Methods, devices and products for intelligent scheduling and collaborative operation of multiple unmanned aerial vehicles (UAVs)

Abstract

This invention provides a method, apparatus, and product for intelligent scheduling and collaborative work of multiple unmanned aerial vehicles (UAVs). The method includes: step S1, receiving tasks to be processed in real time and obtaining the real-time operating status of each UAV in the UAV swarm; step S2, intelligently and dynamically allocating tasks using a pre-established task value scoring model, a UAV task suitability model, and a global multi-task dynamic allocation optimization model; the global multi-task dynamic allocation optimization model is a mixed-integer programming model established under predetermined constraints with the goal of maximizing overall task benefits; step S3, after task allocation, planning routes for each UAV performing the task using a pre-constructed three-dimensional grid path model G(V,E). By combining the mixed-integer programming algorithm with particle swarm optimization, a global task allocation optimization model is constructed, enabling efficient task allocation while also meeting various practical constraints.

Description

Technical Field

[0001] This invention relates to the field of multi-UAV collaborative work technology, and in particular to a method, apparatus and product for realizing intelligent scheduling and collaborative work of multiple UAVs. Background Technology

[0002] Currently, most urban police drone applications operate on a single-drone mission model, relying primarily on manual command or pre-set routes for tasks such as patrolling, evidence collection, and public address systems. There is a lack of systematic intelligent scheduling and coordination mechanisms for multi-drone swarms. In complex urban environments, when multiple drones perform diverse and unpredictable police tasks within limited airspace, they face technical bottlenecks such as flight path conflicts, response delays, unbalanced task allocation, and low coordination efficiency, severely restricting the efficient operational capabilities of police drone swarms. Summary of the Invention

[0003] To address one or more of the problems of the prior art, embodiments of the present invention provide a method, apparatus, and product for realizing intelligent scheduling and collaborative work of multiple unmanned aerial vehicles (UAVs).

[0004] To achieve the above objectives, on the one hand, a method for realizing intelligent scheduling and collaborative work of multiple unmanned aerial vehicles (UAVs) is provided, including:

[0005] Step S1: Receive tasks to be processed in real time. And acquire information about each drone in the drone swarm. The real-time operating status includes: current location, remaining battery power, current load capacity, and configured device capacity information; i=1…m,j=1…n; m is the total number of tasks; n is the total number of drones;

[0006] Step S2 involves intelligently and dynamically allocating tasks using a pre-established task value scoring model, a UAV task suitability model, and a global multi-task dynamic allocation optimization model. The global multi-task dynamic allocation optimization model is a mixed-integer programming model established under predetermined constraints with the objective of maximizing overall task revenue.

[0007] The mixed-integer programming model is as follows:

[0008] ;

[0009] The task value scoring model is as follows:

[0010] ;

[0011] The drone mission suitability model is as follows:

[0012] ;

[0013] For the task Assigned to drones The decision variable takes the value 0 or 1, where 0 represents the drone. Not undertaking this task, 1 indicates a drone. Undertake this task;

[0014] Indicates task Value rating This indicates the pre-defined urgency level of the task; Indicate the importance of pre-defined tasks; This indicates the pre-defined task complexity; , and These are the decision weight parameters obtained through the particle swarm optimization algorithm;

[0015] Indicates drone For the task Compatibility score; Indicates drone With the task Flight distance between locations Indicates task With drones The matching score of the required functional equipment for the mission is used to measure the drone's capabilities. Are they capable of completing the task? Required equipment and functions; , and These are predefined weighting parameters; Indicates drone The remaining battery power, , where 0 indicates low battery and 1 indicates full battery;

[0016] The predetermined constraints of the mixed-integer programming model include: each UAV can only perform one task and the execution of a task cannot exceed the UAV's capability limit; the capability includes payload and endurance.

[0017] Step S3: After the task allocation is completed, a pre-built three-dimensional grid route model G(V, E) is used to plan routes for each UAV performing the task; V is a node, representing a feasible route point, which includes the flight altitude; E is an edge, representing the flight path.

[0018] Preferably, in the method for realizing intelligent scheduling and collaborative work of multiple UAVs, step S3 includes: single-UAV route optimization and multi-route conflict detection; wherein,

[0019] The goal of optimizing the flight path of a single UAV is:

[0020]

[0021] k is the index of each segment in the route; p is the total number of segments contained in the route; The cost of each k-th flight segment; Indicates the length of the flight segment; This indicates the risk coefficient of the flight altitude layer, which is predetermined based on factors such as the density of urban buildings, restrictions and regulations in the flight area, and the type of flight mission. This indicates a high-penalty weighting factor;

[0022] Multi-route conflict detection includes:

[0023] Using a sliding time window Inspect any two drones and Are there any overlapping flight paths? If so, a flight path conflict has been detected.

[0024] If a flight path conflict is detected, the takeoff time or flight segment progress time of the UAV can be adjusted by local time offset, or the temporary altitude corridor can be adjusted by local spatial offset.

[0025] Preferably, the method for realizing intelligent scheduling and collaborative work of multiple UAVs involves establishing a matching matrix between subtask roles and UAVs based on the equipment capabilities required by the subtask roles included in each task, the equipment capabilities of each UAV, and the priority weight of each task when the task includes multiple subtask roles, and then assigning the subtask roles to the corresponding UAVs according to the matching matrix.

[0026] Preferably, in the method for realizing intelligent scheduling and collaborative work of multiple UAVs, the sub-task roles include one or more of the following roles:

[0027] For reconnaissance roles, required equipment capabilities include a high-definition zoom camera;

[0028] For a shouting role, the required equipment includes: a megaphone;

[0029] The lighting role requires equipment capabilities including a high-intensity light projector.

