtual alliance" to jointly bid (extinguish the fire and cool down first, then demolish).

[0066] Performance comparison experimental data of this technical solution On the simulation platform, compared with traditional methods (task completion rate, average task latency): Scene Scale Comparison Algorithm Task completion rate Delay Communication load 10 pieces of equipment, 20 dynamic quests Centralized MILP 85% high Extremely high 10 pieces of equipment, 20 dynamic quests Traditional Contract Network 78% middle high 10 pieces of equipment, 20 dynamic quests Method of the present invention 96% Low middle 20 pieces of equipment, 50 dynamic quests Centralized MILP Unable to be real-time 20 pieces of equipment, 50 dynamic quests Method of the present invention 93% Low-medium middle The technical solution of this application has the following technical advantages: 1. Originality: The dynamic multi-round improved auction mechanism is deeply adapted to the highly dynamic, strongly coupled, and highly real-time scenarios of fire emergency rescue, solving the problems of inflexibility of traditional optimization methods and instability of swarm intelligence methods.

[0067] 2. Novelty: The "Equipment Dynamic Performance Model" and "Task Coupling Relationship Diagram" are proposed as the core inputs of the auction algorithm, making the allocation results more in line with the physical and logical constraints of actual fire fighting.

[0068] 3. Creativity: An adaptive elastic reconfiguration protocol for communication was designed to ensure the robustness of the system in extreme and harsh environments, which is an engineering challenge that is rarely considered in existing laboratory collaborative algorithms.

[0069] 4. Operability: The system adopts a hybrid architecture that combines distributed and centralized approaches. The algorithm module can be embedded into the flight control or mission computer of existing unmanned equipment. Through standard anti-jamming mesh network communication, it is easy to upgrade and modify existing equipment.

[0070] The implementation principle of a collaborative task allocation and control system for fire-fighting unmanned equipment clusters based on dynamic auction algorithm in this application embodiment is as follows: After a fire occurs, the system first accesses the fire scene digital twin model through the task dynamic release module to perceive the fire development trend, changes in environmental conditions, and the distribution of various fire-fighting resources in real time. On this basis, a fire-fighting task set with clear spatiotemporal constraints, resource constraints, and coupling relationships is generated, and the tasks are dynamically updated as the fire scene situation changes. The equipment intelligent agent module on each fire-fighting unmanned equipment obtains its current position, remaining load, remaining energy, and operational capabilities, etc. Based on the constructed dynamic efficiency model, the execution efficiency of each task or task package is evaluated, a corresponding bid value is formed, and it is sent to the dynamic auction coordination module. After receiving bidding information from multiple pieces of equipment, the dynamic auction coordination module first identifies tasks with temporal or collaborative dependencies and combines them into task packages. A multi-round improved auction algorithm is then used to adjudicate bids and resolve conflicts. During the auction process, the module comprehensively considers task completion time, resource consumption, risk exposure, and contribution to overall firefighting efficiency, aiming to maximize global utility in task allocation. The auction coordination module dynamically adjusts the weight parameters in the effectiveness model based on task urgency, risk level, and changes in the fire situation, and broadcasts this information through the cluster communication module to ensure consistency in bidding decisions among all equipment. During the mission execution phase, the elastic cluster communication and reorganization module adaptively switches between collaborative modes such as centralized auction, grouped distributed auction, or autonomous collaboration based on the communication status, ensuring that the cluster can continue to work collaboratively even under conditions of limited or interrupted communication. The mission execution monitoring module monitors the execution status of each piece of equipment in real time and transmits the feedback information back to the dynamic auction coordination module to trigger mission reallocation or cluster reconstruction. Through the above collaborative mechanisms, efficient, adaptive, and robust mission allocation and collaborative control of the firefighting unmanned equipment cluster in complex fire scene environments are achieved.

