A method and system for processing path conflict of a drone cluster

Abstract

The application discloses a kind of unmanned aerial vehicle cluster path conflict processing method and system, belong to unmanned aerial vehicle flight control technical field, the method includes: obtaining the numbering of all unmanned aerial vehicles in cluster, position and other key data and available space resources in region;Then generate cluster member state table and regional available resource vector;Subsequently, collect the single-cluster conflict information of conflict unmanned aerial vehicles in cluster and generate single-cluster conflict information table, while obtaining cross-cluster conflict information by multi-cluster attribution non-cluster head unmanned aerial vehicle and generating multi-cluster conflict information table;If it is determined that there is cross-cluster conflict, generate multiple single-cluster resolution strategies and corresponding cost based on single-cluster conflict information table;Finally, combine multi-cluster conflict information table and single-cluster resolution cost, cluster head and target conflict cluster cluster head carry out distributed game negotiation, and finally obtain global conflict resolution optimal strategy.Through the application, the problem of poor real-time performance and poor adaptability of unmanned aerial vehicle conflict resolution in the prior art can be solved.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) flight control technology, and in particular to a method and system for handling path conflicts in UAV swarms. Background Technology

[0002] With the rapid development of drone technology, its application scenarios are becoming increasingly diverse, evolving from single drone operations to large-scale drone swarm operations. For example, in urban drone logistics delivery scenarios, hundreds or even thousands of drones need to shuttle through urban airspace to complete transportation tasks from warehousing centers to various distribution points; while in the field of agricultural plant protection, large-scale drone swarms can simultaneously and precisely spray large areas of farmland, significantly improving operational efficiency. However, large-scale drone swarm flights face a severe path conflict problem: when a large number of drones fly in the same airspace, factors such as unreasonable flight path planning, dynamic changes in the airspace environment (such as sudden weather, temporary no-fly zones), and fluctuations in the drones' own flight status (such as speed changes, heading deviations) can easily lead to conflicts such as intersecting flight paths and excessively close distances. If these conflicts are not resolved in a timely and effective manner, they may cause collisions, resulting in damage to drones, loss of cargo, and even threatening the safety of ground personnel and facilities.

[0003] Existing technologies mainly include centralized path planning methods and rule-based conflict resolution methods. Centralized path planning methods use a central control node to uniformly acquire flight information from all UAVs, determine conflicts based on preset conflict detection rules, and replan paths for the conflicting UAVs. While this method can achieve theoretically optimal conflict resolution from a global perspective, in large-scale scenarios, the central node needs to process massive amounts of data, resulting in poor real-time performance. Furthermore, if the node fails, the flight safety of the entire cluster will be severely affected, leading to low system reliability. Rule-based conflict resolution methods, on the other hand, pre-define a series of conflict resolution rules, such as priority rules (determining right-of-way based on mission importance or flight altitude) or avoidance rules (e.g., left-side avoidance, deceleration avoidance). When a conflict is detected, the UAV performs avoidance operations according to the rules. These methods are logically simple and easy to implement, but in large-scale scenarios, preset rules cannot cover all complex conflict situations, potentially leading to resolution failure. Moreover, fixed rules lack flexibility and cannot adjust strategies according to dynamic changes in the airspace environment and flight status, resulting in poor adaptability. Summary of the Invention

[0004] This invention provides a method and system for handling path conflicts in a drone swarm, which can solve the problems of poor real-time performance and poor adaptability in existing drone conflict resolution technologies.

[0005] An embodiment of the present invention provides a method for handling path conflicts in a drone swarm, comprising: Obtain the ID, location, heading, speed, cumulative flight distance, mission priority, mission deadline, space resource requirements, and available space resources within the cluster for all UAVs. A cluster member status table is generated based on the number, location, heading, speed, cumulative flight distance, task priority, task deadline, and spatial resource requirements of all UAVs within the cluster. A regional available resource vector is generated based on the available spatial resources within the region. Obtain the single-cluster conflict information uploaded by all conflicting UAVs within the cluster, and generate a single-cluster conflict information table based on the single-cluster conflict information. Obtain cross-cluster conflict information uploaded by non-cluster head UAVs belonging to multiple clusters, and generate a multi-cluster conflict information table based on the cross-cluster conflict information; If, based on the multi-cluster conflict information table, it is determined that there is a cross-cluster conflict, then based on the single-cluster conflict information table, several sets of single-cluster conflict resolution strategies and the single-cluster resolution cost of each corresponding single-cluster conflict resolution strategy are generated. Based on the multi-cluster conflict information table and the single-cluster conflict resolution cost of each single-cluster conflict resolution strategy, a distributed game negotiation is conducted with the cluster head drone of the target conflict drone cluster that has cross-cluster conflicts to obtain the global optimal conflict resolution strategy.

[0006] Furthermore, the cluster leader drone is determined after election calculation based on the communication coverage, remaining battery power, and historical success rate of conflict resolution of the conflicting drones in the conflicting drone cluster; wherein, the conflicting drone cluster is formed when the drone density in a local area is not lower than a preset density value, and the real-time distance between any two drones in the local area is not greater than a preset distance.

[0007] Next, the conflicting drones are defined as any two drones whose spatial resource needs overlap, or whose spatial resource distance at the same time is no greater than a preset spatial resource distance threshold.

[0008] Furthermore, the cross-cluster conflict information is determined in the following way: Non-cluster-head UAVs belonging to multiple clusters compare the spatial resource requirements and available resource vectors in the status tables of members within different clusters. If there are duplicate comparison results, cross-cluster conflict information is obtained.

[0009] Next, based on the multi-cluster conflict information table, if cross-cluster conflicts are determined to exist, several sets of single-cluster conflict resolution strategies and the single-cluster resolution cost of each corresponding single-cluster conflict resolution strategy are generated according to the single-cluster conflict information table, including: Based on the single-cluster conflict information table, Cartesian product is used to iteratively adjust each UAV in its own cluster to generate several sets of single-cluster conflict resolution strategies. The resolution cost of each group of single-cluster conflict resolution strategies is calculated using the following formula, thus obtaining the single-cluster resolution cost corresponding to each group of single-cluster conflicts: ; ; ; in, This indicates the cost of resolving conflicts in a single cluster. Indicates the energy efficiency improvement factor; Indicates the time delay rate; This indicates the cumulative flight distance after strategy adjustment; This indicates the cumulative flight distance before the strategy adjustment; Indicates the task completion time after the strategy adjustment; Indicates the task deadline.

[0010] Furthermore, based on the multi-cluster conflict information table and the single-cluster conflict resolution cost of each single-cluster conflict resolution strategy, a distributed game negotiation is conducted with the cluster leader drone of the target conflict drone cluster with cross-cluster conflicts to obtain the global optimal conflict resolution strategy, including: Based on the multi-cluster conflict information table, target conflicting UAV clusters with cross-cluster conflicts are identified; Calculate the single-cluster conflict resolution cost of each single-cluster conflict resolution strategy within its own cluster, and select the single-cluster conflict resolution strategy with the minimum single-cluster conflict resolution cost as the first target single-cluster conflict resolution strategy. Based on the cluster member state table and the first target single cluster conflict resolution strategy, generate the first cluster UAV trajectory table corresponding to the first target single cluster conflict resolution strategy. The trajectory table of UAVs in the first cluster is compared with the trajectory table of UAVs in the second cluster of the target conflict UAV cluster. The trajectory table of UAVs in the second cluster is generated by the cluster leader UAV of the target conflict UAV cluster based on the second target single cluster conflict resolution strategy with the lowest single cluster resolution cost corresponding to the target conflict UAV cluster. When there is no overlap between the UAV trajectory tables in the first cluster and the UAV trajectory tables in the second cluster, the optimal global conflict resolution strategy is generated based on the conflict resolution strategies for the first and second objectives in a single cluster.

