A multi-unmanned aerial vehicle cooperative reconnaissance task allocation method and system based on time sequence state perception and hierarchical re-planning, a computer readable storage medium and an electronic device

Abstract

This invention provides a method and system for allocating multi-UAV collaborative reconnaissance tasks based on temporal state perception and hierarchical replanning, a computer-readable storage medium, and an electronic device. The method includes: constructing a multi-UAV collaborative reconnaissance scenario and collecting temporal state data of multiple UAVs to build a multi-UAV temporal state database; under a distributed control architecture, dynamically clustering and electing cluster heads based on distance and communication constraints between UAVs to form a multi-level collaborative organizational structure within and between clusters; analyzing the UAV state time series based on the multi-UAV temporal state database to detect risk events such as new tasks, changes in task positions, UAV damage, loss of contact, and task timeouts, and determining the local areas requiring replanning and the affected UAV and task sets; and adaptively selecting a hierarchical incremental replanning strategy based on the event scale and affected scope to complete the dynamic allocation and closed-loop update of multi-UAV collaborative reconnaissance tasks.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) control technology, specifically to a method and system for allocating multi-UAV collaborative reconnaissance tasks based on temporal state perception and hierarchical replanning, a computer-readable storage medium, and an electronic device. Background Technology

[0002] In recent years, multi-UAV collaborative reconnaissance systems have been widely used in military and civilian fields such as battlefield situational surveillance, routine border patrols, natural disaster emergency assessments, and security inspections of key areas, thanks to their advantages of flexible deployment, wide coverage, and strong survivability. Compared with single-UAV operation mode, multi-UAV systems, through reasonable task allocation and collaborative trajectory planning, can complete a larger area coverage within a limited time, significantly improving the probability of target detection and mission execution efficiency.

[0003] Currently, existing technologies for multi-drone task allocation can be mainly divided into three categories: First, centralized global optimization methods, which solve for the globally optimal allocation scheme through integer programming, genetic algorithms, etc. Although this type of method can obtain the theoretical optimal solution, the computational cost increases exponentially with the number of drones and tasks, and is only applicable to static scenarios. Second, distributed allocation methods, which realize distributed task negotiation based on auction algorithms, game theory, consensus algorithms, etc., reduce the pressure of centralized computing, but have problems such as high communication overhead, easy to get trapped in local optima, and difficulty in ensuring global constraint satisfaction. Third, dynamic and static combined planning methods, which make local adjustments based on the initial global planning to deal with environmental changes. However, existing local adjustment strategies mostly rely on greedy rules and lack a systematic perception of the drone's temporal state, making it impossible to accurately judge the scope of event impact. This either frequently triggers global replanning, resulting in insufficient real-time performance, or the local adjustment oversimplifies constraints, which can easily lead to problems such as task timeout, drone energy overload, and local task congestion.

[0004] In real-world collaborative reconnaissance scenarios, dynamic events such as the addition / location change of mission points, drone damage / loss of contact, rapid depletion of remaining energy, and sudden communication interruptions occur frequently, causing the initial static allocation scheme to quickly become ineffective. Existing technologies generally lack a continuous situational awareness mechanism based on multi-drone time-series status data, and have not established a unified framework for incremental replanning triggered hierarchically according to event type and impact scope. This makes it difficult to simultaneously address the real-time nature of task allocation, constraint satisfaction, and computational complexity control in dynamic environments, thus failing to meet the engineering application requirements of multi-drone collaborative reconnaissance in complex and uncertain scenarios. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for multi-UAV collaborative reconnaissance task allocation based on temporal state perception and hierarchical replanning, a computer-readable storage medium, and an electronic device, so as to solve the above-mentioned problems existing in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A multi-UAV cooperative reconnaissance task allocation method based on temporal state awareness and hierarchical replanning includes the following steps: S1. Construct a multi-UAV collaborative reconnaissance scenario, collect time-series status data of multiple UAVs at a preset sampling period, and construct a multi-UAV time-series status database; S2. Under the distributed control architecture, dynamic clustering and cluster head election are performed based on the distance and communication constraints between UAVs to form a multi-level collaborative organizational structure within and between clusters, and the cluster structure and inter-cluster communication topology are maintained in real time. S3. Analyze the time series of UAV status based on the multi-UAV time series status database, detect risk events such as new tasks, changes in task location, UAV damage, loss of contact and task timeout, and determine the local areas that need to be replanned and the affected UAV and task sets. S4. Based on the event scale and the scope of impact, an adaptive hierarchical incremental replanning strategy is selected, which includes single-machine replanning, intra-cluster replanning, inter-cluster replanning, or global replanning. Within the selected level, the task sequence of the corresponding UAV is locally optimized and adjusted, an updated task allocation scheme is generated and sent to the relevant UAVs for execution, and the dynamic allocation and closed-loop update of multi-UAV collaborative reconnaissance tasks are completed.