[0030] Preferably, the method for realizing intelligent scheduling and collaborative work of multiple UAVs further includes:

[0031] When the predetermined dynamic role reassignment conditions are triggered, the role of the drone is dynamically adjusted according to the urgency of the mission and the on-site environment.

[0032] The predetermined dynamic role reassignment conditions include one or more of the following: current device failure alarm and target status change; the target status change includes: target behavior direction change.

[0033] Preferably, the method for realizing intelligent scheduling and collaborative work of multiple UAVs further includes:

[0034] During mission execution, the drones performing each mission establish a data link through a preset communication network and use the data link to periodically broadcast their current position, velocity vector, collected field data, and current role status.

[0035] Preferably, the method for achieving intelligent scheduling and collaborative work of multiple UAVs further includes, when a predetermined rescheduling condition is triggered, using a local incremental reconstruction model to perform real-time adaptive rescheduling of tasks; the local incremental reconstruction model keeps the unaffected tasks unchanged, and only uses the following formula to perform replanning on the affected tasks:

[0036]

[0037] This represents the time cost and adaptation penalty incurred by newly allocating drones; "new" indicates a new allocation. This indicates the cost of switching roles during a mission; switch indicates a role switch.

[0038] Preferably, the method for realizing intelligent scheduling and collaborative work of multiple drones refers to collaborative law enforcement; in step S1, the processing task is received in real time through a multi-source police alarm access interface. The multi-source alarm access interface includes one or more of the following: 110 alarm system, video surveillance platform, and on-duty patrol feedback; the processing task It includes one or more of the following task information: task type, event address, expected execution time limit, and device characteristics required for the task.

[0039] On the other hand, an apparatus for realizing intelligent scheduling and collaborative operation of multiple unmanned aerial vehicles (UAVs) is provided, including a memory and a processor. The memory stores at least one program, which is executed by the processor to implement the steps of the method for realizing intelligent scheduling and collaborative operation of multiple UAVs as described above.

[0040] On the other hand, a system for realizing intelligent scheduling and collaborative work of multiple unmanned aerial vehicles (UAVs) is provided, comprising: multiple UAVs; and, as described herein, an apparatus for realizing intelligent scheduling and collaborative work of multiple UAVs, used for scheduling and collaborative work of the multiple UAVs.

[0041] The above technical solution has the following technical effects:

[0042] The technical solution of this invention combines the Mixed Integer Programming (MILP) algorithm with the existing Particle Swarm Optimization (PSO) algorithm to form a multi-objective, multi-constraint optimization task scheduling framework, breaking through the limitations of existing technologies and realizing a global task allocation optimization model. The global task allocation optimization model can accurately model the constraints between multiple UAVs and tasks, ensuring that the task allocation of each UAV maximizes the task completion rate without overloading, thereby guaranteeing task efficiency and the rational use of UAV resources. This makes task allocation both efficient and compliant with various constraints in practical applications.

[0043] The technical solution of a further embodiment of the present invention, by employing one or more of the following: a collaborative path planning algorithm, a multi-role dynamic division of labor mechanism, and a real-time adaptive task reconstruction mechanism, achieves efficient dynamic scheduling, safe collision avoidance, real-time task succession, and multi-role collaborative law enforcement of multiple UAVs in a limited urban airspace in a complex urban environment. Attached Figure Description

[0044] Figure 1 This is a flowchart illustrating a method for intelligent scheduling and collaborative work of multiple unmanned aerial vehicles (UAVs) according to an embodiment of the present invention.

[0045] Figure 2 This is a schematic diagram of a device for realizing intelligent scheduling and collaborative work of multiple unmanned aerial vehicles (UAVs) according to an embodiment of the present invention. Detailed Implementation

[0046] To further illustrate the various embodiments, the present invention provides accompanying drawings. These drawings are part of the disclosure of the present invention, primarily used to illustrate the embodiments and to explain the operating principles of the embodiments in conjunction with the relevant descriptions in the specification. With reference to these drawings, those skilled in the art should be able to understand other possible implementations and the advantages of the present invention. Components in the drawings are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0047] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments.

[0048] Example 1:

[0049] Figure 1 This is a flowchart illustrating a method for intelligent scheduling and collaborative operation of multiple unmanned aerial vehicles (UAVs) according to an embodiment of the present invention. Figure 1 The method for intelligent scheduling and collaborative operation of multiple unmanned aerial vehicles (UAVs) in this embodiment of the present invention includes:

[0050] Step S1: Receive tasks to be processed in real time. And acquire information about each drone in the drone swarm. The real-time operating status includes: current location, remaining battery power, current load capacity, and configured device capacity information; i=1…m, j=1…n; m is the total number of tasks; n is the total number of drones;

[0051] Step S2 involves intelligently and dynamically allocating tasks using a pre-established task value scoring model, a UAV task suitability model, and a global multi-task dynamic allocation optimization model. The global multi-task dynamic allocation optimization model is a mixed-integer programming model established under predetermined constraints with the objective of maximizing overall task revenue.