[0011] Next, if there is an overlap between the drone trajectory tables in the first cluster and the drone trajectory tables in the second cluster, the overall task priority of the cluster to which the drone belongs is compared with the overall task priority of the cluster of drones that are in conflict with the target. When the overall task priority of its own cluster is lower than the overall task priority of the target conflicting UAV cluster, the single-cluster conflict resolution strategy of its own cluster is updated according to the single-cluster conflict information table of its own cluster and the UAV trajectory table in the second cluster, resulting in several updated single-cluster conflict resolution strategies. Calculate the single-cluster resolution cost of each updated single-cluster conflict resolution strategy, and select the updated single-cluster conflict resolution strategy with the lowest single-cluster resolution cost as the first updated target single-cluster conflict resolution strategy. Based on the first updated target single-cluster conflict resolution strategy and the second target single-cluster conflict resolution strategy, the optimal global conflict resolution strategy is generated.

[0012] Furthermore, the overall task priority is determined using the following formula: ; ; ; ; in, Indicates the task priority of the cluster; This represents the set of drones within the cluster that are currently involved in the conflict; Indicates drone Task priority; This represents the weighting coefficient of drone k; Indicates the task completion rate; Indicates the remaining battery power; Factors indicating the severity of conflict; This represents the weighting coefficient for task completion. This represents the weighting coefficient for the cost of remaining electricity. Indicates the weighting coefficients representing the severity of the conflict; Indicates the current closest distance to other drones; Indicates the collision distance threshold; Indicates the remaining time expected for a collision; Indicates the warning time threshold; An embodiment of the present invention also provides a UAV swarm path conflict handling system, including: a cluster leader UAV within a conflicting UAV cluster, non-cluster leader UAVs belonging to multiple clusters, and non-cluster leader UAVs belonging to a single cluster; The non-cluster-head UAVs belonging to multiple clusters and the non-cluster-head UAVs belonging to a single cluster are used to transmit the UAV's number, location, heading, speed, cumulative flight distance, mission priority, mission deadline, and space resource requirements to the cluster-head UAV. The cluster leader UAV is used to obtain the number, location, heading, speed, cumulative flight distance, mission priority, mission deadline, space resource requirements, and available space resources in the area of ​​all UAVs in the conflicting UAV cluster. Obtain the single-cluster conflict information uploaded by all conflicting UAVs within the cluster, and generate a single-cluster conflict information table based on the single-cluster conflict information. Obtain cross-cluster conflict information uploaded by non-cluster head UAVs belonging to multiple clusters, and generate a multi-cluster conflict information table based on the cross-cluster conflict information; If, based on the multi-cluster conflict information table, it is determined that there is a cross-cluster conflict, then based on the single-cluster conflict information table, several sets of single-cluster conflict resolution strategies and the single-cluster resolution cost of each corresponding single-cluster conflict resolution strategy are generated. Based on the multi-cluster conflict information table and the single-cluster conflict resolution cost of each single-cluster conflict resolution strategy, a distributed game negotiation is conducted with the cluster head drone of the target conflict drone cluster that has cross-cluster conflicts to obtain the global optimal conflict resolution strategy.

[0013] Preferably, the cluster leader drone is determined after election calculation based on the communication coverage, remaining battery power, and historical success rate of conflict resolution of the conflicting drones in the conflicting drone cluster; wherein, the conflicting drone cluster is formed when the drone density in a local area is not lower than a preset density value, and the real-time distance between any two drones in the local area is not greater than a preset distance.

[0014] The following benefits can be obtained by implementing the present invention: This invention first obtains the ID, location, heading, speed, cumulative flight distance, task priority, task deadline, spatial resource requirements, and available spatial resources within a cluster for all UAVs. Based on the aforementioned information of all UAVs within the cluster, a cluster member status table is generated, and an available regional resource vector is generated based on the available spatial resources within the region. Next, it obtains single-cluster conflict information uploaded by all conflicting UAVs within the cluster, and cross-cluster conflict information uploaded by non-cluster head UAVs belonging to multiple clusters. A single-cluster conflict information table is generated based on each single-cluster conflict information, and a multi-cluster conflict information table is generated based on each cross-cluster conflict information. Then, when cross-cluster conflicts are determined based on the multi-cluster conflict information table, several sets of single-cluster conflict resolution strategies and their corresponding single-cluster resolution costs are generated by combining the single-cluster conflict information tables. Finally, based on the multi-cluster conflict information table and the single-cluster resolution costs of each single-cluster conflict resolution strategy, a distributed game negotiation is conducted with the cluster head UAV of the target conflicting UAV cluster that has cross-cluster conflicts, thereby obtaining the optimal global conflict resolution strategy. By dynamically clustering large-scale clusters, the state and inter-cluster interactions are managed by cluster head UAVs, reducing the processing pressure on the central node, improving real-time performance and robustness, and obtaining the globally optimal strategy through distributed game negotiation, thus adapting to the dynamic changes in the airspace environment. Attached Figure Description

[0015] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating a method for handling path conflicts in a drone swarm, as provided in an embodiment of this application.

[0017] Figure 2 This is a system architecture diagram of a drone swarm path conflict handling system provided in one embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0020] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0021] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0022] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0023] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0024] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.

[0025] See Figure 1To address the issues of poor real-time performance and poor adaptability in existing drone conflict resolution technologies, an embodiment of the present invention provides a method for handling drone swarm path conflicts, comprising: S1. Obtain the ID, location, heading, speed, cumulative flight distance, task priority, task deadline, space resource requirements, and available space resources within the cluster for all UAVs. S2. Generate a cluster member status table based on the number, position, heading, speed, cumulative flight distance, task priority, task deadline, and spatial resource requirements of all UAVs in the cluster, and generate a regional available resource vector based on the available spatial resources in the region. Preferably, the cluster leader drone is determined after election calculation based on the communication coverage, remaining battery power, and historical success rate of conflict resolution of the conflicting drones in the conflicting drone cluster; wherein, the conflicting drone cluster is formed when the drone density in a local area is not lower than a preset density value, and the real-time distance between any two drones in the local area is not greater than a preset distance.

[0026] Specifically, the formation of conflict drone clusters has clear conditions. It requires simultaneously meeting two requirements: a drone density of at least two drones per 1000 square meters in a local area, and a real-time distance between any two drones within that local area not exceeding a preset distance. The preset distance threshold is 30 meters in normal scenarios and dynamically adjusted to 50 meters in high-speed flight scenarios. Once these conditions are met, dynamic clustering is triggered, initiating the cluster head election process. The cluster head drone is determined through a scientific calculation using a three-dimensional weighting method. The election is based on three core indicators: the communication coverage range, remaining battery power, and historical conflict resolution success rate of each drone in the conflict drone cluster. Communication coverage range accounts for 30%, remaining battery power for 40%, and historical conflict resolution success rate for 30%. The drone with the highest weighted score is selected as the cluster head.

[0027] It is necessary to comprehensively collect the core information of all UAVs within the cluster and the available space resources within the area. The core information of the UAVs includes their number, location, heading, speed, cumulative flight distance, task priority, task deadline, and space resource requirements. Among them, the task priority is a fixed attribute that is pre-set before the UAV executes the corresponding task, which is used to define the scheduling order of different tasks. The task deadline is the latest time limit for completing the task, which serves as a constraint on UAV task scheduling and space resource occupancy planning. The space resource requirements specifically include horizontal airspace blocks, vertical altitude layers, and occupancy time. The available space resources within the area are specifically real-time data acquired every 5 seconds by the cluster head through 4G / 5G communication with the area resource management node. This data covers available horizontal airspace blocks, vertical altitude layers, available time periods, and available duration, providing a foundation for subsequent data processing.