[0007] In one embodiment, step S1 specifically includes: S11. Construct a multi-UAV collaborative reconnaissance scenario, setting the multi-UAV set as... The reconnaissance missions are as follows The initial position, initial velocity, initial remaining energy, maximum endurance energy, payload capacity, and mission time window constraints of each UAV; S12. During task execution, a preset sampling period is used. Record the three-dimensional position, flight speed, remaining energy, current service task point and completed task set of each UAV to obtain the original time-series state dataset. Perform abnormal data filtering and interpolation to complete the original time-series state dataset to obtain a continuous UAV state sequence. S13, Definition of the A drone at all times The state vector is: in, For the first A drone at all times The three-dimensional spatial coordinates, For the first A drone at all times Flight speed, For the first A drone at all times The remaining energy; concatenate the state vectors of all UAVs to obtain the overall system state vector. : Based on the overall system state vector, a discrete-time dynamic model of the motion state of multiple UAVs and a corresponding observation model are constructed to describe the evolution of the UAV state over time and the observable output.

[0008] In one embodiment, the discrete-time dynamic model and observation model of the multiple UAVs in step S13 are specifically as follows: Discrete-time dynamic model of the state evolution of multiple UAVs over time: in, Indicates time The overall system state vector. Here is the state transition matrix. For the control matrix, For a moment The control input vector, The noise is the process noise, which follows a mean of 0 and a covariance of . Gaussian distribution; Corresponding observation model: in, For a moment The system observation vector; The observation matrix; The observed noise follows a pattern with a mean of 0 and a covariance of . The Gaussian distribution.

[0009] In one embodiment, step S2 specifically includes: S21. Based on the three-dimensional position coordinates of the UAVs in step S13, calculate the coordinates of any two UAVs. and At any moment Euclidean distance The calculation formula is: Within the preset communication radius Constructing an adjacency matrix among UAVs under constraints ,like ≤ Then the adjacent matrix elements =1, otherwise =0; S22. Based on the adjacency matrix, combined with the remaining energy of the UAV, the average distance to neighboring nodes, and the task load index, calculate the cluster head election weight of each UAV. The cluster head election is completed by comparing the weights, and the multiple UAVs are divided into several local clusters. Within each cluster, the cluster head UAV is responsible for intra-cluster task coordination and inter-cluster information interaction. S23. During task execution, monitor the cluster size, cluster radius, and remaining energy of each cluster in real time, preset re-clustering trigger conditions, and when the monitoring data meets the re-clustering trigger conditions, execute the re-clustering operation, update the cluster structure and inter-cluster communication topology, and realize distributed dynamic clustering management.

[0010] In one embodiment, the formula for calculating the cluster head election weight in step S22 is: in, For drones At any moment Cluster head weights, For drones At any moment The remaining energy, For the drone's rated maximum energy, For drones The average distance to its neighboring nodes , These are the preset non-negative weighting coefficients. To prevent extremely small constants with a denominator of zero; The re-clustering condition in step S23 is: satisfying , , Any one of them, where, For the first The number of nodes in a cluster This is the preset threshold for the maximum number of nodes in a single cluster. For the first The communication radius of a cluster, This is the preset maximum communication radius threshold for a single cluster. For the first The remaining energy of the cluster head of each cluster. This is the preset minimum operating energy threshold for the cluster head.