[0052] The mixed-integer programming model is as follows:

[0053] ;

[0054] The task value scoring model is as follows:

[0055] ;

[0056] The drone mission suitability model is as follows:

[0057] ;

[0058] For the task Assigned to drones The decision variable takes the value 0 or 1, where 0 represents the drone. Not undertaking this task, 1 indicates a drone. Undertake this task;

[0059] Indicates task Value rating This indicates the pre-defined urgency level of the task; Indicate the importance of pre-defined tasks; This indicates the pre-defined task complexity; , and These are the decision weight parameters obtained through the particle swarm optimization algorithm;

[0060] Indicates drone For the task Compatibility score; Indicates drone With the task Flight distance between locations Indicates task With drones The matching score of the required functional equipment for the mission is used to measure the drone's capabilities. Are they capable of completing the task? Required equipment and functions; , and These are predefined weighting parameters; Indicates drone The remaining battery power, , where 0 indicates low battery and 1 indicates full battery;

[0061] The predetermined constraints of the mixed-integer programming model include: each drone can only perform one task and the execution of a task cannot exceed the drone's capability limit; capability includes payload and endurance.

[0062] Step S3: After the task allocation is completed, the pre-built three-dimensional grid route model G(V, E) is used to plan routes for each UAV performing the task; V is a node, representing a feasible route point, which includes the flight altitude; E is an edge, representing the flight path.

[0063] Preferably, in the method for intelligent scheduling and collaborative work of multiple UAVs according to the embodiments of the present invention, when a task includes multiple sub-task roles, a matching matrix between sub-task roles and UAVs is established based on the equipment capabilities required by the sub-task roles included in each task, the equipment capabilities of each UAV, and the priority weight of each task, and the sub-task roles are assigned to the corresponding UAVs according to the matching matrix.

[0064] Furthermore, when predetermined dynamic role reassignment conditions are triggered, the role of the drone is dynamically adjusted according to the urgency of the mission and the on-site environment.

[0065] Example 2:

[0066] In this embodiment, the drone is a police drone, and the method of this invention is used to realize intelligent scheduling and collaborative law enforcement of multiple drones.

[0067] The method for intelligent scheduling and collaborative operation of multiple unmanned aerial vehicles (UAVs) in this embodiment of the invention includes the following steps:

[0068] Step 1: Task Access and Real-time Information Aggregation

[0069] The system first receives a set of tasks to be processed in real time from multiple alarm access interfaces (including 110 alarm system, video surveillance platform, and on-duty police feedback). Each task includes information such as task type, event address, expected execution time limit, and required equipment characteristics. Simultaneously, data is continuously collected from each drone in the drone swarm. Real-time running status , including location Remaining battery power Current load capacity and configured device capacity information .

[0070] Step 2: Intelligent Dynamic Task Allocation Algorithm

[0071] Task value scoring model:

[0072]

[0073] in,

[0074] : Mission urgency level;

[0075] Importance of the task;

[0076] Task complexity;

[0077] Decision weight parameters

[0078] The value score of task Ti is a weighted score calculated based on multiple factors such as task priority, urgency, and execution difficulty.

[0079] In one specific implementation, the urgency level of the task is... The tasks are categorized into levels 1-5; this value is based on factors such as the timeliness of the task and the degree of threat to security. Priority is assigned according to preset rules such as the type of incident and emergency response time, or the urgency of the task is assessed based on historical data and statistical models. For example, emergencies such as robbery and fire will receive higher priority (high value), while routine patrol tasks will be given lower priority (low value).

[0080] In this example, Value range: , where 1 represents the lowest priority and 5 represents the highest priority.

[0081] Task importance is used to measure the impact of a task on public safety or the safety of the people. For example, tasks involving personal safety are pre-set as high importance; tasks concerning human life, such as medical emergency and crime fighting, should be assigned a higher importance score.

[0082] In one implementation, the value range is: Ui∈[0,1], where 0 indicates that it is not important and 1 indicates that it is very important.

[0083] Its value is predetermined based on factors such as the type of task and the level of threat at the scene.

[0084] Ci: Task complexity, which can be measured by required resources such as the number of devices. Task complexity considers the amount of resources required to execute the task, equipment requirements, personnel scheduling, and other factors. For example, shooting high-definition video or long-duration reconnaissance missions require more equipment and personnel support.

[0085] In one implementation, its value range is: Ci∈[1,5], where 1 represents a simple task such as simple patrol, and 5 represents a complex task such as deep reconnaissance or coordinated operation of a drone swarm.

[0086] Specifically, the complexity of a task is assessed based on factors such as the difficulty of execution, the required equipment, and time constraints, or by referring to historical data to statistically analyze the complexity of different task types.

[0087] In the specific implementation, the Particle Swarm Optimization (PSO) algorithm is used to adjust the aforementioned weight parameters. , and And to find the optimal task allocation scheme. During the task allocation process, PSO particles continuously iterate to find the optimal solution, thereby maximizing the task value score.

[0088] Unmanned aerial vehicle (UAV) mission suitability model:

[0089]

[0090] in,

[0091] Indicates drone For the task The suitability score is used to calculate the degree of matching between the drone and the mission based on factors such as the drone's location, remaining battery power, and equipment configuration.

[0092] The flight path between the drone and the mission location; that is, the distance from the drone's current location to the mission location, which takes into account the estimated distance of the obstacle avoidance path; in one specific implementation, the path is determined based on the city map, the coordinates of the mission location, and the influence of obstacles such as buildings or no-fly zones, and this distance is calculated using GIS (Geographic Information System) technology or path planning algorithms.

[0093] This score measures whether the drone possesses the necessary equipment and functions to complete the mission. Specific implementations may include features like megaphones, lighting, and zoom lenses. For example, a mission might require a high-definition camera, megaphone, and infrared sensors. In one implementation, the required equipment is assessed based on the nature of the mission, such as reconnaissance or emergency rescue. For instance, patrol missions might not require specialized equipment, while reconnaissance missions might need zoom cameras and thermal imagers. Whether the drone possesses these features is determined by its configuration.