[0028] Based on the collected data on all UAV numbers, locations, headings, speeds, and cumulative flight distances within the cluster, as well as preset task priorities, deadlines, and spatial resource requirements including horizontal airspace blocks, vertical altitude layers, and occupancy times, a cluster member status table is constructed after unified organization and standardization. This status table is distributed across the entire network by the cluster head using UWB ultra-wideband communication, with communication latency controlled within 100ms, and is updated in real-time at a 1-second interval to ensure high consistency of information for all UAV members within the cluster. Simultaneously, based on real-time collected information on available spatial resources within the region, various resource parameters such as available vertical altitude layers and horizontal airspace blocks are first classified and quantified, converting non-standardized resource information into calculable and comparable numerical forms. Based on this, a standardized regional available resource vector is generated. The available resource vector for this region fully embodies the quantitative characteristics of regional airspace resources, specifically reflecting the coverage and remaining available capacity of each vertical height layer, the coordinate boundaries and available area of ​​horizontal airspace blocks, and incorporating the dimensions of available airspace time period and available duration. The available resource vector is synchronously updated every 5 seconds by the cluster head pulling the latest airspace data from the regional resource management node via 4G / 5G communication. It can accurately and in real time characterize the remaining available capacity, spatial distribution range, and effective available duration of allocable airspace resources in the entire region. Subsequently, it can be directly linked and compared with the UAV spatial resource requirements recorded in the status tables of members within each cluster for supply and demand verification. This provides a real-time and accurate data foundation for the rational allocation of resources in multi-cluster collaborative games, ensuring that the overall game decision-making is scientific and rational.

[0029] This invention scientifically sets dynamic clustering conditions, uses a three-dimensional weighting method to elect the optimal cluster head, and relies on the cluster head to complete core operations such as information collection, data processing, and resource updates. This effectively ensures the consistency of information within the cluster and the real-time nature of regional airspace resource data, providing accurate and reliable data support for resource allocation in multi-cluster collaborative games. It significantly improves the scientificity and rationality of game decisions, realizes the orderly advancement of intra-cluster management and multi-cluster collaboration, effectively avoids airspace resource allocation conflicts, greatly improves the stability, efficiency, and safety of UAV swarm flight, and further optimizes the operational efficiency of UAV swarms in various scenarios.

[0030] S3. Obtain the single-cluster conflict information uploaded by all conflicting UAVs within the cluster, and generate a single-cluster conflict information table based on the single-cluster conflict information. Preferably, the conflicting drones are any two drones whose spatial resource needs overlap, or whose spatial resource distance at the same time is no greater than a preset spatial resource distance threshold.

[0031] Specifically, the determination of conflicting drones follows the intra-cluster conflict detection logic. This refers to any two drones within the same cluster, both being cluster members (including the cluster head and ordinary drones). First, the cluster head's distributed intra-cluster member status table is obtained, then the spatial resource requirements of both drones are extracted to determine the conflict. A drone is considered a conflicting drone if one of the following two conditions is met: First, their spatial resource requirements overlap, meaning both drones occupy the same horizontal airspace and vertical height resources at the same time, resulting in overlapping resource usage and thus intra-cluster conflict. Second, at the same time, the spatial resources occupied by both drones are adjacent, and the distance between adjacent resources is no greater than a preset threshold. This threshold is typically 15 meters and can be dynamically adjusted according to drone size. For large drones, the threshold is set to 20 meters; in this case, even if they do not occupy the same resources, it is still considered intra-cluster conflict.

[0032] After determining the conflicting drones, the cluster head receives various conflict data uploaded in real time by all drones in the cluster that have autonomously identified flight conflicts within the cluster via the real-time status link within the cluster. Each reported data includes key information such as the number of the drone and the target drone, the estimated conflict location, the estimated conflict time, the heading, and the speed. The cluster head continuously summarizes this reported information and organizes it according to a unified standard, continuously collecting and sorting out all conflict entries within the cluster and completing standardized integration processing, thereby forming a complete and standardized single-cluster conflict information table.

[0033] This invention employs intra-cluster conflict detection logic to autonomously determine UAV conflicts. Combining the dual criteria of overlapping airspace resources and distance constraints between adjacent resources, it accurately identifies flight conflict behaviors between the cluster head and ordinary UAVs. Simultaneously, it leverages the real-time status link within the cluster to rapidly collect core data reported by conflicting UAVs. After standardization and integration, it generates a single-cluster conflict information table, which can comprehensively, accurately, and efficiently analyze various flight conflict situations within the cluster, avoid the risk of close-range collisions between UAVs within the cluster, and provide complete and reliable data support for multi-cluster collaborative game decision-making and subsequent conflict resolution. This effectively improves the safety and orderliness of UAV swarm flight operations.

[0034] S4. Obtain cross-cluster conflict information uploaded by non-cluster head UAVs belonging to multiple clusters, and generate a multi-cluster conflict information table based on the cross-cluster conflict information. Preferably, the conflicting drone cluster also includes: non-cluster leader drones belonging to multiple clusters; Preferably, the cross-cluster conflict information is determined in the following way: Non-cluster-head UAVs belonging to multiple clusters compare the spatial resource requirements and available resource vectors in the status tables of members within different clusters. If there are duplicate comparison results, cross-cluster conflict information is obtained.

[0035] Specifically, within a conflict-ridden drone cluster, in addition to regular cluster members, there are also non-cluster head drones belonging to multiple clusters. These drones belong to multiple clusters simultaneously and function as both information relays and conflict detection systems. During cross-cluster information exchange, non-cluster head drones belonging to multiple clusters act as information relay nodes, forwarding their own cluster member status tables (which have been anonymized and retain only cross-cluster conflict-related data) to the cluster heads of the other clusters they belong to every 2 seconds. This constructs a multi-cluster information exchange network, providing a data foundation for cross-cluster conflict detection.

[0036] The determination of cross-cluster conflict information is accomplished by non-cluster head UAVs with multiple cluster affiliations: Assuming X is a non-cluster head UAV with multiple cluster affiliations and responsible for cross-cluster conflict comparison and judgment, Y and Z are two UAVs belonging to different clusters and having cross-cluster resource conflicts with each other. X is only responsible for global data retrieval and conflict assessment, and does not necessarily participate in the conflict relationship between Y and Z. A non-cluster leader UAV X with multi-cluster affiliation is responsible for comparing and judging cross-cluster conflicts. X retrieves the status tables of the members in each of its respective clusters and extracts the spatial resource demand data of all UAVs in each cluster, including horizontal airspace blocks, vertical altitude layers, and occupation time. Then, it matches the available regional resource vector, which includes available horizontal airspace blocks, vertical altitude layers, available time periods, and available duration. Using horizontal airspace blocks, vertical altitude layers, and time as the three comparison dimensions, X performs overlap verification on the spatial resource demands of UAVs in different clusters and simultaneously matches and verifies the airspace, altitude, and time period resources of the regional available resource vector. When there are overlapping airspace blocks, overlapping altitude layers, conflicts between resource occupation time and available time periods, or repeated occupation of the same available spatial resource by multiple clusters, UAVs Y and Z with overlapping resource occupation can be identified and cross-cluster conflict information can be generated. The cross-cluster conflict information includes the numbers of UAVs Y and Z that are in conflict, as well as the predicted conflict location and predicted conflict time obtained through comparison and deduction. It also includes the heading and speed of UAVs Y and Z respectively.

[0037] Subsequently, each cluster head obtains the cross-cluster conflict information uploaded by non-cluster head UAVs belonging to multiple clusters via the cross-cluster collaborative link, and summarizes and integrates all the cross-cluster conflict information to finally generate a multi-cluster conflict information table, providing data support for subsequent conflict resolution work such as multi-cluster collaborative game.

[0038] This invention utilizes a non-cluster-head UAV belonging to multiple clusters, combining information relay and conflict detection functions. Based on this, a multi-cluster information exchange network is constructed, synchronously comparing the intra-cluster member status tables and regional available resource vectors of different clusters. This enables accurate and efficient determination of cross-cluster conflicts, effectively avoiding misjudgments caused by static comparisons. Simultaneously, it reduces data collection and transmission redundancy, lowers UAV computing power consumption, and generates a multi-cluster conflict information table from the cross-cluster conflict information reported by each cluster head. This provides accurate, standardized, and efficient data support for subsequent multi-cluster collaborative game theory, significantly improving the efficiency and accuracy of cross-cluster conflict detection, ensuring the order and safety of multi-cluster UAV swarm flight, and promoting the orderly development of cross-cluster collaborative management of UAV swarms.

