A low-energy consumption joint optimization method for a UAV group performing a reconnaissance task

Abstract

This invention discloses a low-energy joint optimization method for UAV swarms performing communication and reconnaissance computing tasks. The method includes: acquiring UAV swarm status information, user communication and reconnaissance computing task requirements, and system resource constraints within the task area; and constructing a low-energy joint optimization model with the objective of minimizing total system energy consumption. Based on the model, the method implements the following: given the UAV deployment locations, it obtains the matching association results between UAVs and user communication and reconnaissance computing tasks using distance edge weights and auction matching algorithms; given the matching association results, it determines the allocation results of communication resources, detection resources, and computing resources; and updates the UAV deployment locations using particle swarm optimization, iteratively solving until convergence conditions are met. This yields the joint optimization results for output task matching association, resource allocation, and deployment locations. This invention can reduce the energy consumption of UAV swarm systems under multi-service concurrency conditions, improve the utilization efficiency of communication and reconnaissance computing resources, and enhance the adaptability of UAV swarm deployment.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicle (UAV) swarm control technology, specifically relating to a low-energy joint optimization method for UAV swarms performing communication and reconnaissance tasks, a computer-readable storage medium, and an electronic device. Background Technology

[0002] In complex areas where ground base stations are limited, ground users typically have multiple mission requirements, including temporary communication access, key area reconnaissance, and near-end computing. Unmanned aerial vehicles (UAVs) offer advantages such as maneuverability, rapid deployment, and adjustable coverage, making them suitable as aerial service platforms to provide ground users with communication, reconnaissance, and computing services.

[0003] In existing technologies, separate models are typically established and optimized for problems such as UAV-assisted communication, UAV disaster monitoring, and UAV-assisted mobile edge computing. However, in real-world continuous support scenarios, communication, detection, and computing tasks often coexist, and a single user may simultaneously request one or more services from these three areas. Optimizing for a single service is insufficient to meet the demands of multi-service collaborative services.

[0004] Given the limited energy and onboard resources of drones, optimizing task association, resource allocation, or deployment location in isolation can easily lead to problems such as unreasonable allocation of drone service tasks, low resource utilization, high energy consumption during drone maneuvers, and difficulty in guaranteeing task service quality. Therefore, there is an urgent need for an energy consumption-aware deployment optimization method that can jointly optimize drone swarm task matching and association, communication, reconnaissance, and computing resource allocation, and drone deployment location under concurrent communication, reconnaissance, and computing tasks. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a low-energy joint optimization method for UAV swarms performing communication, reconnaissance, and computing tasks. This method integrates flight and hovering energy consumption, communication service energy consumption, computing processing energy consumption, and detection service energy consumption into a unified optimization objective. By constructing a low-energy joint optimization model aimed at minimizing the total energy consumption of the UAV swarm system, and through user task allocation association, communication, reconnaissance, and computing resource allocation, and spatial deployment scheduling, it achieves efficient utilization of airborne resources and coordinated optimization of low-energy networking. This enables low-energy collaborative services under multi-service concurrency conditions, improving the utilization efficiency of communication, reconnaissance, and computing resources and the adaptability of UAV swarm deployment.

[0006] This invention is achieved using the following technical solution: This invention provides a low-energy joint optimization method for UAV swarms performing communication and reconnaissance tasks, comprising the following steps: A low-energy joint optimization method for UAV swarms performing reconnaissance tasks is characterized by the following steps: Step S1: Based on the UAV swarm status information, user communication and reconnaissance task requirements information, and system resource constraint information, construct a low-energy joint optimization model with the goal of minimizing the total energy consumption of the UAV swarm system. Step S2: The low-energy joint optimization model is divided into three sub-problems: communication and reconnaissance task matching and correlation optimization, communication and reconnaissance resource allocation optimization, and UAV deployment location optimization. The sub-problems are solved alternately in the order of communication and reconnaissance task matching and correlation, communication and reconnaissance resource allocation, and UAV deployment location. Step S3: Given the deployment location of the UAV, solve the sub-problem of matching and association optimization of the communication and reconnaissance task to obtain the matching and association results between the UAV and the user's communication and reconnaissance task. Step S4: Given the deployment location of the UAV and the matching association result, solve the sub-problem of resource allocation optimization for communication and reconnaissance computing to obtain the resource allocation result of communication and reconnaissance computing. Step S5: Given the matching association result and the resource allocation result, solve the UAV deployment location optimization sub-problem to obtain the updated UAV deployment location result; Step S6: Based on the updated UAV deployment location results, iteratively execute steps S3 to S5 until the preset convergence condition is met or the preset maximum number of iterations is reached, and output the joint optimization results of the final task matching association, communication resource allocation and UAV deployment location of the UAV swarm communication calculation.