[0094] The remaining battery power of the drone. The remaining battery power will affect whether the drone can complete the mission, especially long-term missions such as reconnaissance and surveillance, which may be unable to be performed due to insufficient battery power.

[0095] Real-time monitoring of the battery status of each drone, or estimation of power consumption by calculating the relationship between mission time, distance, and power consumption. For example, the relationship between flight time and power consumption can be obtained through historical data or battery performance tests.

[0096] In one specific implementation, the range of values ​​is: , where 0 indicates low battery and 1 indicates full battery.

[0097] This invention combines the Mixed Integer Programming (MILP) algorithm with the existing Particle Swarm Optimization (PSO) algorithm to form a multi-objective, multi-constraint optimization task scheduling framework. In the task allocation process, traditional PSO is often used only for optimizing a single objective. However, this invention overcomes the limitations of existing technologies by introducing multiple constraints such as task urgency, UAV capabilities, battery power, and flight time, and constructs a global task allocation optimization model.

[0098] Mixed Integer Programming (MILP) can accurately model the constraints between multiple UAVs and tasks, ensuring that the task allocation for each UAV maximizes task completion rate without overloading, thus guaranteeing task efficiency and the rational use of UAV resources. The combination of MILP and task allocation involves adjusting task weights through PSO in the global task allocation optimization model while using MILP to ensure the constraints during the allocation process. This combination allows for efficient task allocation while also meeting various constraints in real-world applications.

[0099] Global multi-task dynamic allocation optimization model:

[0100] To maximize the overall task benefit, the following mixed-integer programming model is established:

[0101]

[0102] in,

[0103] Indicates task Assigned to drones The decision variable takes the value 0 or 1, indicating whether the drone undertakes the task; 0 indicates that it does not undertake the task, and 1 indicates that it undertakes the task.

[0104] The application of MILP here: Mixed Integer Programming (MILP) is mainly reflected in the constraints. These constraints, modeled using MILP, ensure the feasibility of task allocation while maximizing task completion rates and the rational use of UAV resources.

[0105] constraint:

[0106] Each drone can only perform one task: ;

[0107] The execution of the mission must not exceed the drone's capability limitations, such as payload and flight time.

[0108]

[0109] in, Indicates drone Execute the task The required power, For drones Maximum battery capacity limit.

[0110] The distributed multi-agent reinforcement learning (MARL) algorithm is used for online adaptive training and solving to achieve efficient allocation of large-scale tasks.

[0111] Step 3: Cooperative route planning and collision avoidance algorithm

[0112] After the mission is assigned, the system plans a specific route for each mission drone to ensure flight safety, efficient obstacle avoidance, and coordination of airspace conflicts.

[0113] 3D route modeling:

[0114] This invention, based on traditional two-dimensional path planning algorithms, introduces a three-dimensional grid-based flight path model. This effectively considers the differences in building heights within the urban environment and the altitude limitations of UAV flights, making flight path planning more targeted and adaptable. Urban space is abstracted as a three-dimensional grid map. In this context, node V represents a legal candidate path point (excluding building obstacles), and edge E represents a connection between adjacent reachable nodes.

[0115] In existing technologies, most path planning algorithms (such as A* and Dijkstra's algorithm) focus primarily on path optimization in two-dimensional space, especially in open spaces or simple environments. However, in complex urban environments, particularly in densely populated urban areas with tall buildings, traditional two-dimensional path planning algorithms often cannot handle complex factors such as building height, flight level, and dynamic obstacle avoidance. Therefore, the ability of existing technologies to perform three-dimensional path planning in urban environments is relatively limited, especially lacking three-dimensional rasterized models that can effectively handle flight height restrictions and building obstructions.

[0116] The core innovation of this invention is the abstraction of urban space into a three-dimensional grid map (G(V, E)). This means that the space not only has horizontal coordinates but also considers the height (Z-axis) of each grid cell, thus more accurately representing the constraints of buildings, high-altitude obstacles, and flight levels in the city. This three-dimensional model provides a comprehensive representation of complex urban airspace and can be optimized according to obstacles at different height levels.

[0117] Nodes and edges of a raster graph:

[0118] Node V: Represents a feasible waypoint in path planning, including possible flight levels, i.e., altitudes. The elevation value of the node reflects the flight altitude constraints.

[0119] Edge E: Represents the flight path that the drone may take during flight. The weight of the edge can be assigned based on the flight distance, the risk of the flight level (such as low-altitude flight, risk of densely built-up urban areas), and the safe flight distance.

[0120] Single UAV route optimization objectives:

[0121]

[0122] in,

[0123] k: The index of each segment in the path. In other words, the route is composed of multiple segments, each segment corresponding to the flight path from one node (location) to the next node (location). Therefore, k represents each segment of the journey, and the journey from the starting point to the destination is divided into multiple segments for computational optimization.

[0124] p: The total number of flight segments in the path, i.e., the number of flight segments contained in the complete path from the start point to the end point. Specifically, p is the number of flight segments that need to be optimized in the entire path. In a 3D raster map, a path is composed of multiple raster cells connected together, and the flight path between each raster cell is a flight segment. The final path optimization will be performed on the sum of these flight segments.

[0125] Cost(Segmentk) is the cost of the k-th segment. Cost is the optimization objective, usually expressed as the "cost" of the segment, which is the factor that needs to be considered when the UAV performs flight on that segment, specifically including penalties for flight distance and flight altitude.