[0039] S5. If it is determined that there is a cross-cluster conflict based on the multi-cluster conflict information table, generate several sets of single-cluster conflict resolution strategies and the single-cluster resolution cost of each corresponding single-cluster conflict resolution strategy based on the single-cluster conflict information table. Preferably, based on the single-cluster conflict information table, the Cartesian product is used to iteratively adjust each UAV in its own cluster to generate several sets of single-cluster conflict resolution strategies. The resolution cost of each group of single-cluster conflict resolution strategies is calculated using the following formula, thus obtaining the single-cluster resolution cost corresponding to each group of single-cluster conflicts: ; ; ; in, This indicates the cost of resolving conflicts in a single cluster. Indicates the energy efficiency improvement factor; Indicates the time delay rate; This indicates the cumulative flight distance after strategy adjustment; This indicates the cumulative flight distance before the strategy adjustment; Indicates the task completion time after the strategy adjustment; Indicates the task deadline.

[0040] Specifically, after determining the existence of cross-cluster conflicts based on the multi-cluster conflict information table, each cluster head will initiate a single-cluster conflict resolution strategy generation process based on the single-cluster conflict information table. Using the task priority of UAVs within the cluster as the core principle, it prioritizes adjusting low-priority UAVs while preserving the original flight paths of high-priority UAVs as much as possible, and iteratively adjusts the UAVs within the cluster using a Cartesian product method. Illustratively, adjustable discrete values ​​are first generated for each conflicting UAV within the range of ±5°~±30° heading, ±10%~±30% speed, and ±1~±2 vertical altitude layers. Then, 3-5 complete conflict resolution sampling strategies are generated through full combinatorial matching, providing comprehensive samples for subsequent strategy optimization. For example, suppose there are two conflicting drones U1 and U2 in the cluster. The possible adjustment actions generated for U1 are heading +10° and speed +10% and heading +20° and speed +20%. The possible adjustment actions generated for U2 are altitude +1 layer and altitude +2 layers. After Cartesian product operation, four complete strategies will be obtained: (U1 heading +10° and speed +10%, U2 altitude +1 layer), (U1 heading +10° and speed +10%, U2 altitude +2 layers), (U1 heading +20° and speed +20%, U2 altitude +1 layer), and (U1 heading +20° and speed +20%, U2 altitude +2 layers). Then, 3-5 groups (4 groups in this case) are selected from these for subsequent cost evaluation to ensure that all possible adjustment combinations are covered.

[0041] After policy generation is complete, each cluster head will quantify the single-cluster resolution cost of each policy group using a unified weighted cost function, the calculation formula of which is: Among them, the energy efficiency improvement coefficient Defined as the ratio of the cumulative flight distance after strategy adjustment to the cumulative flight distance before adjustment, time delay rate. Defined as the ratio of task completion time to task deadline after strategy adjustment, this value is used to comprehensively evaluate the overall cost of each strategy in terms of energy consumption and task delay, providing a quantitative basis for subsequent strategy selection.

[0042] This invention, after a cross-cluster conflict occurs, prioritizes task priorities and adjusts low-priority UAVs to ensure the stability of high-priority task paths. It also uses a Cartesian product full combination sampling method to generate multiple single-cluster conflict resolution strategies. Quantitative evaluation is then completed through a weighted cost function of energy consumption and task delay. This approach ensures comprehensive coverage of resolution solutions while efficiently selecting the optimal resolution strategy that balances energy consumption control and task timeliness. It effectively avoids the one-sidedness of a single adjustment scheme, significantly improves the scientific nature and feasibility of conflict resolution, and ensures the safety and mission execution efficiency of UAV swarm cross-cluster collaborative flight.

[0043] S6. Based on the multi-cluster conflict information table and the single-cluster conflict resolution cost of each single-cluster conflict resolution strategy, a distributed game negotiation is conducted with the cluster head drone of the target conflict drone cluster that has cross-cluster conflicts to obtain the global optimal conflict resolution strategy.

[0044] Preferably, the target conflicting UAV clusters with cross-cluster conflicts are determined based on the multi-cluster conflict information table; Calculate the single-cluster conflict resolution cost of each single-cluster conflict resolution strategy within its own cluster, and select the single-cluster conflict resolution strategy with the minimum single-cluster conflict resolution cost as the first target single-cluster conflict resolution strategy. Based on the cluster member state table and the first target single cluster conflict resolution strategy, generate the first cluster UAV trajectory table corresponding to the first target single cluster conflict resolution strategy. The trajectory table of UAVs in the first cluster is compared with the trajectory table of UAVs in the second cluster of the target conflict UAV cluster. The trajectory table of UAVs in the second cluster is generated by the cluster leader UAV of the target conflict UAV cluster based on the second target single cluster conflict resolution strategy with the lowest single cluster resolution cost corresponding to the target conflict UAV cluster. When there is no overlap between the UAV trajectory tables in the first cluster and the UAV trajectory tables in the second cluster, the optimal global conflict resolution strategy is generated based on the conflict resolution strategies for the first and second objectives in a single cluster.

[0045] Next, if there is an overlap between the drone trajectory tables in the first cluster and the drone trajectory tables in the second cluster, the overall task priority of the cluster to which the drone belongs is compared with the overall task priority of the cluster of drones that are in conflict with the target. When the overall task priority of its own cluster is lower than the overall task priority of the target conflicting UAV cluster, the single-cluster conflict resolution strategy of its own cluster is updated according to the single-cluster conflict information table of its own cluster and the UAV trajectory table in the second cluster, resulting in several updated single-cluster conflict resolution strategies. Calculate the single-cluster resolution cost of each updated single-cluster conflict resolution strategy, and select the updated single-cluster conflict resolution strategy with the lowest single-cluster resolution cost as the first updated target single-cluster conflict resolution strategy. Based on the first updated target single-cluster conflict resolution strategy and the second target single-cluster conflict resolution strategy, the optimal global conflict resolution strategy is generated.

[0046] Preferably, the overall task priority is determined by the following formula: ; ; ; ; in, Indicates the task priority of the cluster; This represents the set of drones within the cluster that are currently involved in the conflict; Indicates drone Task priority; Indicates drone Weighting coefficients; Indicates the task completion rate; Indicates the remaining battery power; Factors indicating the severity of conflict; This represents the weighting coefficient for task completion. This represents the weighting coefficient for the cost of remaining electricity. Indicates the weighting coefficients representing the severity of the conflict; Indicates the current closest distance to other drones; Indicates the collision distance threshold; Indicates the remaining time expected for a collision; This indicates the warning time threshold.

[0047] Specifically, after identifying cross-cluster conflicts, the current cluster head (denoted as cluster A) will conduct distributed game negotiation with the cluster head of the target conflict cluster (denoted as cluster B) based on the multi-cluster conflict information table and the single-cluster resolution cost of each single-cluster conflict resolution strategy, and finally determine the optimal global conflict resolution strategy.

[0048] First, cluster A, based on the multi-cluster conflict information table, identifies the target conflict cluster (i.e., cluster B) with cross-cluster conflicts. Then, it calculates the single-cluster resolution cost of all single-cluster conflict resolution strategies within its own cluster, selecting the strategy with the lowest cost as the first target single-cluster conflict resolution strategy (denoted as S1). Based on the cluster member status table and S1, it generates the corresponding first cluster-wide UAV trajectory table (denoted as T1). Simultaneously, cluster B completes the same process, determining its own second target single-cluster conflict resolution strategy with the lowest cost (denoted as S2), and generating the corresponding second cluster-wide UAV trajectory table (denoted as T2). Cluster A compares T1 and T2; if the trajectories have no spatial resource overlap, it directly combines S1 and S2 to generate the globally optimal conflict resolution strategy.