[0007] According to a further description of the present invention, preferably, step S1 specifically includes the following steps: The drone swarm status information includes the current location, service capacity, remaining energy, maximum number of services, and available resource budget for each drone; among which, to distinguish the service functions of different drones, the service capacity vector of drone m is obtained: ; in, , , Both are binary variables, representing the drone respectively. Does it possess communication service capabilities, detection service capabilities, and computing service capabilities? When the corresponding variable takes a value of 1, it indicates that the drone... It possesses the corresponding service capabilities; when the corresponding variable takes a value of 0, it indicates that the drone... Lacking the corresponding service capabilities; based on the aforementioned UAV service capability vector, the UAV set is divided into a communication UAV set. Detection drone collection and computing drone collection ; Represents communication tasks. Represents a reconnaissance mission. Represents a computational task; The communication, detection, and computation tasks are decomposed into specific user sub-tasks, including each user's location in the current time slot, service request type, and corresponding task parameters; to describe whether a user requests a communication task, detection task, or computation task, the service request vector of user n in time slot t is obtained: ; in, , , Both are binary variables, representing users respectively. In the time slot Whether to request communication, detection, and computation tasks; when the corresponding variable is 1, it indicates that the user... Request the corresponding task; when the corresponding variable is 0, it indicates that the user... No corresponding task is requested; assuming the drone ensemble is... The user set is The time slot set is ; The user's communication and detection task requirements include communication task requirements, detection task requirements, and calculation task requirements. The communication task requirement information is represented by a communication task tuple as follows: ; in, Indicates user In the time slot The amount of communication task data to be received. Indicates user In the time slot The maximum allowed completion time for a communication task. Indicates user In the time slot The minimum access signal-to-interference-plus-noise ratio (SINR) threshold required to complete the communication task request; The detection task requirement information is represented by a detection task tuple as follows: ; in, Indicates user In the time slot The corresponding detection target area, Indicates user In the time slot The amount of baseline information requested in the detection task required to complete the detection task. Indicates user In the time slot The maximum allowed completion time for a detection mission; The computation task requirement information is represented by a computation task tuple as follows: ; in, Indicates user In the time slot The amount of data to be processed in the computational tasks. Indicates user In the time slot The computational intensity required per unit bit of data Indicates user In the time slot The maximum allowable completion time for the calculation task; Based on UAV swarm status information, user communication and surveillance task requirements, and system resource constraints, a low-energy joint optimization model is constructed with the goal of minimizing the total energy consumption of the UAV swarm system; among which, UAVs In the time slot The total energy consumption of the system is expressed as: ; in, Represents flight status; Represents the communication status; Indicates drone In the time slot The energy consumption during flight and hovering. Indicates drone In the time slot The energy consumption of performing communication tasks; Indicates drone In the time slot The energy consumption of performing computational tasks. Indicates drone In the time slot The energy consumption for performing detection tasks; The low-energy joint optimization model is expressed as follows:

[0008] Among them, the optimization variables are ; This represents the set of drone deployment locations. Indicates in time slot The set of matching related variables, Indicates in time slot The set of downlink bandwidth allocation variables, Indicates in time slot The set of allocation variables for downlink transmit power. Indicates in time slot The set of uplink bandwidth allocation variables, Indicates in time slot The set of CPU frequency allocation variables for computational tasks. Indicates in time slot The set of CPU frequency allocation variables for detection tasks; Represents the downlink; Represents the uplink; The system resource constraint information includes constraints applicable to the low-energy joint optimization model. These constraints include unique task association constraints, maximum number of UAV services constraints, resource budget constraints, safe distance constraints between UAVs, task area boundary constraints, and adjacent time slot travel distance constraints.

[0009] According to a further description of the present invention, preferably, step S2 specifically includes: The low-energy joint optimization model is divided into matching and correlation optimization sub-problems, communication and reconnaissance computing resource allocation optimization sub-problems, and UAV deployment optimization sub-problems according to the type of optimization variables. At a given drone deployment location Under the condition of matching associated variables To construct the matching association optimization subproblem for the optimization variables. :

[0010] in Let P1 represent the total energy consumption of UAV m in time slot t. Subproblem P1 is to obtain the minimum total energy consumption of M UAVs in all time slots. The constraints of the communication, reconnaissance, and computation task matching and association optimization subproblem include unique association constraints for communication tasks, unique association constraints for computation tasks, unique association constraints for reconnaissance tasks, and constraints on the maximum number of times the UAV can serve. At a given drone deployment location and matching associated variables Under the given conditions, the communication and surveillance computing resource allocation optimization sub-problem is constructed using the communication and surveillance computing resource allocation variable as the optimization variable. :

[0011] The constraints of the communication and reconnaissance computing resource allocation optimization sub-problem include downlink bandwidth budget constraints, downlink transmit power budget constraints, uplink bandwidth budget constraints, airborne CPU frequency budget constraints, and communication and reconnaissance computing task completion constraints. Given matching associated variables Under the conditions of resource allocation results and reconnaissance, the deployment location of the UAV is considered. To optimize the variables, construct the UAV deployment optimization sub-problem. :

[0012] The constraints of the UAV deployment optimization sub-problem include the safe distance constraints between UAVs, the boundary constraints of the mission area, the travel distance constraints between adjacent time slots, and the completion constraints of the communication and reconnaissance task. The matching and correlation optimization sub-problem, the communication and reconnaissance computing resource allocation optimization sub-problem, and the UAV deployment optimization sub-problem are solved alternately in that order.