[0011] In one embodiment, step S3 specifically includes: S31. Extract the trajectory sequences of each UAV from the multi-UAV time-series status database. Task execution state sequence and remaining energy sequence Among them, trajectory sequence ; S32, for each task Set reconnaissance time window ,in, This is the earliest time when reconnaissance can be conducted for the mission. For the latest reconnaissance moment of the mission, various dynamic events are identified based on preset criteria: When the task is predicted or the actual completion time satisfy When this occurs, it is marked as a task timeout risk event. The preset timeout warning threshold; When the task set is added or removed, or when the task location changes, it is marked as a task addition / location change event; When the remaining energy of the drone is lower than the preset safe return threshold or the continuous loss of contact exceeds the preset duration, it is marked as a drone damage / failure event. S33. For the identified dynamic events, analyze the spatial impact range of the events and determine the subset of drones affected. and task subset Based on the cluster structure involved, the events are labeled as single-machine events, intra-cluster events, or inter-cluster events, and used as inputs for incremental replanning.

[0012] In one embodiment, S4 specifically includes: S41. Adaptively match the replanning hierarchy based on the event type and the affected scope: When the affected subset of drones When only a single drone is involved, enable single-drone replanning; When the affected subset of drones When all elements are within the same cluster, intra-cluster replanning is enabled. When the affected subset of drones When spanning multiple clusters, inter-cluster replanning should be enabled; When local replanning cannot meet the mission time window constraints or UAV energy constraints, global replanning is enabled. S42. Within the selected replanning level, for the affected subset of drones... With task subset Construct a local task allocation subproblem and introduce binary decision variables: Define the task allocation state before replanning as follows: drones To the mission The flight path distance is The unit task adjustment cost coefficient is drones Execute the task Energy consumption is A task sequence optimization model is established with the preconditions of unique task allocation, UAV energy constraint, and task time window constraint. The task sequence optimization model aims to minimize the comprehensive cost function. S43. Under the premise of satisfying the constraints, the task sequence optimization model is solved by heuristic search or local search algorithm, an updated task allocation scheme is generated, and the scheme is sent to the relevant UAVs for execution. The execution results are written back to the multi-UAV time-series state database to complete the closed-loop iteration.

[0013] In one embodiment, the comprehensive cost function in step S42 is: in, , , These are the preset non-negative weighting coefficients. The total flight path length cost or total energy consumption cost for multiple UAVs performing a mission is calculated using the following formula: This is a penalty for exceeding the preset time limit when the task is completed. The calculation formula for the task sequence adjustment and communication overhead caused by replanning is as follows: In one embodiment, the constraints of the task sequence optimization model specifically include: Unique task allocation constraint: This ensures that each reconnaissance mission is assigned to only one drone. Drone energy constraints: ,in, For the drone in step S13 At any moment The remaining energy; Task time window constraints: This ensures that the mission is completed within the reconnaissance time window set in step S32.

[0014] This invention also provides a multi-UAV cooperative reconnaissance task allocation system based on temporal state awareness and hierarchical incremental replanning, used to execute the above-mentioned multi-UAV cooperative reconnaissance task allocation method based on temporal state awareness and hierarchical incremental replanning, the system comprising: The time-series state database construction module is used to build a multi-UAV collaborative reconnaissance scenario. It collects time-series state data of multiple UAVs at a preset sampling period, completes data preprocessing and state modeling, and builds a multi-UAV time-series state database. The distributed dynamic clustering module is used to construct adjacency relationships based on distance and communication constraints between UAVs under a distributed control architecture, perform cluster head election and dynamic clustering, form a multi-level collaborative organizational structure, and maintain the cluster structure and inter-cluster communication topology in real time. The time-series status perception module is used to extract the time series of UAV status from the multi-UAV time-series status database, identify new tasks, changes in task location, UAV damage / loss of contact and risk events of task timeout, and determine the scope of the event's impact and the set of UAVs and tasks affected. The hierarchical incremental replanning module is used to adaptively match the replanning level according to the event scale and the affected scope, build and solve a local task allocation optimization model within the selected level, and generate an updated task allocation scheme. The task execution and evaluation module is used to distribute task allocation plans to relevant UAVs for execution, collect execution feedback data and write it back to the time-series status database, and at the same time calculate the task completion rate, energy utilization rate and replanning cost to complete system performance evaluation and parameter optimization.

[0015] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method for allocating multi-UAV collaborative reconnaissance tasks based on temporal state perception and hierarchical incremental replanning.

[0016] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-described method for multi-UAV cooperative reconnaissance task allocation based on temporal state perception and hierarchical incremental replanning.