[0049] If the comparison reveals that T1 and T2 still have overlapping resources, they will negotiate based on the overall task priority of the cluster. Before negotiation, each cluster will first calculate its own overall task priority using a formula. This priority is determined by the maximum value of the product of the task priority and the dynamic weight coefficient among the drones participating in the conflict within the cluster. (Where k represents the number of drones involved in the conflict within the cluster). Weighting coefficients The calculation is based on a combination of task completion rate, remaining battery power, and conflict severity factors, using the following formula: ,and It can be dynamically configured according to the scenario; the conflict severity factor is calculated from the normalized result of the current distance and the remaining collision time, and the formula is as follows: This is used to quantify the urgency and danger of conflicts. During negotiation, the cluster with the lower overall task priority acts as the active adjuster. If cluster A's priority is lower than cluster B's, cluster A will update its single-cluster conflict resolution strategy based on its own single-cluster conflict information table and cluster B's trajectory table T2, generating several updated strategies. Then, the single-cluster resolution cost of these strategies is recalculated, and the scheme with the lowest cost is selected as the first updated target single-cluster conflict resolution strategy (denoted as S1 update). Finally, combined with cluster B's S2, a globally optimal conflict resolution strategy without resource conflicts is formed. The entire negotiation process lasts no more than 1 second per round. By iteratively adjusting the strategies of lower-priority clusters, the cumulative value of the global resolution cost is gradually reduced until a globally optimal strategy combination is obtained.

[0050] This invention employs a distributed game-theoretic negotiation model among multiple conflict clusters. Each cluster independently generates its own resolution strategy and quantifies the resolution cost. Preliminary conflict verification is achieved through cross-referencing the initial flight trajectories of both clusters, enabling rapid assessment of the adaptability of the multi-cluster autonomous optimization strategy. This invention introduces standardized cluster-wide task priority calculation rules, using multi-dimensional parameters such as task progress, device power consumption, and conflict urgency as evaluation criteria. When spatial conflicts still exist in the initial trajectories of both clusters, it automatically distinguishes between primary and secondary clusters, adjusting only the lower-priority clusters to iteratively update their resolution schemes, without requiring simultaneous changes to the flight plans of the entire cluster. This collaborative adjustment method significantly reduces the scope of strategy modifications and computational overhead, effectively controlling the resolution cost of individual clusters and avoiding increased energy consumption and task delays caused by blind adjustments. Through multiple rounds of lightweight iterative negotiation, it rapidly converges, ultimately outputting a global conflict resolution scheme with no spatial interference and optimal overall cost. This approach balances the flight safety, task execution stability, and overall operational efficiency of multi-cluster UAV swarms, improving the intelligent management and control level of cross-cluster collaborative operations.

[0051] Accordingly, such as Figure 2 As shown, one embodiment of this application also provides a drone swarm path conflict handling system, including: a cluster leader drone within a conflicting drone cluster, non-cluster leader drones belonging to multiple clusters, and non-cluster leader drones belonging to a single cluster. The non-cluster-head UAVs belonging to multiple clusters and the non-cluster-head UAVs belonging to a single cluster are used to transmit the UAV's number, location, heading, speed, cumulative flight distance, mission priority, mission deadline, and space resource requirements to the cluster-head UAV. The cluster leader UAV is used to obtain the number, location, heading, speed, cumulative flight distance, mission priority, mission deadline, space resource requirements, and available space resources in the area of ​​all UAVs in the conflicting UAV cluster. Obtain the single-cluster conflict information uploaded by all conflicting UAVs within the cluster, and generate a single-cluster conflict information table based on the single-cluster conflict information. Obtain cross-cluster conflict information uploaded by non-cluster head UAVs belonging to multiple clusters, and generate a multi-cluster conflict information table based on the cross-cluster conflict information; If, based on the multi-cluster conflict information table, it is determined that there is a cross-cluster conflict, then based on the single-cluster conflict information table, several sets of single-cluster conflict resolution strategies and the single-cluster resolution cost of each corresponding single-cluster conflict resolution strategy are generated. Based on the multi-cluster conflict information table and the single-cluster conflict resolution cost of each single-cluster conflict resolution strategy, a distributed game negotiation is conducted with the cluster head drone of the target conflict drone cluster that has cross-cluster conflicts to obtain the global optimal conflict resolution strategy.

[0052] Preferably, the cluster leader drone is determined after election calculation based on the communication coverage, remaining battery power, and historical success rate of conflict resolution of the conflicting drones in the conflicting drone cluster; wherein, the conflicting drone cluster is formed when the drone density in a local area is not lower than a preset density value, and the real-time distance between any two drones in the local area is not greater than a preset distance.

[0053] Specifically, the formation of conflict drone clusters has clear conditions. It requires simultaneously meeting two requirements: a drone density of at least two drones per 1000 square meters in a local area, and a real-time distance between any two drones within that local area not exceeding a preset distance. The preset distance threshold is 30 meters in normal scenarios and dynamically adjusted to 50 meters in high-speed flight scenarios. Once these conditions are met, dynamic clustering is triggered, initiating the cluster head election process. The cluster head drone is determined through a scientific calculation using a three-dimensional weighting method. The election is based on three core indicators: the communication coverage range, remaining battery power, and historical conflict resolution success rate of each drone in the conflict drone cluster. Communication coverage range accounts for 30%, remaining battery power for 40%, and historical conflict resolution success rate for 30%. The drone with the highest weighted score is selected as the cluster head.

[0054] It is necessary to comprehensively collect the core information of all UAVs within the cluster and the available space resources within the area. The core information of the UAVs includes their number, location, heading, speed, cumulative flight distance, task priority, task deadline, and space resource requirements. Among them, the task priority is a fixed attribute that is pre-set before the UAV executes the corresponding task, which is used to define the scheduling order of different tasks. The task deadline is the latest time limit for completing the task, which serves as a constraint on UAV task scheduling and space resource occupancy planning. The space resource requirements specifically include horizontal airspace blocks, vertical altitude layers, and occupancy time. The available space resources within the area are specifically real-time data acquired every 5 seconds by the cluster head through 4G / 5G communication with the area resource management node. This data covers available horizontal airspace blocks, vertical altitude layers, available time periods, and available duration, providing a foundation for subsequent data processing.

[0055] Based on the collected data on all UAV numbers, locations, headings, speeds, and cumulative flight distances within the cluster, as well as preset task priorities, deadlines, and spatial resource requirements including horizontal airspace blocks, vertical altitude layers, and occupancy times, a cluster member status table is constructed after unified organization and standardization. This status table is distributed across the entire network by the cluster head using UWB ultra-wideband communication, with communication latency controlled within 100ms, and is updated in real-time at a 1-second interval to ensure high consistency of information for all UAV members within the cluster. Simultaneously, based on real-time collected information on available spatial resources within the region, various resource parameters such as available vertical altitude layers and horizontal airspace blocks are first classified and quantified, converting non-standardized resource information into calculable and comparable numerical forms. Based on this, a standardized regional available resource vector is generated. The available resource vector for this region fully embodies the quantitative characteristics of regional airspace resources, specifically reflecting the coverage and remaining available capacity of each vertical height layer, the coordinate boundaries and available area of ​​horizontal airspace blocks, and incorporating the dimensions of available airspace time period and available duration. The available resource vector is synchronously updated every 5 seconds by the cluster head pulling the latest airspace data from the regional resource management node via 4G / 5G communication. It can accurately and in real time characterize the remaining available capacity, spatial distribution range, and effective available duration of allocable airspace resources in the entire region. Subsequently, it can be directly linked and compared with the UAV spatial resource requirements recorded in the status tables of members within each cluster for supply and demand verification. This provides a real-time and accurate data foundation for the rational allocation of resources in multi-cluster collaborative games, ensuring that the overall game decision-making is scientific and rational.

[0056] This invention scientifically sets dynamic clustering conditions, uses a three-dimensional weighting method to elect the optimal cluster head, and relies on the cluster head to complete core operations such as information collection, data processing, and resource updates. This effectively ensures the consistency of information within the cluster and the real-time nature of regional airspace resource data, providing accurate and reliable data support for resource allocation in multi-cluster collaborative games. It significantly improves the scientificity and rationality of game decisions, realizes the orderly advancement of intra-cluster management and multi-cluster collaboration, effectively avoids airspace resource allocation conflicts, greatly improves the stability, efficiency, and safety of UAV swarm flight, and further optimizes the operational efficiency of UAV swarms in various scenarios.

[0057] Preferably, the conflicting drones are any two drones whose spatial resource needs overlap, or whose spatial resource distance at the same time is no greater than a preset spatial resource distance threshold.