[0013] According to a further description of the present invention, preferably, step S3 specifically includes: Based on the current drone deployment location and user location, calculate the drone... With users Spatial distance between And construct matching edge weights using the negative values ​​of the spatial distance: ; in, Indicates the task type. This represents the weight values ​​of the different edges formed by different users and drones during the matching process; Based on the maximum number of services per drone Expanding the drone nodes to generate a virtual drone node set transforms the capacity-constrained one-to-many matching problem into a one-to-one maximum weight matching problem. The objective function of the one-to-one maximum weight matching problem is:

[0014] in, For task type Unloading drones With users In the time slot Related variables; Indicates drone Matching users Weight value at time For executable tasks A collection of drones, For having a task The set of users who require it; An auction matching algorithm is used to solve the one-to-one maximum weight matching problem. In the auction matching algorithm, unmatched user tasks select bidding nodes based on their net income from each virtual drone node. Virtual drone nodes update node prices and establish or replace matching relationships based on bidding until the user task matching association is completed or a preset termination condition is met. The matching association result is constructed based on the final matching relationship.

[0015] According to a further description of the present invention, preferably, step S4 specifically includes: The user {n} is a large set, which includes user I for communication tasks, user j for detection tasks, and user k for computation tasks. For communication tasks, given the known deployment locations of drones and the matching association results, determine the drones... Assigned to communication users in communication tasks downlink bandwidth and downlink transmit power The downlink transmission rate is calculated according to the following formula: ; in, Indicates drone With users The equivalent channel power gain between Indicates the noise power spectral density; Under the conditions of satisfying the minimum access signal-to-interference-plus-noise ratio threshold and the maximum completion time of the communication task, determine the minimum feasible downlink bandwidth and minimum feasible downlink transmit power of the communication task: ; ; For detection missions, determine the drone Detection users assigned to detection tasks CPU frequency detection And determine the minimum feasible detection CPU frequency based on the detection and processing latency constraints: ; in, Indicates the strength of the detection calculation. Indicates drone When performing a detection task, detect the user The amount of detection data generated; For computational tasks, determine the drone. Assigned to users in computing tasks Uplink bandwidth And calculate CPU frequency The uplink unloading rate is calculated using the following formula: ; in, Indicates computing user Uplink transmit power; Given uplink bandwidth Under the given conditions, determine the minimum feasible CPU frequency based on the total computational task latency constraint: ; The computational resource allocation problem is transformed into a single-variable optimization problem concerning uplink bandwidth. The uplink bandwidth allocation result is determined by Lagrange dual update and one-dimensional search, and the computational CPU frequency allocation result is recovered based on the minimum feasible computational CPU frequency. The minimum feasible downlink bandwidth, minimum feasible downlink transmit power, minimum feasible detection CPU frequency, uplink bandwidth allocation result, and calculated CPU frequency allocation result are used as the communication detection computing resource allocation result.

[0016] According to a further description of the present invention, preferably, step S5 specifically includes: The particle swarm optimization algorithm is used to solve the subproblem of optimizing the deployment location of the UAV; where each particle represents a three-dimensional deployment scheme of the UAV, and the i-th particle represents the t-th particle. The position vector of each particle is represented as: ; in, Indicates the first The first particle The three-dimensional deployment location of the drone; Using the drone deployment location of the previous time slot as the center, the particle position is initialized with random perturbation. Construct a particle fitness function based on total system energy consumption and business constraint default terms: ; in, Indicates the first Sub-iteration particles Total system energy consumption under the corresponding deployment scheme , , These represent the violation terms of the communication task constraints, the computation task constraints, and the detection task constraints, respectively. This is the penalty coefficient; Update the individual and group historical best positions based on the fitness values ​​of each particle, and update the particle velocity and position according to the following formula: ; ; in, Indicates the first In the nth iteration The speed of each particle Indicates the first The individual best position in the history of each particle. This indicates the group's historical best position. Indicates inertia weight, It is an individual learning factor. It is a group learning factor. It is the first random number. It is the second random number; When the particle swarm iteration reaches the preset number of deployment optimization iterations, or when the fitness change meets the preset deployment convergence condition, the UAV 3D deployment scheme corresponding to the swarm's historical best position is used as the updated UAV deployment position.

[0017] According to a further description of the present invention, preferably, step S6 specifically includes: Initialize the drone deployment location and calculate the initial total system energy consumption; In the In the alternating iterations, the matching association optimization sub-problem is solved based on the current drone deployment location to obtain the... This match is associated with the results; Based on the current drone deployment location and the Solving the communication and detection resource allocation optimization sub-problem using the matching and association results of the first step yields the first... The results of resource allocation for secondary surveillance; Based on the The result of the second matching association and the first The results of the second reconnaissance resource allocation are used to solve the UAV deployment optimization sub-problem, resulting in the first... The updated drone deployment locations; According to the The updated drone deployment location results, the The result of the second matching association and the first The calculation of the next detection resource allocation result is as follows: The total energy consumption of the subsystem is calculated, and the change in total system energy consumption between two adjacent alternating iterations is also calculated: ; when Less than or equal to the preset convergence threshold, or When the preset maximum number of iterations is reached, the alternating iteration stops, and the final task matching and association results, communication resource allocation results, detection resource allocation results, computing resource allocation results, and UAV deployment location results are output.

[0018] Based on the same inventive concept, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above method.

[0019] Based on the same inventive concept, 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, wherein the processor executes the computer program to implement the steps of the above method.

[0020] Compared with the prior art, the beneficial effects of the present invention are: 1) This invention constructs a low-energy joint optimization model with the goal of minimizing the total energy consumption of the UAV swarm system. It integrates flight and hovering energy consumption, communication service energy consumption, computing processing energy consumption and detection service energy consumption into the optimization objective, which can realize low-energy collaborative services under multi-service concurrent conditions.