[0126] Indicates the length of the flight segment

[0127] This indicates the risk coefficient at a specific height level, such as the penalty for densely built-up areas at lower altitudes.

[0128] 1. Determination method

[0129] The magnitude of the value is related to the following factors:

[0130] Flight altitude: When drones fly at lower altitudes, such as near densely built-up areas, the risks are greater, and the cost of the flight path should increase. This is because low-altitude flights typically face more obstacles, building obstructions, and more complex airflow conditions, thus increasing the difficulty and risk of the flight.

[0131] For example, if a drone flies between high-rise buildings, the buildings may obstruct the view and create obstacles, increasing the cost of the flight path. Therefore, areas with lower flight altitudes will have higher penalty values.

[0132] Urban building density: In areas with dense high-rise buildings, the risk of flight paths is greater, which needs to be adjusted with a penalty factor. If the drone's flight path passes through these dense areas, a higher penalty will be imposed. value.

[0133] Flight area restrictions and regulations: Some urban areas may have flight restrictions, such as no-fly zones and protected areas around airports. These areas usually have a significant impact on the choice of flight paths, therefore, flying in these areas... A higher penalty value is also required to avoid crossing these sensitive areas.

[0134] Types of flight missions: Different missions have different altitude requirements. For example, reconnaissance missions may require higher altitudes, while low-altitude photography missions may require lower altitudes. In such cases, the altitude requirements for different missions will affect the severity of the penalty.

[0135] 2. Range of values

[0136] It is a weighting factor, and its value can usually be set between 0 and 1:

[0137] 0 indicates no penalty; flight altitude has no effect on path cost.

[0138] 1 indicates the maximum penalty; the flight altitude limit has the strongest impact on path cost.

[0139] 3. Factors affecting the magnitude of the value

[0140] The value will be adjusted based on the following factors:

[0141] Flight altitude:

[0142] At lower altitudes, the penalty value increases because low-altitude flight involves more obstacles and higher risks.

[0143] At higher altitudes, the penalty value is lower because flight conditions are better and there are fewer obstacles and less difficulty in flying.

[0144] Building density:

[0145] In densely built-up urban areas, the risks of flying are high, so strong penalties are required for flight paths;

[0146] Flying is safer in open or low-density areas, with lower penalties.

[0147] Flight mission type:

[0148] For high-risk missions such as counter-terrorism and emergency rescue, the flight altitude is often lower, and the penalty for the flight path may be greater.

[0149] For low-risk missions such as routine patrols and environmental monitoring, the flight altitude may be higher, and the corresponding penalties are smaller.

[0150] Airspace control requirements:

[0151] For no-fly zones or areas that require airspace control, flight paths may need to be specially avoided, and such paths will be subject to higher penalty coefficients.

[0152] The penalty coefficient is lower in areas of free flight.

[0153] Indicates a high-penalty weighting factor

[0154] Multi-aircraft route conflict detection:

[0155] This invention employs a multi-drone route conflict detection algorithm. Unlike the simple obstacle avoidance of single-drone route planning in existing technologies, this invention can dynamically detect and adjust the routes of each drone when multiple drones are operating in cooperation, ensuring that no collisions occur when multiple drones are performing tasks simultaneously.

[0156] The system uses a sliding time window. Inspect the overlapping area of ​​the flight paths of any two drones:

[0157]

[0158] Mathematical symbols used to represent existence; The mathematical symbol for representing intersection; Mathematical symbols indicating constraints imposed by subsequent conditions; That is to say and Does a conflict exist? The expression on the right indicates whether there exist segments k and k′ that satisfy the condition that... and Located in the same time window At that time, and The distance between them is less than the predetermined safe distance. If the expression is satisfied, then there is a conflict; otherwise, there is no conflict.

[0159] k and k′: These two parameters represent the flight segment indices. Each drone's flight path is divided into multiple segments, each with its own index. k and k′ represent the individual flight segments of the two drones' flight paths within a time window. For example, if a drone's flight path runs from time T0 to time T1, this path can be divided into multiple segments. Each segment has a time index (e.g., k), indicating its position within the flight path.

[0160] and These two parameters represent the location and time of the flight segment, which is the specific flight path of each segment in a multi-aircraft route. This represents the k-th segment of the flight path of the j-th UAV. Indicates the first The k′ segment of the drone's flight path. By checking whether these two segments overlap, it is determined whether there is a flight path conflict.

[0161] This indicates the predetermined safe distance.

[0162] In multi-drone collaboration, the flight path of each drone changes continuously over time, potentially leading to collisions or conflicts both temporally and spatially. A sliding time window is used to check if the flight paths of two drones overlap in time and space, and dynamically adjusts routes or flight strategies when a conflict is detected to avoid collisions.

[0163] Using a sliding time window How does it work to check the overlapping area of ​​the flight paths of any two drones?

[0164] 1. Time window:

[0165] Set a time window This indicates the time frame within which drone path overlap needs to be considered. Each drone's path changes over time; therefore, within a given time window, the algorithm checks for path overlaps. For drones Flight time in segment k; For drones Flight time in segment k′;

[0166] 2. Route overlap detection:

[0167] For two drones and The sliding time window checks whether their flight paths overlap within that time period. If the distance between their flight paths is less than a set safe margin and their flight time periods overlap, a flight path conflict is considered to exist.

[0168] 3. Dynamically adjust flight routes:

[0169] If a conflict is detected, the system will automatically adjust the flight path of one or more drones to avoid overlapping areas, and adjust flight time or altitude to prevent a collision.