[0058] Specifically, the determination of conflicting drones follows the intra-cluster conflict detection logic. This refers to any two drones within the same cluster, both being cluster members (including the cluster head and ordinary drones). First, the cluster head's distributed intra-cluster member status table is obtained, then the spatial resource requirements of both drones are extracted to determine the conflict. A drone is considered a conflicting drone if one of the following two conditions is met: First, their spatial resource requirements overlap, meaning both drones occupy the same horizontal airspace and vertical height resources at the same time, resulting in overlapping resource usage and thus intra-cluster conflict. Second, at the same time, the spatial resources occupied by both drones are adjacent, and the distance between adjacent resources is no greater than a preset threshold. This threshold is typically 15 meters and can be dynamically adjusted according to drone size. For large drones, the threshold is set to 20 meters; in this case, even if they do not occupy the same resources, it is still considered intra-cluster conflict.

[0059] After determining the conflicting drones, the cluster head receives various conflict data uploaded in real time by all drones in the cluster that have autonomously identified flight conflicts within the cluster via the real-time status link within the cluster. Each reported data includes key information such as the number of the drone and the target drone, the estimated conflict location, the estimated conflict time, the heading, and the speed. The cluster head continuously summarizes this reported information and organizes it according to a unified standard, continuously collecting and sorting out all conflict entries within the cluster and completing standardized integration processing, thereby forming a complete and standardized single-cluster conflict information table.

[0060] This invention employs intra-cluster conflict detection logic to autonomously determine UAV conflicts. Combining the dual criteria of overlapping airspace resources and distance constraints between adjacent resources, it accurately identifies flight conflict behaviors between the cluster head and ordinary UAVs. Simultaneously, it leverages the real-time status link within the cluster to rapidly collect core data reported by conflicting UAVs. After standardization and integration, it generates a single-cluster conflict information table, which can comprehensively, accurately, and efficiently analyze various flight conflict situations within the cluster, avoid the risk of close-range collisions between UAVs within the cluster, and provide complete and reliable data support for multi-cluster collaborative game decision-making and subsequent conflict resolution. This effectively improves the safety and orderliness of UAV swarm flight operations.

[0061] Preferably, the conflicting drone cluster also includes: non-cluster leader drones belonging to multiple clusters; Preferably, the cross-cluster conflict information is determined in the following way: Non-cluster-head UAVs belonging to multiple clusters compare the spatial resource requirements and available resource vectors in the status tables of members within different clusters. If there are duplicate comparison results, cross-cluster conflict information is obtained.

[0062] Specifically, within a conflict-ridden drone cluster, in addition to regular cluster members, there are also non-cluster head drones belonging to multiple clusters. These drones belong to multiple clusters simultaneously and function as both information relays and conflict detection systems. During cross-cluster information exchange, non-cluster head drones belonging to multiple clusters act as information relay nodes, forwarding their own cluster member status tables (which have been anonymized and retain only cross-cluster conflict-related data) to the cluster heads of the other clusters they belong to every 2 seconds. This constructs a multi-cluster information exchange network, providing a data foundation for cross-cluster conflict detection.

[0063] The determination of cross-cluster conflict information is accomplished by non-cluster head UAVs with multiple cluster affiliations: Assuming X is a non-cluster head UAV with multiple cluster affiliations and responsible for cross-cluster conflict comparison and judgment, Y and Z are two UAVs belonging to different clusters and having cross-cluster resource conflicts with each other. X is only responsible for global data retrieval and conflict assessment, and does not necessarily participate in the conflict relationship between Y and Z. A non-cluster leader UAV X with multi-cluster affiliation is responsible for comparing and judging cross-cluster conflicts. X retrieves the status tables of the members in each of its respective clusters and extracts the spatial resource demand data of all UAVs in each cluster, including horizontal airspace blocks, vertical altitude layers, and occupation time. Then, it matches the available regional resource vector, which includes available horizontal airspace blocks, vertical altitude layers, available time periods, and available duration. Using horizontal airspace blocks, vertical altitude layers, and time as the three comparison dimensions, X performs overlap verification on the spatial resource demands of UAVs in different clusters and simultaneously matches and verifies the airspace, altitude, and time period resources of the regional available resource vector. When there are overlapping airspace blocks, overlapping altitude layers, conflicts between resource occupation time and available time periods, or repeated occupation of the same available spatial resource by multiple clusters, UAVs Y and Z with overlapping resource occupation can be identified and cross-cluster conflict information can be generated. The cross-cluster conflict information includes the numbers of UAVs Y and Z that are in conflict, as well as the predicted conflict location and predicted conflict time obtained through comparison and deduction. It also includes the heading and speed of UAVs Y and Z respectively.

[0064] Subsequently, each cluster head obtains the cross-cluster conflict information uploaded by non-cluster head UAVs belonging to multiple clusters via the cross-cluster collaborative link, and summarizes and integrates all the cross-cluster conflict information to finally generate a multi-cluster conflict information table, providing data support for subsequent conflict resolution work such as multi-cluster collaborative game.

[0065] This invention utilizes a non-cluster-head UAV belonging to multiple clusters, combining information relay and conflict detection functions. Based on this, a multi-cluster information exchange network is constructed, synchronously comparing the intra-cluster member status tables and regional available resource vectors of different clusters. This enables accurate and efficient determination of cross-cluster conflicts, effectively avoiding misjudgments caused by static comparisons. Simultaneously, it reduces data collection and transmission redundancy, lowers UAV computing power consumption, and generates a multi-cluster conflict information table from the cross-cluster conflict information reported by each cluster head. This provides accurate, standardized, and efficient data support for subsequent multi-cluster collaborative game theory, significantly improving the efficiency and accuracy of cross-cluster conflict detection, ensuring the order and safety of multi-cluster UAV swarm flight, and promoting the orderly development of cross-cluster collaborative management of UAV swarms.

[0066] Preferably, based on the single-cluster conflict information table, the Cartesian product is used to iteratively adjust each UAV in its own cluster to generate several sets of single-cluster conflict resolution strategies. The resolution cost of each group of single-cluster conflict resolution strategies is calculated using the following formula, thus obtaining the single-cluster resolution cost corresponding to each group of single-cluster conflicts: ; ; ; in, This indicates the cost of resolving conflicts in a single cluster. Indicates the energy efficiency improvement factor; Indicates the time delay rate; This indicates the cumulative flight distance after strategy adjustment; This indicates the cumulative flight distance before the strategy adjustment; Indicates the task completion time after the strategy adjustment; Indicates the task deadline.

[0067] Specifically, after determining the existence of cross-cluster conflicts based on the multi-cluster conflict information table, each cluster head will initiate a single-cluster conflict resolution strategy generation process based on the single-cluster conflict information table. Using the task priority of UAVs within the cluster as the core principle, it prioritizes adjusting low-priority UAVs while preserving the original flight paths of high-priority UAVs as much as possible, and iteratively adjusts the UAVs within the cluster using a Cartesian product method. Illustratively, adjustable discrete values ​​are first generated for each conflicting UAV within the range of ±5°~±30° heading, ±10%~±30% speed, and ±1~±2 vertical altitude layers. Then, 3-5 complete conflict resolution sampling strategies are generated through full combinatorial matching, providing comprehensive samples for subsequent strategy optimization. For example, suppose there are two conflicting drones U1 and U2 in the cluster. The possible adjustment actions generated for U1 are heading +10° and speed +10% and heading +20° and speed +20%. The possible adjustment actions generated for U2 are altitude +1 layer and altitude +2 layers. After Cartesian product operation, four complete strategies will be obtained: (U1 heading +10° and speed +10%, U2 altitude +1 layer), (U1 heading +10° and speed +10%, U2 altitude +2 layers), (U1 heading +20° and speed +20%, U2 altitude +1 layer), and (U1 heading +20° and speed +20%, U2 altitude +2 layers). Then, 3-5 groups (4 groups in this case) are selected from these for subsequent cost evaluation to ensure that all possible adjustment combinations are covered.