[0021] 2) This invention decomposes the original joint optimization problem into a communication and reconnaissance task matching and correlation optimization subproblem, a communication and reconnaissance resource allocation optimization subproblem, and a UAV deployment location optimization subproblem, thereby reducing the solution complexity of simultaneously processing discrete matching variables, continuous resource allocation variables, and three-dimensional deployment location variables.

[0022] 3) This invention constructs matching edge weights based on spatial distance and solves the one-to-many matching problem with capacity constraints through node expansion and auction matching algorithms, so that the UAV and user communication and reconnaissance tasks form a matching relationship that is more in line with spatial proximity and service capacity constraints.

[0023] 4) This invention allocates resources for communication tasks, detection tasks and computing tasks respectively. While meeting task constraints such as communication latency, access signal-to-interference-plus-noise ratio, detection processing latency and total computing latency, it reduces unnecessary bandwidth, transmission power and CPU frequency consumption and improves the utilization efficiency of communication, detection and computing resources.

[0024] 5) This invention uses particle swarm optimization algorithm to update the three-dimensional deployment position of UAVs, and continuously coordinates the UAV service tasks, resource status and spatial location through alternating iterations of matching association, resource allocation and deployment optimization, so that the UAV swarm can adapt to changes in user location and task requirements and improve deployment adaptability in continuous support scenarios. Attached Figure Description

[0025] Figure 1 This is an overall flowchart of the low-energy joint optimization method for UAV swarms performing communication and reconnaissance tasks according to the present invention; Figure 2 This is a schematic diagram of a drone swarm communication and reconnaissance service scenario for the low-energy joint optimization method of drone swarms for performing communication and reconnaissance tasks according to the present invention. Figure 3 This is a flowchart of the matching and correlation optimization subproblem of the low-energy joint optimization method for UAV swarms performing reconnaissance tasks according to the present invention. Figure 4This is a schematic diagram of the decomposition of the communication and reconnaissance computing resource allocation optimization subproblem in the low-energy joint optimization method for UAV swarms performing communication and reconnaissance computing tasks according to the present invention. Figure 5 A comparison chart showing the changes in total system energy consumption of the low-energy joint optimization method for UAV swarms performing reconnaissance tasks according to the present invention; Figure 6 This is a comparison chart of the average energy consumption of a single UAV among the low-energy joint optimization method for UAV swarms performing reconnaissance tasks according to the present invention and different ablation schemes. Detailed Implementation

[0026] The invention will now be described in further detail with reference to the accompanying drawings.

[0027] Example 1: like Figure 1 As shown in the overall flowchart, this embodiment provides a low-energy joint optimization method for UAV swarms performing communication and reconnaissance tasks. Through user task allocation association, communication and reconnaissance resource allocation, and spatial deployment scheduling, it achieves efficient utilization of airborne resources and collaborative optimization of low-energy networking. Compared with other methods, this method has superior energy efficiency. Specifically, it includes the following steps: Step S1: Based on the UAV swarm status information, user communication and reconnaissance task requirements information, and system resource constraint information, construct a low-energy joint optimization model; like Figure 2 As shown in the schematic diagram of the UAV swarm communication, reconnaissance, and computing service scenario, within the task area, ground users may simultaneously have communication, reconnaissance, and computing task requirements. However, the service capabilities, available resources, and deployment locations of different UAVs within the swarm are not entirely the same. To achieve task matching, resource allocation, and deployment location optimization between UAVs and user communication, reconnaissance, and computing tasks in subsequent steps, this embodiment first obtains the UAV swarm status information, user communication, reconnaissance, and computing task requirement information, and system resource constraint information for the current time slot.

[0028] Obtain the drone swarm status information, including the current location, service capacity, remaining energy, maximum number of services, and available resource budget for each drone. Specifically, to differentiate the service functions of different drones, obtain the service capacity vector of drone m: ; in, , , Both are binary variables, representing the drone respectively. Does it possess communication service capabilities, detection service capabilities, and computing service capabilities? When the corresponding variable takes a value of 1, it indicates that the drone... It possesses the corresponding service capabilities; when the corresponding variable takes a value of 0, it indicates that the drone... It lacks the corresponding service capabilities. Based on the aforementioned UAV service capability vector, the UAV set is divided into a communication UAV set. Detection drone collection and computing drone collection . Represents communication tasks. Represents a reconnaissance mission. Represents a computational task.

[0029] Obtain user communication, detection, and computation task request information, including each user's location in the current time slot, service request type, and corresponding task parameters. To describe whether a user has requested a communication task, detection task, or computation task, obtain the service request vector for user n in time slot t: ; in, , , Both are binary variables, representing users respectively. In the time slot Whether to request communication, detection, and computation tasks. When the corresponding variable is 1, it indicates that the user... Request the corresponding task; when the corresponding variable is 0, it indicates that the user... Do not request the corresponding task.

[0030] Let the set of drones be The user set is The time slot set is ; The user's communication and detection task requirements include communication task requirements, detection task requirements, and calculation task requirements. Based on the user service demand vector, the user set is divided into a communication user set. Detect user sets and calculate user set .