[0017] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention constructs a multi-UAV time-series state database and performs unified discrete-time dynamic modeling of the UAV's position, speed, remaining energy, and mission execution status. This enables continuous and accurate situational awareness of the entire process of multi-UAV collaborative reconnaissance, providing reliable data support for mission allocation and adjustment in dynamic environments and significantly improving the system's response sensitivity to dynamic environmental changes.

[0018] 2. This invention adopts a distributed control architecture and a dynamic clustering mechanism. By combining the cluster head election strategy based on remaining energy, spatial distance, and task load, it realizes the adaptive hierarchical organization of UAV swarms. While ensuring the overall coordination and consistency of the system, it pushes a large number of computing and communication requirements to be completed within the cluster, which significantly reduces the communication load and computing overhead caused by centralized global scheduling, and improves the scalability and operating efficiency of large-scale UAV swarms in complex scenarios.

[0019] 3. The event detection mechanism based on time-series state sequences of the present invention can accurately identify various dynamic events such as new tasks / location changes, drone damage / loss of contact, and task timeout risks, and quantify and analyze the spatial impact range of the events, thereby achieving precision in replanning triggering, avoiding the performance loss caused by indiscriminate global replanning, and eliminating the constraint violation problem caused by insufficient local adjustments.

[0020] 4. The hierarchical incremental replanning strategy of the present invention can adaptively match the optimal replanning level according to the event scale and impact range, and only optimize and adjust the affected local UAVs and task sets. Under the premise of strictly meeting the task time window constraints and UAV energy constraints, it significantly reduces the computational complexity and response latency of replanning, and significantly improves the completion rate and system operation stability of multi-UAV collaborative reconnaissance missions in dynamic environments. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall process of the multi-UAV cooperative reconnaissance task allocation method based on temporal state perception and hierarchical incremental replanning of the present invention. Figure 2 This is a schematic diagram illustrating the triggering and execution logic of the hierarchical incremental replanning strategy of the present invention; Figure 3 This is a schematic diagram of the hierarchical collaborative organization structure of multiple unmanned aerial vehicles under the distributed control architecture of the present invention; Figure 4 This is a time-series simulation diagram of the execution process of a multi-UAV collaborative reconnaissance mission in an embodiment of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] In the description of this invention, it should be noted that the terms "upper," "lower," "inner," "outer," "front end," "rear end," "both ends," "one end," and "the other end," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0024] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installed," "equipped with," "connected," etc., should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0025] Example 1 like Figures 1 to 4 As shown in this embodiment, the multi-UAV cooperative reconnaissance task allocation method based on temporal state awareness and hierarchical replanning includes the following steps: S1. Construction of Multi-UAV Collaborative Reconnaissance Scenarios and Establishment of Temporal Status Database To construct a multi-UAV collaborative reconnaissance scenario, time-series state data of multiple UAVs are collected at a preset sampling period to build a multi-UAV time-series state database. The specific steps are as follows: S11. Construct a multi-UAV collaborative reconnaissance scenario, setting the multi-UAV set as... The reconnaissance missions are as follows The initial position, initial velocity, initial remaining energy, maximum endurance energy, payload capacity, and mission time window constraints of each UAV; S12. During task execution, a preset sampling period is used. (0.5s) Record the three-dimensional position, flight speed, remaining energy, current service task point and completed task set of each UAV to obtain the original time-series state dataset. Perform abnormal data filtering and interpolation to complete the original time-series state dataset to obtain a continuous UAV state sequence. S13, Definition of the A drone at all times The state vector is: in, For the first A drone at all times The three-dimensional spatial coordinates, For the first A drone at all times Flight speed, For the first A drone at all times The remaining energy; concatenate the state vectors of all UAVs to obtain the overall system state vector. : Based on the overall system state vector, a discrete-time dynamic model of the motion state of multiple UAVs and a corresponding observation model are constructed to describe the evolution process of the UAV state over time and the observable output. The discrete-time dynamic model and observation model for multiple UAVs are as follows: Discrete-time dynamic model of the state evolution of multiple UAVs over time: in, Indicates time The overall system state vector. Here is the state transition matrix. For the control matrix, For a moment The control input vector, The noise is the process noise, which follows a mean of 0 and a covariance of . Gaussian distribution; Corresponding observation model: in, For a moment The system observation vector; The observation matrix; The observed noise follows a pattern with a mean of 0 and a covariance of . The Gaussian distribution.