[0068] After policy generation is complete, each cluster head will quantify the single-cluster resolution cost of each policy group using a unified weighted cost function, the calculation formula of which is: Among them, the energy efficiency improvement coefficient Defined as the ratio of the cumulative flight distance after strategy adjustment to the cumulative flight distance before adjustment, time delay rate. Defined as the ratio of task completion time to task deadline after strategy adjustment, this value is used to comprehensively evaluate the overall cost of each strategy in terms of energy consumption and task delay, providing a quantitative basis for subsequent strategy selection.

[0069] This invention, after a cross-cluster conflict occurs, prioritizes task priorities and adjusts low-priority UAVs to ensure the stability of high-priority task paths. It also uses a Cartesian product full combination sampling method to generate multiple single-cluster conflict resolution strategies. Quantitative evaluation is then completed through a weighted cost function of energy consumption and task delay. This approach ensures comprehensive coverage of resolution solutions while efficiently selecting the optimal resolution strategy that balances energy consumption control and task timeliness. It effectively avoids the one-sidedness of a single adjustment scheme, significantly improves the scientific nature and feasibility of conflict resolution, and ensures the safety and mission execution efficiency of UAV swarm cross-cluster collaborative flight.

[0070] Preferably, the target conflicting UAV clusters with cross-cluster conflicts are determined based on the multi-cluster conflict information table; Calculate the single-cluster conflict resolution cost of each single-cluster conflict resolution strategy within its own cluster, and select the single-cluster conflict resolution strategy with the minimum single-cluster conflict resolution cost as the first target single-cluster conflict resolution strategy. Based on the cluster member state table and the first target single cluster conflict resolution strategy, generate the first cluster UAV trajectory table corresponding to the first target single cluster conflict resolution strategy. The trajectory table of UAVs in the first cluster is compared with the trajectory table of UAVs in the second cluster of the target conflict UAV cluster. The trajectory table of UAVs in the second cluster is generated by the cluster leader UAV of the target conflict UAV cluster based on the second target single cluster conflict resolution strategy with the lowest single cluster resolution cost corresponding to the target conflict UAV cluster. When there is no overlap between the UAV trajectory tables in the first cluster and the UAV trajectory tables in the second cluster, the optimal global conflict resolution strategy is generated based on the conflict resolution strategies for the first and second objectives in a single cluster.

[0071] Next, if there is an overlap between the drone trajectory tables in the first cluster and the drone trajectory tables in the second cluster, the overall task priority of the cluster to which the drone belongs is compared with the overall task priority of the cluster of drones that are in conflict with the target. When the overall task priority of its own cluster is lower than the overall task priority of the target conflicting UAV cluster, the single-cluster conflict resolution strategy of its own cluster is updated according to the single-cluster conflict information table of its own cluster and the UAV trajectory table in the second cluster, resulting in several updated single-cluster conflict resolution strategies. Calculate the single-cluster resolution cost of each updated single-cluster conflict resolution strategy, and select the updated single-cluster conflict resolution strategy with the lowest single-cluster resolution cost as the first updated target single-cluster conflict resolution strategy. Based on the first updated target single-cluster conflict resolution strategy and the second target single-cluster conflict resolution strategy, the optimal global conflict resolution strategy is generated.

[0072] Preferably, the overall task priority is determined by the following formula: ; ; ; ; in, Indicates the task priority of the cluster; This represents the set of drones within the cluster that are currently involved in the conflict; Indicates drone Task priority; Indicates drone Weighting coefficients; Indicates the task completion rate; Indicates the remaining battery power; Factors indicating the severity of conflict; This represents the weighting coefficient for task completion. This represents the weighting coefficient for the cost of remaining electricity. Indicates the weighting coefficients representing the severity of the conflict; Indicates the current closest distance to other drones; Indicates the collision distance threshold; Indicates the remaining time expected for a collision; This indicates the warning time threshold.

[0073] Specifically, after identifying cross-cluster conflicts, the current cluster head (denoted as cluster A) will conduct distributed game negotiation with the cluster head of the target conflict cluster (denoted as cluster B) based on the multi-cluster conflict information table and the single-cluster resolution cost of each single-cluster conflict resolution strategy, and finally determine the optimal global conflict resolution strategy.

[0074] First, cluster A, based on the multi-cluster conflict information table, identifies the target conflict cluster (i.e., cluster B) with cross-cluster conflicts. Then, it calculates the single-cluster resolution cost of all single-cluster conflict resolution strategies within its own cluster, selecting the strategy with the lowest cost as the first target single-cluster conflict resolution strategy (denoted as S1). Based on the cluster member status table and S1, it generates the corresponding first cluster-wide UAV trajectory table (denoted as T1). Simultaneously, cluster B completes the same process, determining its own second target single-cluster conflict resolution strategy with the lowest cost (denoted as S2), and generating the corresponding second cluster-wide UAV trajectory table (denoted as T2). Cluster A compares T1 and T2; if the trajectories have no spatial resource overlap, it directly combines S1 and S2 to generate the globally optimal conflict resolution strategy.

[0075] If the comparison reveals that T1 and T2 still have overlapping resources, they will negotiate based on the overall task priority of the cluster. Before negotiation, each cluster will first calculate its own overall task priority using a formula. This priority is determined by the maximum value of the product of the task priority and the dynamic weight coefficient among the drones participating in the conflict within the cluster. (Where k represents the number of drones involved in the conflict within the cluster). Weighting coefficients The calculation is based on a combination of task completion rate, remaining battery power, and conflict severity factors, using the following formula: ,and It can be dynamically configured according to the scenario; the conflict severity factor is calculated from the normalized result of the current distance and the remaining collision time, and the formula is as follows: This is used to quantify the urgency and danger of conflicts. During negotiation, the cluster with the lower overall task priority acts as the active adjuster. If cluster A's priority is lower than cluster B's, cluster A will update its single-cluster conflict resolution strategy based on its own single-cluster conflict information table and cluster B's trajectory table T2, generating several updated strategies. Then, the single-cluster resolution cost of these strategies is recalculated, and the scheme with the lowest cost is selected as the first updated target single-cluster conflict resolution strategy (denoted as S1 update). Finally, combined with cluster B's S2, a globally optimal conflict resolution strategy without resource conflicts is formed. The entire negotiation process lasts no more than 1 second per round. By iteratively adjusting the strategies of lower-priority clusters, the cumulative value of the global resolution cost is gradually reduced until a globally optimal strategy combination is obtained.

[0076] This invention employs a distributed game-theoretic negotiation model among multiple conflict clusters. Each cluster independently generates its own resolution strategy and quantifies the resolution cost. Preliminary conflict verification is achieved through cross-referencing the initial flight trajectories of both clusters, enabling rapid assessment of the adaptability of the multi-cluster autonomous optimization strategy. This invention introduces standardized cluster-wide task priority calculation rules, using multi-dimensional parameters such as task progress, device power consumption, and conflict urgency as evaluation criteria. When spatial conflicts still exist in the initial trajectories of both clusters, it automatically distinguishes between primary and secondary clusters, adjusting only the lower-priority clusters to iteratively update their resolution schemes, without requiring simultaneous changes to the flight plans of the entire cluster. This collaborative adjustment method significantly reduces the scope of strategy modifications and computational overhead, effectively controlling the resolution cost of individual clusters and avoiding increased energy consumption and task delays caused by blind adjustments. Through multiple rounds of lightweight iterative negotiation, it rapidly converges, ultimately outputting a global conflict resolution scheme with no spatial interference and optimal overall cost. This approach balances the flight safety, task execution stability, and overall operational efficiency of multi-cluster UAV swarms, improving the intelligent management and control level of cross-cluster collaborative operations.

[0077] It is understood that the above system item embodiments correspond to the method item embodiments of the present invention, and can implement the method for handling path conflicts of UAV clusters provided by any of the above method item embodiments of the present invention.