[0031] For communication tasks, obtain the amount of data the communication user needs to receive in the current time slot, the maximum task completion time, and the minimum access signal-to-interference-plus-noise ratio (SINR) threshold. The communication task requirement information is represented by a communication task tuple as follows: ; in, This represents the amount of communication data that user n needs to receive in time slot t. This indicates the maximum allowed completion time for the communication task. Indicates user In the time slot The minimum access signal-to-interference-plus-noise ratio (SINR) threshold required to initiate a communication task request.

[0032] For a detection task, the system obtains the user's location, the minimum amount of detection information required to complete the task, and the maximum time limit for completion. The detection task requirements are represented by a detection task tuple: ; in, Indicates user In the time slot The corresponding detection target area, Indicates user In the time slot The amount of baseline information requested in the detection task required to complete the detection task. Indicates user In the time slot The maximum allowed completion time for the detection mission.

[0033] For computational tasks, obtain the amount of data to be processed by the user, the computational intensity required per unit bit of data, and the maximum completion time for the computational task. The computational task requirements are represented by a computational task tuple as follows: ; in, Indicates user In the time slot The amount of data to be processed in the computational tasks. Indicates user In the time slot The computational intensity required per unit bit of data Indicates user In the time slot The maximum allowable completion time for the computation task.

[0034] Therefore, the UAV swarm status information, user communication and reconnaissance task requirement information, and system resource constraint information obtained in step S1 together constitute the input data for the subsequent low-energy joint optimization model, providing a data foundation for the joint optimization of matching correlation variables, communication and reconnaissance resource allocation variables, and UAV deployment locations.

[0035] Based on the UAV swarm status information, user communication and reconnaissance task requirements, and system resource constraints, a low-energy joint optimization model is constructed with the goal of minimizing the total energy consumption of the UAV swarm system. The low-energy joint optimization model uses communication and reconnaissance task matching variables, communication and reconnaissance resource allocation variables, and UAV deployment locations as optimization variables.

[0036] In this embodiment, the drone In the time slot Total energy consumption includes flight and hovering energy consumption, communication service energy consumption, computing processing energy consumption, and detection service energy consumption for drones. In the time slot The total energy consumption of the system is expressed as: ; in, Represents flight status. Represents the communication status. Indicates drone In the time slot The total energy consumption of the system; Indicates drone In the time slot Energy consumption during flight and hovering; Indicates drone In the time slot Energy consumption for communication services generated during the execution of communication tasks; Indicates drone In the time slot The energy consumed by computational processing in performing computational tasks; Indicates drone In the time slot Energy consumption of detection services generated during the execution of detection tasks.

[0037] The flight and hovering energy consumption can be determined based on the UAV's flight power, hovering power, flight time, and hovering time, and is expressed as follows: ; in, Indicates drone Flight speed The flight power below, Indicates drone In the time slot Flight time within the range, Indicates the hovering power of the drone. Indicates drone In the time slot Hovering time within.

[0038] The energy consumption for the communication service can be determined based on the downlink transmit power allocated by the UAV for the communication task and the transmission time of the communication task, expressed as:

[0039] in, Represents a set of communication users. Indicates drone With communication users In the time slot Communication-related variables, Indicates drone Assigned to communication users downlink transmit power, Indicates drone To communication users Transmission time for transmitting communication task data.

[0040] The computational processing energy consumption can be determined based on the CPU frequency and computation time allocated to the computational task by the UAV, expressed as follows:

[0041] in, This indicates the computation of the user set. Indicates drone With computing users In the time slot Calculate the related variables. Indicates drone The effective switching capacitor coefficient of the chip, Indicates drone Processing computing users The computational processing time required for the computational task Indicates drone Assigned to computing users Calculate CPU frequency.

[0042] The energy consumption of the detection service includes the energy consumption for detection data acquisition and the energy consumption for detection data processing, expressed as follows: ; in, Indicates drone In the time slot Energy consumption for detection data acquisition Indicates drone In the time slot The energy consumption for processing detection data. Furthermore, the energy consumption for acquiring detection data can be expressed as:

[0043] The energy consumption for detecting data processing can be expressed as:

[0044] in, Indicates the set of users to be detected. Indicates drone With the detection user In the time slot Detection of related variables, Indicates the energy consumption per detection and data collection. Indicates drone Complete the detection mission Number of detections required Indicates drone Handling detection tasks The detection data processing time Indicates drone Assigned to detection tasks Detects CPU frequency.

[0045] Based on the above energy consumption composition, the following low-energy joint optimization model is constructed. :

[0046] Among them, the optimization variables are ; This represents the set of drone deployment locations. Indicates in time slot The set of matching related variables, Indicates in time slot The set of downlink bandwidth allocation variables, Indicates in time slot The set of allocation variables for downlink transmit power. Indicates in time slot The set of uplink bandwidth allocation variables, Indicates in time slot The set of CPU frequency allocation variables for computational tasks. Indicates in time slot The set of CPU frequency allocation variables for detection tasks; Represents the downlink; Represents the uplink; The system resource constraint information includes constraints applicable to the low-energy joint optimization model. These constraints include unique task association constraints, maximum number of UAV services constraints, resource budget constraints, safe distance constraints between UAVs, task area boundary constraints, and adjacent time slot travel distance constraints.