[0170] Conflict resolution and synchronization correction:

[0171] An improved Conflicting Basis Search (ICBS) algorithm is used:

[0172] Adjusting takeoff or flight segment progress through local time offsetting, i.e., time staggering:

[0173]

[0174] This is a drone. The original takeoff time or segment progress time on the k-th flight segment is the start time of the UAV executing the k-th flight segment. For example, suppose the UAV... The plan is to The k-th segment of the journey begins at a specific time. After calculation and conflict detection, this time may need to be adjusted.

[0175] This is the adjusted takeoff time or segment progress time. When a route conflict occurs, the algorithm will modify this value to adjust the drone's flight path. The takeoff time of the kth segment is adjusted to avoid conflict with other drones.

[0176] This is the adjustment amount for staggered flight times. This amount represents the increment in time for the adjustment, which is determined based on the results of conflict detection and airspace capacity to ensure that the takeoff times of the two flight segments do not overlap. It can be a positive or negative value:

[0177] A positive value indicates a delayed takeoff time; a negative value indicates an earlier takeoff time. The magnitude of this adjustment is typically calculated dynamically based on factors such as the drone's flight speed, mission urgency, and airspace availability.

[0178] Temporary altitude corridors are adjusted by local track spatial offset, i.e., altitude misalignment:

[0179]

[0180] in,

[0181] Original flight altitude, representing the drone's altitude. The flight altitude during the k-th segment of the journey.

[0182] The adjusted flight altitude is used to avoid flight path conflicts with other drones.

[0183] The altitude shift adjustment determines whether the flight altitude increases or decreases, ensuring that the flight paths are staggered.

[0184] Recursively insert conflicting nodes and replan until the global conflict resolution converges.

[0185] Step 4: Dynamic Division of Labor among Multi-Machine Roles

[0186] The multi-drone role dynamic task allocation mechanism of this invention dynamically assigns task roles to each drone based on the drone's mission requirements and equipment configuration, such as reconnaissance, broadcasting, and illumination. Unlike the common single-task execution mode in the prior art, this invention can dynamically adjust the drone's role according to the urgency of the mission and the on-site environment, ensuring efficient execution.

[0187] Role definition and device mapping:

[0188] Evidence collection and investigation (reconnaissance aircraft) High-definition zoom cameras are preferred.

[0189] Loudspeaker deterrence (loudspeaker) Equipped with a megaphone device;

[0190] Lighting support (lighting machine) ): Equipped with floodlight components;

[0191] Communication relay (relay machine) It has a relay communication module;

[0192] Role allocation algorithm:

[0193] For each task Subtask requirement matrix Based on the allocated drone equipment capabilities Matching weight:

[0194]

[0195] in, Indicates drone Have you been assigned a task role? .

[0196] This indicates the priority weight of each device submodule. Indicates drone Does it have a role? Required equipment .

[0197] The final result is a role-drone matching matrix, such as the optimal matching matrix, which ensures that the role required for each task is undertaken by the appropriate drone.

[0198] The following example illustrates the role allocation algorithm used in this embodiment of the invention.

[0199] Suppose there are 3 missions and 4 drones. The missions require the roles of reconnaissance, broadcasting, and illumination, respectively. The capabilities and configurations of the drones are as follows:

[0200] Tasks and Roles

[0201] Task T1: Requires a reconnaissance role and a high-definition zoom camera;

[0202] Task T2: Requires a character who can shout out messages, and requires a megaphone;

[0203] Task T3: Requires a lighting character and a high-intensity light projector;

[0204] Drones and Capabilities

[0205] drone equipment

[0206] UAV1 high-definition zoom camera, megaphone

[0207] UAV2 megaphone, lighting components

[0208] UAV3 HD zoom camera, lighting components

[0209] UAV4 lighting components

[0210] Fit calculation

[0211] Fit is calculated based on whether the drone possesses the equipment required for the role. Fit score. Calculate using the allocation algorithm above:

[0212] Adaptability matrix calculation (based on device matching): Table 1 is an example of an adaptability matrix, where the rows correspond to UAV1 to UAV4; the columns correspond to tasks T1 to T3; as shown in Table 1, a value of 1 indicates that the corresponding UAV is adapted to the corresponding task; a value of 0 indicates that the corresponding UAV is not adapted to the corresponding task.

[0213] Table 1: Example of Fit Matrix

[0214]

[0215] The UAV1 is equipped with reconnaissance equipment (high-definition zoom camera) and a loudspeaker, so it can perform mission T1 (reconnaissance) and mission T2 (loudspeaker), but not mission T3 (lighting).

[0216] The UAV2 is equipped with a megaphone and lighting components, making it suitable for mission T2 (megaphone) and mission T3 (lighting).

[0217] The UAV3 is equipped with reconnaissance equipment (high-definition zoom camera) and lighting components, making it suitable for missions T1 (reconnaissance) and T3 (lighting).

[0218] UAV4 can only perform task T3 (lighting).

[0219] Weighted score

[0220] To optimize role allocation based on the urgency and importance of tasks, we can further weight the matrix. For example, we can give task T1 (reconnaissance) a higher weight because it may be an urgent task.

[0221] Assume the weights are as follows:

[0222] Task T1 (Reconnaissance): Urgent and important, weight 0.5

[0223] Task T2 (Shouting): Weight 0.3

[0224] Task T3 (Lighting): Weight 0.2

[0225] Based on the task suitability and task weight, the final score for each UAV is calculated, and the match with the highest score is selected. Table 2 shows an example of a score calculation. Rows correspond to UAVs UAV1 to UAV4; columns correspond to tasks T1 to T3, and the values ​​in the table represent the weight of the corresponding task for the corresponding UAV.