[0078] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for handling path conflicts in a drone swarm, applicable to the head drone within a conflicting drone swarm, characterized in that, The conflicting drone cluster also includes: non-cluster leader drones belonging to multiple clusters; The method for handling path conflicts in a drone swarm includes Obtain the ID, location, heading, speed, cumulative flight distance, mission priority, mission deadline, space resource requirements, and available space resources within the cluster for all UAVs. A cluster member status table is generated based on the number, location, heading, speed, cumulative flight distance, task priority, task deadline, and spatial resource requirements of all UAVs within the cluster. A regional available resource vector is generated based on the available spatial resources within the region. Obtain the single-cluster conflict information uploaded by all conflicting UAVs within the cluster, and generate a single-cluster conflict information table based on the single-cluster conflict information. Obtain cross-cluster conflict information uploaded by non-cluster head UAVs belonging to multiple clusters, and generate a multi-cluster conflict information table based on the cross-cluster conflict information; If, based on the multi-cluster conflict information table, it is determined that there is a cross-cluster conflict, then based on the single-cluster conflict information table, several sets of single-cluster conflict resolution strategies and the single-cluster resolution cost of each corresponding single-cluster conflict resolution strategy are generated. Based on the multi-cluster conflict information table and the single-cluster conflict resolution cost of each single-cluster conflict resolution strategy, a distributed game negotiation is conducted with the cluster head drone of the target conflict drone cluster that has cross-cluster conflicts to obtain the global optimal conflict resolution strategy.

2. The method for handling path conflicts in a drone swarm as described in claim 1, characterized in that, The cluster leader drone is determined by election calculation based on the communication coverage, remaining battery power, and historical success rate of conflict resolution of the conflicting drones in the conflicting drone cluster; wherein, the conflicting drone cluster is formed when the drone density in a local area is not lower than a preset density value, and the real-time distance between any two drones in the local area is not greater than a preset distance.

3. The method for handling path conflicts in a drone swarm as described in claim 2, characterized in that, The conflicting drones are any two drones whose spatial resource needs overlap, or whose spatial resource distance at the same time is no greater than a preset spatial resource distance threshold.

4. The method for handling path conflicts in a drone swarm as described in claim 3, characterized in that, The cross-cluster conflict information is determined in the following way: Non-cluster-head UAVs belonging to multiple clusters compare the spatial resource requirements and available resource vectors in the status tables of members within different clusters. If there are duplicate comparison results, cross-cluster conflict information is obtained.

5. The method for handling path conflicts in a drone swarm as described in claim 4, characterized in that, If, based on the multi-cluster conflict information table, cross-cluster conflicts are determined to exist, several sets of single-cluster conflict resolution strategies and the single-cluster resolution cost of each corresponding single-cluster conflict resolution strategy are generated based on the single-cluster conflict information table, including: Based on the single-cluster conflict information table, Cartesian product is used to iteratively adjust each UAV in its own cluster to generate several sets of single-cluster conflict resolution strategies. The resolution cost of each group of single-cluster conflict resolution strategies is calculated using the following formula, thus obtaining the single-cluster resolution cost corresponding to each group of single-cluster conflicts: ； ； ； in, This indicates the cost of resolving conflicts in a single cluster. Indicates the energy efficiency improvement factor; Indicates the time delay rate; This indicates the cumulative flight distance after strategy adjustment; This indicates the cumulative flight distance before the strategy adjustment; Indicates the task completion time after the strategy adjustment; Indicates the task deadline.

6. The method for handling path conflicts in a drone swarm as described in claim 5, characterized in that, Based on the multi-cluster conflict information table and the single-cluster conflict resolution cost of each single-cluster conflict resolution strategy, a distributed game negotiation is conducted with the cluster leader drone of the target conflict drone cluster with cross-cluster conflicts to obtain the global optimal conflict resolution strategy, including: Based on the multi-cluster conflict information table, target conflicting UAV clusters with cross-cluster conflicts are identified; Calculate the single-cluster conflict resolution cost of each single-cluster conflict resolution strategy within its own cluster, and select the single-cluster conflict resolution strategy with the minimum single-cluster conflict resolution cost as the first target single-cluster conflict resolution strategy. Based on the cluster member state table and the first target single cluster conflict resolution strategy, generate the first cluster UAV trajectory table corresponding to the first target single cluster conflict resolution strategy. The trajectory table of UAVs in the first cluster is compared with the trajectory table of UAVs in the second cluster of the target conflict UAV cluster. The trajectory table of UAVs in the second cluster is generated by the cluster leader UAV of the target conflict UAV cluster based on the second target single cluster conflict resolution strategy with the lowest single cluster resolution cost corresponding to the target conflict UAV cluster. When there is no overlap between the UAV trajectory tables in the first cluster and the UAV trajectory tables in the second cluster, the optimal global conflict resolution strategy is generated based on the conflict resolution strategies for the first and second objectives in a single cluster.

7. The method for handling path conflicts in a drone swarm as described in claim 6, characterized in that, Also includes: If the trajectory tables of UAVs in the first cluster overlap with those in the second cluster, compare the overall task priority of the cluster to which the UAV belongs with the overall task priority of the cluster of UAVs that are in conflict with the target. When the overall task priority of its own cluster is lower than the overall task priority of the target conflicting UAV cluster, the single-cluster conflict resolution strategy of its own cluster is updated according to the single-cluster conflict information table of its own cluster and the UAV trajectory table in the second cluster, resulting in several updated single-cluster conflict resolution strategies. Calculate the single-cluster resolution cost of each updated single-cluster conflict resolution strategy, and select the updated single-cluster conflict resolution strategy with the lowest single-cluster resolution cost as the first updated target single-cluster conflict resolution strategy. Based on the first updated target single-cluster conflict resolution strategy and the second target single-cluster conflict resolution strategy, the optimal global conflict resolution strategy is generated.

8. The method for handling path conflicts in a drone swarm as described in claim 6, characterized in that, The overall task priority is determined by the following formula: ； ； ； ； in, Indicates the task priority of the cluster; This represents the set of drones within the cluster that are currently involved in the conflict; Indicates drone Task priority; This represents the weighting coefficient of drone k; Indicates the task completion rate; Indicates the remaining battery power; Factors indicating the severity of conflict; This represents the weighting coefficient for task completion. This represents the weighting coefficient for the cost of remaining electricity. Indicates the weighting coefficients representing the severity of the conflict; Indicates the current closest distance to other drones; Indicates the collision distance threshold; Indicates the remaining time expected for a collision; This indicates the warning time threshold.

9. A system for handling path conflicts in a drone swarm, characterized in that, include: The cluster leader drone within a conflicting drone cluster, non-cluster leader drones belonging to multiple clusters, and non-cluster leader drones belonging to a single cluster; The non-cluster-head UAVs belonging to multiple clusters and the non-cluster-head UAVs belonging to a single cluster are used to transmit the UAV's number, location, heading, speed, cumulative flight distance, mission priority, mission deadline, and space resource requirements to the cluster-head UAV. The cluster leader UAV is used to obtain the number, location, heading, speed, cumulative flight distance, mission priority, mission deadline, space resource requirements, and available space resources in the area of ​​all UAVs in the conflicting UAV cluster. Obtain the single-cluster conflict information uploaded by all conflicting UAVs within the cluster, and generate a single-cluster conflict information table based on the single-cluster conflict information. Obtain cross-cluster conflict information uploaded by non-cluster head UAVs belonging to multiple clusters, and generate a multi-cluster conflict information table based on the cross-cluster conflict information; If, based on the multi-cluster conflict information table, it is determined that there is a cross-cluster conflict, then based on the single-cluster conflict information table, several sets of single-cluster conflict resolution strategies and the single-cluster resolution cost of each corresponding single-cluster conflict resolution strategy are generated. Based on the multi-cluster conflict information table and the single-cluster conflict resolution cost of each single-cluster conflict resolution strategy, a distributed game negotiation is conducted with the cluster head drone of the target conflict drone cluster that has cross-cluster conflicts to obtain the global optimal conflict resolution strategy.

10. The UAV swarm path conflict handling system as described in claim 9, characterized in that, The cluster leader drone is determined by election calculation based on the communication coverage, remaining battery power, and historical success rate of conflict resolution of the conflicting drones in the conflicting drone cluster; wherein, the conflicting drone cluster is formed when the drone density in a local area is not lower than a preset density value, and the real-time distance between any two drones in the local area is not greater than a preset distance.

Images