[0047] In the above optimization model In this context, constraints C1 to C3 are task uniqueness constraints, ensuring that each communication, detection, and computation task is provided by at most one UAV with the corresponding service capability within the same time slot, preventing the same task from being repeatedly assigned. Constraint C4 is a UAV service capacity constraint, limiting the total number of communication, detection, and computation services performed by each UAV within the same time slot, ensuring that the sum of the number of communication, detection, and computation services performed by each UAV does not exceed the maximum number of services a UAV can provide. Constraints C5 and C6 are communication resource budget constraints. Constraint C5 limits the total downlink bandwidth allocated to communication tasks by the UAV to no more than its maximum downlink bandwidth budget. Constraint C6 is used to limit the total downlink transmit power allocated by the UAV to communication tasks to not exceed its maximum downlink transmit power budget. Constraint C7 is an uplink offload resource budget constraint, used to limit the total uplink bandwidth allocated by the UAV to computing tasks to not exceed its maximum uplink bandwidth budget. Constraint C8 is an onboard computing resource budget constraint, used to limit the sum of CPU frequencies allocated to computing and detection tasks by the UAV to no more than its maximum allocatable onboard CPU frequency. Constraint C9 is a safety distance constraint between drones, used to ensure that the spatial distance between any two drones is not less than a preset safety distance. To reduce the risk of collisions between drones, constraint C10 is a task area boundary constraint used to ensure the deployment position of the drones. Always located in the preset task area Within. Constraint C11 is an adjacent time slot movement distance constraint, used to limit the movement distance of the UAV between adjacent time slots to no more than a preset maximum movement distance. To meet the constraints of drone mobility.

[0048] Step S2: Decompose the low-energy joint optimization model into three optimization sub-problems and solve them alternately; Since the low-energy-consumption joint optimization model simultaneously includes discrete communication and reconnaissance task matching and correlation variables, continuous communication and reconnaissance resource allocation variables, and continuous UAV 3D deployment location variables, and the total system energy consumption is highly coupled with the task correlation, resource allocation results, and UAV deployment location, directly solving this model is highly complex. Therefore, this embodiment decomposes the low-energy-consumption joint optimization model into communication and reconnaissance task matching and correlation optimization sub-problems, communication and reconnaissance resource allocation optimization sub-problems, and UAV deployment location optimization sub-problems, and solves the decomposed sub-problems alternately in the order of communication and reconnaissance task matching and correlation, communication and reconnaissance resource allocation, and UAV deployment location. In the In each alternation iteration, based on the current drone deployment location First, a matching association optimization subproblem is constructed. This matching association optimization subproblem focuses on matching association variables. To optimize variables, the service correspondence between the UAV and user communication tasks, detection tasks, and computing tasks is determined.

[0049] At a given drone deployment location Under the condition of matching associated variables To construct the matching association optimization subproblem for the optimization variables. Its expression is:

[0050] Among them, constraints C1 to C4 are all constraints defined by the low-energy joint optimization model P0 in step S1. Let P1 represent the total energy consumption of UAV m in time slot t. Subproblem P1 is to obtain the minimum total energy consumption of M UAVs in all time slots. The constraints of the communication and reconnaissance task matching and association optimization subproblem include unique association constraints for communication tasks, unique association constraints for computation tasks, unique association constraints for reconnaissance tasks, and constraints on the maximum number of times the UAV can serve; wherein, the constraints of the communication and reconnaissance resource allocation optimization subproblem include downlink bandwidth budget constraints, downlink transmit power budget constraints, uplink bandwidth budget constraints, airborne CPU frequency budget constraints, and communication and reconnaissance task completion constraints. After obtaining the number Secondary matching association results Then, based on the current deployment location of the drone and matching results A sub-problem for optimizing the allocation of communication, detection, and computing resources is constructed. This sub-problem uses communication resources, detection resources, and computing resources as optimization variables to determine the downlink bandwidth, downlink transmit power, uplink bandwidth, computing CPU frequency, and detection CPU frequency allocated to each UAV for associated tasks; given the UAV deployment location... and matching associated variables Under the given conditions, the communication and surveillance computing resource allocation optimization sub-problem is constructed using the communication and surveillance computing resource allocation variable as the optimization variable. Its expression is:

[0051] Among them, constraints C5 to C8 are all constraints defined by the low-energy joint optimization model P0 in step 1. Meanwhile, to ensure that various communication tasks can meet their corresponding quality of service requirements, the communication resource allocation optimization sub-problem also satisfies task completion constraints. Specifically, for communication tasks, the minimum access signal-to-interference-plus-noise ratio threshold requirement must be met simultaneously. Maximum completion time requirement for communication tasks For computational tasks, the total completion time requirement, consisting of uplink offloading latency and edge computing processing latency, must be met. For detection tasks, the time limit for processing detection data must be met. ;in, Indicates drone With users The equivalent channel power gain between Represents the noise power spectral density. Indicates drone To communication users downlink transmission rate Indicates computing user To drones Uplink unloading rate of unloading computing tasks. Indicates the strength of the detection calculation. Indicates drone Execute reconnaissance mission The amount of detection data generated.