[0226] Table 2: Example of scoring calculation

[0227]

[0228] Final matching matrix

[0229] Based on the total score of each drone, we can select the match with the highest score to ensure the efficiency of task allocation; Table 3 shows an example of drones assigned for each task.

[0230] Table 3: Example of Drone Allocation

[0231]

[0232] Dynamic role reassignment triggering mechanism: When a predetermined dynamic role reassignment condition is triggered, the role of the UAV is dynamically adjusted according to the urgency of the task and the on-site environment; wherein, the predetermined dynamic role reassignment condition includes one or more of the following: current equipment failure alarm and target status change; the target status change includes: target behavior direction change.

[0233] The conditions for dynamic role reassignment are, for example:

[0234] Current equipment failure alarm;

[0235] Sudden changes in the situation, such as a change in the target's escape direction;

[0236] The formation reconstruction automatically notifies the role reassignment module to update the matching matrix and trigger role switching.

[0237] Step 5: Collaborative Task Execution and Information Sharing Mechanism

[0238] During the operation, each mission drone establishes a low-latency data link through a predetermined communication network, such as a 5G private network or an ad hoc network protocol, and periodically broadcasts the following information about the drone: current location. Velocity vector Collect on-site perception data, including: video, audio, images and / or recognition results;

[0239] Role status identifier, used to indicate the current role status, such as normal, abnormal, or switching;

[0240] The command center integrates all UAV data streams in real time to form an overall situational map for tactical command and decision support.

[0241] Step Six: Real-time Adaptive Task Reconstruction Algorithm

[0242] When the predetermined rescheduling conditions are triggered, a local incremental reconfiguration model is used to perform real-time adaptive reconfiguration of the task.

[0243] In one specific implementation, the rescheduling triggering conditions include one or more of the following:

[0244] New emergency calls were received;

[0245] Warning of drone equipment failure or critical battery level;

[0246] The tactical situation on the ground has changed significantly.

[0247] Local incremental reconstruction model:

[0248] The local incremental task reconfiguration mechanism proposed in this invention can quickly update task allocation and flight paths when UAV malfunctions, new tasks are inserted, or emergencies occur, ensuring the continuity and efficiency of task execution. Compared with the static task allocation and path planning of existing technologies, this mechanism has a stronger dynamic adaptability.

[0249] Keep unaffected tasks unchanged, and only perform rapid replanning on affected tasks:

[0250]

[0251] The time cost and adaptation penalty of newly allocated drones; "new" means newly allocated.

[0252] The cost of switching roles during a mission, such as taking over from a backup machine; switch means switching.

[0253] When missions change or UAVs malfunction, adjustments are made only to the affected mission portions, reducing the computational overhead of global reconfiguration. A multi-round local priority queue heuristic search algorithm (Hybrid Local Replanning, HLR) is employed to generate the latest mission assignments and route planning adjustments within a very short latency, ensuring uninterrupted missions and maintaining situational awareness.

[0254] Examples of application scenarios of this invention:

[0255] In the event of a serious brawl in the city's core commercial area, the system receives real-time alarm data, automatically scores and prioritizes the incident, and dispatches nearby drones equipped with loudspeakers to issue warnings from the air. Drones with high-magnification zoom capabilities are quickly deployed for close-range high-altitude reconnaissance and image capture, while drones with lighting components illuminate the scene to facilitate evidence collection. The system globally plans all flight paths synchronously, dynamically avoiding temporarily restricted-fly zones near high-rise buildings to ensure safe, non-overlapping, and conflict-free operation of multiple drones. If a reconnaissance drone experiences low battery during the mission, the system immediately dispatches a standby drone to seamlessly take over, ensuring uninterrupted reconnaissance. If a new type of illegal parking incident occurs in a nearby area, the system balances the remaining workload globally and immediately dispatches nearby idle drones for rapid response and handling, achieving dynamic, multi-emergency adaptive closed-loop control.

[0256] Example 3:

[0257] This invention also provides a device for realizing intelligent scheduling and collaborative work of multiple unmanned aerial vehicles (UAVs), such as... Figure 2 As shown, the device includes a processor 201, a memory 202, a bus 203, and a computer program stored in the memory 202 and executable on the processor 201. The processor 201 includes one or more processing cores. The memory 202 is connected to the processor 201 via the bus 203. The memory 202 is used to store program instructions. When the processor executes the computer program, it implements the steps in the above-described method embodiment of Embodiment 1 of the present invention.

[0258] Furthermore, as an executable solution, the device can be a computer unit, which may be a desktop computer, laptop, handheld computer, or cloud server, etc. The computer unit may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above-described structure of the computer unit is merely an example and does not constitute a limitation on the computer unit. It may include more or fewer components, or combine certain components, or use different components. For example, the computer unit may also include input / output devices, network access devices, buses, etc., and this embodiment of the invention does not limit this.

[0259] Furthermore, as an executable solution, the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc. The processor is the control center of the computer unit, connecting various parts of the entire computer unit via various interfaces and lines.

[0260] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the computer unit by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function; the data storage area may store data created based on the use of the mobile phone, etc. In addition, the memory may include high-speed random access memory and may also include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital card (SD card), flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0261] Example 4:

[0262] The present invention also provides a system for realizing intelligent scheduling and collaborative work of multiple unmanned aerial vehicles (UAVs), comprising: multiple UAVs; and the aforementioned apparatus for realizing intelligent scheduling and collaborative work of multiple UAVs, used for scheduling and collaborative work of the multiple UAVs.

[0263] Example 5:

[0264] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps described above.