[0026] S2, Dynamic Clustering and Collaborative Organization Construction under Distributed Control Architecture Under the distributed control architecture, dynamic clustering and cluster head election are performed based on the distance and communication constraints between UAVs, forming a multi-level collaborative organizational structure within and between clusters, and maintaining the cluster structure and inter-cluster communication topology in real time. The specific steps are as follows: S21. Based on the three-dimensional position coordinates of the UAVs in step S13, calculate the coordinates of any two UAVs. and At any moment Euclidean distance The calculation formula is: Within the preset communication radius Constructing an adjacency matrix among UAVs under constraints ,like ≤ Then the adjacent matrix elements =1, otherwise =0; S22. Based on the adjacency matrix, combined with the remaining energy of the UAV, the average distance to neighboring nodes, and the task load index, calculate the cluster head election weight of each UAV. The cluster head election is completed by comparing the weights, and the multiple UAVs are divided into several local clusters. Within each cluster, the cluster head UAV is responsible for intra-cluster task coordination and inter-cluster information interaction. The formula for calculating the cluster head election weight is: in, For drones At any moment Cluster head weights, For drones At any moment The remaining energy, For the drone's rated maximum energy, For drones The average distance to its neighboring nodes , These are the preset non-negative weighting coefficients. To prevent extremely small constants with a denominator of zero; S23. During the task execution process, monitor the cluster size, cluster radius and remaining energy of the cluster head of each cluster in real time, preset the re-clustering trigger conditions, and when the monitoring data meets the re-clustering trigger conditions, execute the re-clustering operation, update the cluster structure and inter-cluster communication topology, and realize distributed dynamic clustering management. The condition for re-clustering is: satisfying , , Any one of them, where, For the first The number of nodes in a cluster This is the preset threshold for the maximum number of nodes in a single cluster. For the first The communication radius of a cluster, This is the preset threshold for the maximum communication radius of a single cluster. For the first The remaining energy of the cluster head of each cluster. This is the preset minimum operating energy threshold for the cluster head.

[0027] S3. Event detection and replanning triggering based on temporal state awareness Based on a multi-UAV time-series state database, the system analyzes UAV state time series to detect risk events such as new missions, changes in mission location, UAV damage, loss of contact, and mission timeouts. This identifies local areas requiring replanning, as well as the affected UAV and mission sets. The specific steps are as follows: S31. Extract the trajectory sequences of each UAV from the multi-UAV time-series status database. Task execution state sequence and remaining energy sequence Among them, trajectory sequence ; S32, for each task Set reconnaissance time window ,in, This is the earliest time when reconnaissance can be conducted for the mission. For the latest reconnaissance moment of the mission, various dynamic events are identified based on preset criteria: When the task is predicted or the actual completion time satisfy When this occurs, it is marked as a task timeout risk event. The preset timeout warning threshold; When the task set is added or removed, or when the task location changes, it is marked as a task addition / location change event; When the remaining energy of the drone is lower than the preset safe return threshold or the continuous loss of contact exceeds the preset duration, it is marked as a drone damage / failure event. S33. For the identified dynamic events, analyze the spatial impact range of the events and determine the subset of drones affected. and task subset Based on the cluster structure involved, the events are labeled as single-machine events, intra-cluster events, or inter-cluster events, and used as inputs for incremental replanning.

[0028] S4, Multi-level Incremental Collaborative Task Replanning Based on the event scale and affected scope, a hierarchical incremental replanning strategy is adaptively selected, employing single-drone replanning, intra-cluster replanning, inter-cluster replanning, or global replanning. Within the selected level, the task sequences of the corresponding UAVs are locally optimized and adjusted, generating an updated task allocation scheme and distributing it to the relevant UAVs for execution. This completes the dynamic allocation and closed-loop update of multi-UAV collaborative reconnaissance tasks. The specific steps are as follows: S41. Adaptively match the replanning hierarchy based on the event type and the affected scope: When the affected subset of drones When only a single drone is involved, enable single-drone replanning; When the affected subset of drones When all elements are within the same cluster, intra-cluster replanning is enabled. When the affected subset of drones When spanning multiple clusters, inter-cluster replanning should be enabled; When local replanning cannot meet the mission time window constraints or UAV energy constraints, global replanning is enabled. S42. Within the selected replanning level, for the affected subset of drones... With task subset Construct a local task allocation subproblem and introduce binary decision variables: Define the task allocation state before replanning as follows: drones To the mission The flight path distance is The unit task adjustment cost coefficient is drones Execute the task Energy consumption is A task sequence optimization model is established with the preconditions of unique task allocation, UAV energy constraint, and task time window constraint. The task sequence optimization model aims to minimize the comprehensive cost function. S43. Under the premise of satisfying the constraints, the task sequence optimization model is solved by heuristic search or local search algorithm, an updated task allocation scheme is generated, and the scheme is sent to the relevant UAVs for execution. The execution results are written back to the multi-UAV time-series state database to complete the closed-loop iteration.