[0052] After obtaining the number After the resource allocation results of the second communication and surveillance calculation, based on the first The result of the second matching association and the first Based on the resource allocation results of the secondary reconnaissance calculation, a sub-problem of UAV deployment optimization is constructed. This UAV deployment optimization sub-problem is based on the UAV deployment location. To optimize variables used to update the 3D deployment positions of each UAV within the mission area; given matching and correlated variables Under the conditions of resource allocation results and reconnaissance, the deployment location of the UAV is considered. To optimize the variables, construct the UAV deployment optimization sub-problem. Its expression is:

[0053] Among them, constraints C9 to C11 are all constraints defined by the low-energy joint optimization model P0 in step 1. The constraints of the UAV deployment optimization sub-problem include the safety distance constraints between UAVs, the task area boundary constraints, the movement distance constraints between adjacent time slots, and the communication and reconnaissance task completion constraints; they are solved alternately in the order of the matching association optimization sub-problem, the communication and reconnaissance resource allocation optimization sub-problem, and the UAV deployment optimization sub-problem.

[0054] Therefore, in each round of alternating iterations, this embodiment first optimizes the matching and correlation variables given the UAV deployment location, then optimizes the communication and surveillance resource allocation variables given the UAV deployment location and the matching and correlation results, and finally optimizes the UAV deployment location given the matching and correlation results and the communication and surveillance resource allocation results. After completing one round of alternating iterations, the matching and correlation optimization subproblem, the communication and surveillance resource allocation optimization subproblem, and the UAV deployment optimization subproblem are solved based on the updated UAV deployment location.

[0055] Specifically, let the first The total energy consumption of the system obtained by the alternating iterations is: , No. The total energy consumption of the system obtained by the alternating iterations is: Then, the change in total system energy consumption between two consecutive alternating iterations is: ; when If the number of alternating iterations is less than or equal to the preset convergence threshold, or if the number of alternating iterations reaches the preset maximum number of iterations, the alternating iterations are stopped; otherwise, the current updated drone deployment location is used as the input for the next round of alternating iterations, and the solution is continued in the order of matching association optimization, communication and surveillance resource allocation optimization, and drone deployment optimization.

[0056] By using the above-mentioned splitting and alternating solution method, this embodiment transforms the original joint optimization problem into three interrelated but relatively well-defined sub-problems, thereby reducing the difficulty of simultaneously handling discrete matching variables, continuous resource allocation variables, and UAV deployment location variables.

[0057] Step S3: Given the deployment location of the UAV, solve the sub-problem of matching and association optimization of the communication and reconnaissance task to obtain the matching and association results between the UAV and the user's communication and reconnaissance task. Based on the matching association optimization sub-problem constructed in step S2, closer UAVs typically have better air-to-ground link conditions, which helps reduce communication transmission energy consumption and computational offloading overhead. For detection tasks, closer UAVs can obtain better observation geometry, which helps reduce the cost of completing the detection task. Therefore, this embodiment constructs the matching edge weights between UAVs and user tasks using negative values ​​of spatial distance: Based on the current drone deployment location and user location, calculate the drone... With users Spatial distance between And construct matching edge weights using the negative values ​​of the spatial distance: ; in, Indicates the user task type. This represents the weight values ​​of the different edges formed by different users and drones during the matching process; Indicates time slot drones With users Spatial distance; Indicates drone Service users of Matching edge weights for different task types. A larger edge weight indicates a more powerful drone. With users The smaller the spatial distance between them, the more suitable the drone is for performing the corresponding user's task.

[0058] Based on the maximum number of services per drone Expanding the drone nodes to generate a virtual drone node set transforms the capacity-constrained one-to-many matching problem into a one-to-one maximum weight matching problem. The objective function of the one-to-one maximum weight matching problem is:

[0059] in, For task type Unloading drones With users In the time slot Related variables; Indicates drone Matching users Weight value at time For executable tasks A collection of drones, For having a task The set of users who require it; An auction matching algorithm is used to solve the one-to-one maximum weight matching problem. In the auction matching algorithm, unmatched user tasks select bidding nodes based on their net income from each virtual drone node. Virtual drone nodes update node prices and establish or replace matching relationships based on bidding until the user task matching association is completed or a preset termination condition is met. The matching association result is constructed based on the final matching relationship.

[0060] Since each drone can serve multiple user tasks simultaneously, and each user task is associated with at most one drone, this matching problem is a one-to-many matching problem with capacity constraints. For ease of solution, this embodiment uses the maximum number of services each drone can provide. Expand the drone nodes. For example... Figure 3 As shown, for drones structure One virtual drone node: ; Each virtual drone node corresponds to a drone. One available service slot, and inherit the drone The matching edge weights for each user task. By extending the processing through this node, the original one-to-many matching problem with capacity constraints is transformed into a one-to-one maximum weight matching problem.

[0061] Subsequently, the auction matching algorithm is used to solve the one-to-one maximum weight matching problem. The specific process includes the following steps: Initialize the price, matching set, and unmatched task set for each virtual drone node. For any unmatched user task... Calculate its relationship with each virtual drone node Net income: ; in, Represents virtual drone nodes The current node price in the auction process.

[0062] For unmatched user tasks Select the virtual drone node with the highest net profit as the optimal bidding node: ; And select the virtual drone node with the highest net profit besides the optimal bidding node as the second-best bidding node: ; Calculate the bidding increment based on the difference between the best and second-best net profits: ; in, This sets the auction step size parameter. Subsequently, the price of the optimal virtual drone node is updated: ; If the optimal virtual drone node is already occupied by another user task, the original matching task is released and added back to the unmatched task set; if the optimal virtual drone node is not occupied, a matching relationship is directly established between the current user task and the optimal virtual drone node.