[0265] Although the invention has been specifically shown and described in conjunction with preferred embodiments, those skilled in the art should understand that various changes in form and detail may be made to the invention without departing from the spirit and scope of the invention as defined in the appended claims, all of which shall be within the scope of protection of the invention.

Claims

1. A method for implementing intelligent scheduling and cooperative work of multiple unmanned aerial vehicles, characterized in that, Comprising: Step S1: Receive tasks to be processed in real time. And acquire information about each drone in the drone swarm. The real-time operating status includes: current location, remaining battery power, current load capacity, and configured device capacity information; i=1…m, j=1…n; m is the total number of tasks; n is the total number of drones; Step S2, using a pre-established task value scoring model, a UAV task adaptation degree model, and a global multi-task dynamic allocation optimization model to intelligently dynamically allocate tasks; the global multi-task dynamic allocation optimization model is a mixed integer programming model established under predetermined constraints and with the goal of maximizing overall task revenue; wherein, The mixed integer programming model is: ； The task value scoring model is: ； The UAV task adaptation degree model is: ； for the task assigned to the drone decision variable whose value is 0 or 1, 0 indicating that the drone does not undertake the task, 1 indicating that the drone undertakes the task; a value score indicative of the task an urgency level of the task; an importance of the task; a complexity of the task; and decision weight parameters obtained by a particle swarm optimization algorithm;​​ Indicates drone For the task Compatibility score; Indicates drone With the task Flight distance between locations Indicates task With drones The matching score of the required functional equipment for the mission is used to measure the drone's capabilities. Are they capable of completing the task? Required equipment and functions; , and These are predefined weighting parameters; Indicates drone The remaining battery power, , where 0 indicates low battery and 1 indicates full battery; The predetermined constraints of the mixed integer programming model include: each UAV can only perform one task and the execution of a task cannot exceed the capacity limit of the UAV; the capacity includes load and endurance time; Step S3, after task allocation is completed, using a pre-constructed three-dimensional grid route model G(V, E) to plan routes for each UAV performing a task; V is a node, representing a feasible route point, which includes a flight altitude; E is an edge, representing a flight path. 2.The method of claim 1, wherein, The step S3 includes: single UAV route optimization and multi-route conflict detection; wherein, The goal of single UAV route optimization is: k is the index of each leg in the flight path; p is the total number of legs contained in the flight path; is the cost of the kth leg; represents the leg distance length; represents the altitude layer risk coefficient predetermined according to the flight altitude layer, the density of urban buildings, the restrictions and regulations of the flight area, and the type of flight mission; represents the altitude penalty weight factor; Multi-route conflict detection includes: Using a sliding time window checking whether there is a track overlap area between any two UAVs; and and if there is, determining that a flight path conflict is detected; In the case of determining that a route conflict is detected, the takeoff time or the segment progress time of the UAV is adjusted through local time offset or the temporary altitude corridor is adjusted through local space offset. 3.The method of claim 1, wherein, When a task includes multiple sub-task roles, a matching matrix between sub-task roles and UAVs is established according to the equipment capabilities required by each sub-task role, the equipment capabilities possessed by each UAV, and the priority weight of each task, and the sub-task roles are allocated to the corresponding UAVs according to the matching matrix.

4. The method of claim 3, wherein, The sub-task roles include one or more of the following roles: The required equipment capabilities of the reconnaissance role include a high-definition zoom camera; The required equipment capabilities of the shouting role include a megaphone; The required equipment capabilities of the lighting role include a strong light projector.

5. The method of claim 3, wherein, Further comprising: When a predetermined dynamic role reallocation condition is triggered, dynamically adjusting the roles of the UAVs according to the urgency of the task and the on-site environment; Wherein, the predetermined dynamic role reallocation condition includes one or more of the following: a current device failure alarm and a target state change; the target state change includes a target behavior direction change.

6. The method of claim 1, wherein, Further comprising: During task execution, the UAVs performing each task establish a data link through a pre-set communication network and periodically broadcast their current position, velocity vector, collected on-site data, and current role state using the data link.

7. The method for implementing multi-UAV intelligent scheduling and cooperative work according to claim 1, further comprising, when a predetermined rescheduling condition is triggered, using a local incremental reconstruction model to perform real-time adaptive reconstruction of the task; the local incremental reconstruction model keeps the unaffected tasks unchanged and only performs re-planning on the affected tasks using the following formula: represents the time cost of the newly assigned drone generation and the adaptation penalty; new denotes newly assigned; represents the cost incurred by performing a role switch in a task; switch represents a role switch. 8.The method of claim 1, wherein, The cooperation is for cooperative law enforcement; in the step S1, the processing task is received in real time through a multi-source police information access interface The multi-source police information access interface includes one or more of the following: a 110 alarm system, a video monitoring platform, and road enforcement feedback; the processing task includes one or more of the following task information: task type, event address, expected execution time limit, and device characteristics required for the task.

9. An apparatus for implementing multi-unmanned aerial vehicle intelligent scheduling and cooperative work, characterized in that, Comprising a memory and a processor, the memory storing at least one program, the at least one program being executed by the processor to implement the steps of the method for implementing multi-UAV intelligent scheduling and cooperative work according to any one of claims 1 to 8.

10. A system for implementing intelligent scheduling and cooperative work of multiple unmanned aerial vehicles, characterized in that, Comprising: Multiple UAVs; And, The device for realizing multi-unmanned aerial vehicle intelligent scheduling and cooperative work according to claim 9 is used for scheduling and cooperative work of the multiple unmanned aerial vehicles.

Images