[0029] The overall cost function is: in, , , These are the preset non-negative weighting coefficients. The total flight path length cost or total energy consumption cost for multiple UAVs performing a mission is calculated using the following formula: This is a penalty for exceeding the preset time limit when the task is completed. The calculation formula for the task sequence adjustment and communication overhead caused by replanning is as follows: The specific constraints of the task sequence optimization model are as follows: Unique task allocation constraint: This ensures that each reconnaissance mission is assigned to only one drone. Drone energy constraints: ,in, For the drone in step S13 At any moment The remaining energy; Task time window constraints: This ensures that the mission is completed within the reconnaissance time window set in step S32.

[0030] Example 2 This embodiment discloses a multi-UAV cooperative reconnaissance task allocation system based on temporal state awareness and hierarchical incremental replanning, used to execute the multi-UAV cooperative reconnaissance task allocation method based on temporal state awareness and hierarchical replanning described in Embodiment 1 above. The system includes: The time-series state database construction module is used to build a multi-UAV collaborative reconnaissance scenario. It collects time-series state data of multiple UAVs at a preset sampling period, completes data preprocessing and state modeling, and builds a multi-UAV time-series state database. The distributed dynamic clustering module is used to construct adjacency relationships based on distance and communication constraints between UAVs under a distributed control architecture, perform cluster head election and dynamic clustering, form a multi-level collaborative organizational structure, and maintain the cluster structure and inter-cluster communication topology in real time. The time-series status perception module is used to extract the time series of UAV status from the multi-UAV time-series status database, identify new tasks, changes in task location, UAV damage / loss of contact and risk events of task timeout, and determine the scope of the event's impact and the set of UAVs and tasks affected. The hierarchical incremental replanning module is used to adaptively match the replanning level according to the event scale and the affected scope, build and solve a local task allocation optimization model within the selected level, and generate an updated task allocation scheme. The task execution and evaluation module is used to distribute task allocation plans to relevant UAVs for execution, collect execution feedback data and write it back to the time-series status database, and at the same time calculate the task completion rate, energy utilization rate and replanning cost to complete system performance evaluation and parameter optimization.

[0031] Example 3 The computer-readable storage medium disclosed in this embodiment stores a computer program. When the computer program is executed by a processor, it implements the multi-UAV cooperative reconnaissance task allocation method based on temporal state perception and hierarchical incremental replanning as described in Embodiment 1 above.

[0032] Example 4 The electronic device disclosed in this embodiment includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the multi-UAV cooperative reconnaissance task allocation method based on temporal state perception and hierarchical incremental replanning as described in Embodiment 1 above.

[0033] Any aspects of this invention not described in detail are well-known to those skilled in the art.

[0034] Finally, it should be noted that the above specific embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the embodiments, those skilled in the art should understand that modifications and equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications and substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for allocating multi-UAV cooperative reconnaissance tasks based on temporal state perception and hierarchical replanning, characterized in that, Includes the following steps: S1. Construct a multi-UAV collaborative reconnaissance scenario, collect time-series status data of multiple UAVs at a preset sampling period, and construct a multi-UAV time-series status database; S2. Under the distributed control architecture, dynamic clustering and cluster head election are performed based on the distance and communication constraints between UAVs to form a multi-level collaborative organizational structure within and between clusters, and the cluster structure and inter-cluster communication topology are maintained in real time. S3. Analyze the time series of UAV status based on the multi-UAV time series status database, detect risk events such as new tasks, changes in task location, UAV damage, loss of contact and task timeout, and determine the local areas that need to be replanned and the affected UAV and task sets. S4. Based on the event scale and the scope of impact, an adaptive hierarchical incremental replanning strategy is selected, which includes single-machine replanning, intra-cluster replanning, inter-cluster replanning, or global replanning. Within the selected level, the task sequence of the corresponding UAV is locally optimized and adjusted, an updated task allocation scheme is generated and sent to the relevant UAVs for execution, and the dynamic allocation and closed-loop update of multi-UAV collaborative reconnaissance tasks are completed.