[0071] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A collaborative task allocation and control system for firefighting unmanned equipment clusters based on a dynamic auction algorithm, characterized in that, include: The task dynamic release module is used to generate and dynamically update a set of fire-fighting tasks with spatiotemporal constraints, resource constraints and coupling relationships based on real-time situation information obtained from the fire scene digital twin model. The equipment intelligent agent module is deployed on each fire-fighting unmanned equipment to build a dynamic performance model based on the current position, remaining load, remaining energy and operation capability of the fire-fighting unmanned equipment, and to calculate the bid value for the task set or task package accordingly. The dynamic auction coordination module is used to execute multi-round improved auction algorithms, receive bidding information from various equipment intelligent agent modules, resolve conflicting bids, and complete task allocation based on the principle of maximizing global utility. The flexible cluster communication and reorganization module is used to switch between different collaborative modes according to changes in communication status, and to maintain the collaborative operation capability of the fire-fighting unmanned equipment cluster in the event of communication restrictions or interruptions.

2. The system according to claim 1, characterized in that, The dynamic performance model in the equipment intelligent agent module is represented as follows: Where EbaseiE_{base_i}Ebasei represents the basic operational capability of the equipment, Li(t) represents the load, Bi(t) represents the remaining energy, and D(task,posi(t) represents the distance or path cost from the current position of the equipment to the mission position.

3. The system according to claim 2, characterized in that, The weighting coefficients α, β, γ, and δ are adaptively adjusted by the dynamic auction coordination module according to the urgency of the task, the risk level, and changes in the fire situation, and broadcast to each equipment intelligent agent module through the cluster communication module.

4. The system according to claim 1, characterized in that, The method for the dynamic auction coordination module to handle tasks with coupled relationships includes: Identify multiple tasks that have temporal or cooperative dependencies; The multiple tasks are combined to form a task package; Allows a single firefighting drone or a virtual alliance of multiple firefighting drones to... The task package is submitted as a whole for bidding.

5. The system according to claim 4, characterized in that, When adjudicating bids for a task package, the dynamic auction coordination module comprehensively evaluates the overall estimated completion time, resource consumption, risk exposure, and contribution to the overall task completion of the task package, and calculates the corresponding bid utility value. The dynamic auction coordination module employs an improved auction algorithm with multiple rounds of iteration. In each round of auction, it dynamically adjusts task priorities and bidding constraints based on the allocation results of the previous round to improve overall allocation stability and convergence speed.

6. The system according to claim 1, characterized in that, The elastic cluster communication and reassembly module supports at least the following three collaboration modes: Centralized auction model with full communication coverage; Grouped distributed auction mode under communication-constrained conditions; An autonomous collaborative model based on commitment mechanisms and local negotiation under conditions of communication interruption; In the fully centralized auction mode, the dynamic auction coordination module centrally collects bidding information for all equipment and uniformly completes task allocation decisions.

7. The system according to claim 6, characterized in that, In the grouped distributed auction mode under the aforementioned communication-restricted conditions, firefighting unmanned equipment is divided into several sub-clusters according to geographical location or task relevance. Each sub-cluster independently executes a local auction and completes task allocation.

8. The system according to claim 6, characterized in that, In the autonomous collaborative mode under the condition of communication interruption, each fire-fighting unmanned equipment autonomously decides to execute tasks based on pre-set commitment rules, task priorities and local situation information, and performs state synchronization after communication is restored.

9. The system according to claim 1, characterized in that, The task dynamic release module can cancel, split, or reorganize assigned tasks according to changes in the fire situation and trigger a new round of dynamic auction process. The firefighting unmanned equipment includes firefighting drones, firefighting unmanned vehicles, firefighting robots, or combinations thereof, and different types of firefighting unmanned equipment correspond to different capability parameters in the dynamic performance model.

10. The system according to any one of claims 1 to 9, characterized in that, The system further includes a task execution monitoring module, which is used to monitor the task execution status of each fire-fighting unmanned equipment in real time and send the execution feedback information to the dynamic auction coordination module to support task reallocation and cluster reconstruction.

Images