2. The multi-UAV cooperative reconnaissance task allocation method according to claim 1, characterized in that, Step S1 specifically includes: S11. Construct a multi-UAV collaborative reconnaissance scenario, setting the multi-UAV set as... The reconnaissance missions are as follows The initial position, initial velocity, initial remaining energy, maximum endurance energy, payload capacity, and mission time window constraints of each UAV; S12. During task execution, a preset sampling period is used. Record the three-dimensional position, flight speed, remaining energy, current service task point and completed task set of each UAV to obtain the original time-series state dataset. Perform abnormal data filtering and interpolation to complete the original time-series state dataset to obtain a continuous UAV state sequence. S13, Definition of the A drone at all times The state vector is: in, For the first A drone at all times The three-dimensional spatial coordinates, For the first A drone at all times Flight speed, For the first A drone at all times The remaining energy; concatenate the state vectors of all UAVs to obtain the overall system state vector. : Based on the overall system state vector, a discrete-time dynamic model of the motion state of multiple UAVs and a corresponding observation model are constructed to describe the evolution of the UAV state over time and the observable output.

3. The multi-UAV collaborative reconnaissance task allocation method according to claim 2, characterized in that, The discrete-time dynamic model and observation model for multiple UAVs in step S13 are as follows: Discrete-time dynamic model of the state evolution of multiple UAVs over time: in, Indicates time The overall system state vector. Here is the state transition matrix. For the control matrix, For a moment The control input vector, The noise is the process noise, which follows a mean of 0 and a covariance of . Gaussian distribution; Corresponding observation model: in, For a moment The system observation vector; The observation matrix; The observed noise follows a pattern with a mean of 0 and a covariance of . The Gaussian distribution.

4. The multi-UAV cooperative reconnaissance task allocation method according to claim 2, characterized in that, Step S2 specifically includes: S21. Based on the three-dimensional position coordinates of the UAVs in step S13, calculate the coordinates of any two UAVs. and At any moment Euclidean distance The calculation formula is: Within the preset communication radius Constructing an adjacency matrix among UAVs under constraints ,like ≤ Then the adjacent matrix elements =1, otherwise =0; S22. Based on the adjacency matrix, combined with the remaining energy of the UAV, the average distance to neighboring nodes, and the task load index, calculate the cluster head election weight of each UAV. The cluster head election is completed by comparing the weights, and the multiple UAVs are divided into several local clusters. Within each cluster, the cluster head UAV is responsible for intra-cluster task coordination and inter-cluster information interaction. S23. During task execution, monitor the cluster size, cluster radius, and remaining energy of each cluster in real time, preset re-clustering trigger conditions, and when the monitoring data meets the re-clustering trigger conditions, execute the re-clustering operation, update the cluster structure and inter-cluster communication topology, and realize distributed dynamic clustering management. The formula for calculating the cluster head election weight in step S22 is as follows: in, For drones At any moment Cluster head weights, For drones At any moment The remaining energy, For the drone's rated maximum energy, For drones The average distance to its neighboring nodes , These are the preset non-negative weighting coefficients. To prevent extremely small constants with a denominator of zero; The re-clustering condition in step S23 is: satisfying , , Any one of them, where, For the first The number of nodes in a cluster This is the preset threshold for the maximum number of nodes in a single cluster. For the first The communication radius of a cluster, This is the preset threshold for the maximum communication radius of a single cluster. For the first The remaining energy of the cluster head of each cluster. This is the preset minimum operating energy threshold for the cluster head.

5. The multi-UAV collaborative reconnaissance task allocation method according to claim 2, characterized in that, Step S3 specifically includes: S31. Extract the trajectory sequences of each UAV from the multi-UAV time-series status database. Task execution state sequence and remaining energy sequence Among them, trajectory sequence ; S32, for each task Set reconnaissance time window ,in, This is the earliest time when reconnaissance can be conducted for the mission. For the latest reconnaissance moment of the mission, various dynamic events are identified based on preset criteria: When the task is predicted or the actual completion time satisfy When this occurs, it is marked as a task timeout risk event. The preset timeout warning threshold; When the task set is added or removed, or when the task location changes, it is marked as a task addition / location change event; When the remaining energy of the drone is lower than the preset safe return threshold or the continuous loss of contact exceeds the preset duration, it is marked as a drone damage / failure event. S33. For the identified dynamic events, analyze the spatial impact range of the events and determine the subset of drones affected. and task subset Based on the cluster structure involved, the events are labeled as single-machine events, intra-cluster events, or inter-cluster events, and used as inputs for incremental replanning.

6. The multi-UAV cooperative reconnaissance task allocation method according to claim 5, characterized in that, S4 specifically includes: S41. Adaptively match the replanning hierarchy based on the event type and the affected scope: When the affected subset of drones When only a single drone is involved, enable single-drone replanning; When the affected subset of drones When all elements are within the same cluster, intra-cluster replanning is enabled. When the affected subset of drones When spanning multiple clusters, inter-cluster replanning should be enabled; When local replanning cannot meet the mission time window constraints or UAV energy constraints, global replanning is enabled. S42. Within the selected replanning level, for the affected subset of drones... With task subset Construct a local task allocation subproblem and introduce binary decision variables: Define the task allocation state before replanning as follows: drones To the mission The flight path distance is The unit task adjustment cost coefficient is drones Execute the task Energy consumption is A task sequence optimization model is established with the preconditions of unique task allocation, UAV energy constraint, and task time window constraint. The task sequence optimization model aims to minimize the comprehensive cost function. S43. Under the premise of satisfying the constraints, the task sequence optimization model is solved by heuristic search or local search algorithm, an updated task allocation scheme is generated, and the scheme is sent to the relevant UAVs for execution. The execution results are written back to the multi-UAV time-series state database to complete the closed-loop iteration.

7. The multi-UAV cooperative reconnaissance task allocation method according to claim 6, characterized in that, The comprehensive cost function in step S42 is: in, , , These are the preset non-negative weighting coefficients. The total flight path length cost or total energy consumption cost for multiple UAVs performing a mission is calculated using the following formula: This is a penalty for exceeding the preset time limit when the task is completed. The calculation formula for the task sequence adjustment and communication overhead caused by replanning is as follows: The constraints of the task sequence optimization model specifically include: Unique task allocation constraint: This ensures that each reconnaissance mission is assigned to only one drone. Drone energy constraints: ,in, For the drone in step S13 At any moment The remaining energy; Task time window constraints: This ensures that the mission is completed within the reconnaissance time window set in step S32.

8. A multi-UAV cooperative reconnaissance task allocation system based on temporal state perception and hierarchical replanning, characterized in that, The system is used to perform the multi-UAV cooperative reconnaissance task allocation method according to any one of claims 1-8, the system comprising: The time-series state database construction module is used to build a multi-UAV collaborative reconnaissance scenario. It collects time-series state data of multiple UAVs at a preset sampling period, completes data preprocessing and state modeling, and builds a multi-UAV time-series state database. The distributed dynamic clustering module is used to construct adjacency relationships based on distance and communication constraints between UAVs under a distributed control architecture, perform cluster head election and dynamic clustering, form a multi-level collaborative organizational structure, and maintain the cluster structure and inter-cluster communication topology in real time. The time-series status perception module is used to extract the time series of UAV status from the multi-UAV time-series status database, identify new tasks, changes in task location, UAV damage / loss of contact and risk events of task timeout, and determine the scope of the event's impact and the set of UAVs and tasks affected. The hierarchical incremental replanning module is used to adaptively match the replanning level according to the event scale and the affected scope, build and solve a local task allocation optimization model within the selected level, and generate an updated task allocation scheme. The task execution and evaluation module is used to distribute task allocation plans to relevant UAVs for execution, collect execution feedback data and write it back to the time-series status database, and at the same time calculate the task completion rate, energy utilization rate and replanning cost to complete system performance evaluation and parameter optimization.

9. A computer-readable storage medium, characterized in that, When the computer program is executed by the processor, it implements the multi-UAV cooperative reconnaissance task allocation method based on temporal state perception and hierarchical incremental replanning as described in any one of claims 1-7.

10. An electronic device, characterized in that, The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the multi-UAV cooperative reconnaissance task allocation method based on temporal state awareness and hierarchical incremental replanning as described in any one of claims 1-7.

Images