[0063] Repeat the above net profit calculation, bidding, price update, and matching replacement process until the unmatched task set is empty, or the change in current auction profit is less than a preset termination threshold. Finally, map the matching results of the virtual drone nodes back to the original drone nodes to construct the matching association set for the current time slot. .

[0064] Step S4: Given the deployment location of the UAV and the matching association result, solve the sub-problem of resource allocation optimization for communication and reconnaissance computing to obtain the resource allocation result of communication and reconnaissance computing. like Figure 4 As shown, since communication tasks, detection tasks and computing tasks occupy different types of resources, this embodiment further decomposes the sub-problem of optimizing the allocation of communication, detection and computing resources into communication resource allocation, detection resource allocation and computing resource allocation.

[0065] Let users {n} be a large set, which includes users of communication tasks. I Users of detection tasks j and users of computing tasks k ; For communication tasks, given the known deployment locations of drones and the matching association results, determine the drones... Assigned to communication users in communication tasks downlink bandwidth and downlink transmit power For drones Associated communication users The drone To communication users The downlink transmission rate is: ; in, Indicates drone With users The equivalent channel power gain between Indicates the noise power spectral density; The allocation of communication task resources needs to meet the minimum access signal-to-interference-plus-noise ratio (SINR) threshold and the maximum completion time requirement of the communication task. Since the energy consumption of communication services is related to downlink transmit power and communication transmission time, this embodiment selects the minimum feasible downlink bandwidth and minimum feasible downlink transmit power to enable the communication task to be completed, while meeting the quality of service requirements. Under the conditions of meeting the minimum access SINR threshold and the maximum completion time requirement of the communication task, the minimum feasible downlink bandwidth and minimum feasible downlink transmit power of the communication task are determined as follows: ; ; This yields the downlink bandwidth allocation and downlink transmit power allocation results for each communication task.

[0066] For drones Related detection tasks Since the energy consumption for detection processing increases with the detection CPU frequency, the minimum feasible detection CPU frequency is selected as the detection resource allocation result, while satisfying the detection latency constraint. For the detection task, the UAV is determined... Detection users assigned to detection tasks CPU frequency detection And determine the minimum feasible detection CPU frequency based on the detection and processing latency constraints: ; in, Indicates the strength of the detection calculation. Indicates drone When performing a detection task, detect the user The amount of detection data generated; This leads to the development of drones. The set of detection CPU frequencies allocated to the detection tasks it serves. Then, the drone is calculated. Remaining airborne CPU frequency budget after detection resource allocation:

[0067] in, Indicates drone The maximum allocatable CPU frequency. The remaining onboard CPU frequency budget is used for resource allocation for subsequent computing tasks.

[0068] For computational tasks, determine the drone. Assigned to users in computing tasks Uplink bandwidth And calculate CPU frequency For drones Associated computing users The user To drones The uplink transfer rate for unloading computing tasks is: ; in, Indicates computing user Uplink transmit power; To ensure that the computing task can be completed within the maximum time limit Completed within a given uplink bandwidth Under the given conditions, determine the minimum feasible CPU frequency based on the total computational task latency constraint: ; The computational resource allocation problem is transformed into a single-variable optimization problem concerning uplink bandwidth. The uplink bandwidth allocation result is determined by Lagrange dual update and one-dimensional search, and the computational CPU frequency allocation result is recovered based on the minimum feasible computational CPU frequency. The minimum feasible downlink bandwidth, minimum feasible downlink transmit power, minimum feasible detection CPU frequency, uplink bandwidth allocation result, and calculated CPU frequency allocation result are used as the communication detection computing resource allocation result.

[0069] Since computational processing power consumption increases with CPU frequency, the minimum feasible CPU frequency is selected as the computational frequency allocation value given the uplink offload bandwidth. This embodiment constructs a Lagrangian function to address the computational resource allocation problem and introduces the dual variable corresponding to the uplink bandwidth constraint. The dual variable corresponding to the remaining CPU frequency constraint Given dual variables, construct a relation for each computational task. The one-dimensional stationary point equation is obtained, and the corresponding uplink unloading bandwidth is solved using a binary search method; then, the dual variables are updated according to the constraint residuals: ; ; in, Indicates the dual update step size. , This indicates the current dual iteration number. When the change in the dual variable is less than the preset resource allocation convergence threshold, or when the preset maximum resource allocation iteration number is reached, the iteration stops, and the optimal uplink offloading bandwidth is output. Then, substituting the optimal uplink offload bandwidth into the minimum feasible CPU frequency calculation expression, we obtain the calculated CPU frequency allocation result. .

[0070] After the above-mentioned allocation of communication resources, detection resources, and computational resources, the allocation results of communication, detection, and computational resources for the current time slot are obtained: ; in, This represents the set of downlink bandwidth allocation results for a communication task. This represents the set of downlink transmit power allocation results for a communication task. This represents the set of uplink offload bandwidth allocation results for computational tasks. This represents the set of CPU frequency allocation results for computational tasks. This represents the set of CPU frequency allocation results for the detection task. The CPU resource allocation results are used as input to the UAV deployment optimization sub-problem in step S5.

[0071] Step S5: Solve the drone deployment optimization subproblem to obtain the updated drone deployment locations; Given the matching and association results and the resource allocation results, the sub-problem of optimizing the UAV deployment location is solved to obtain the updated UAV deployment location results.

[0072] Based on the UAV deployment optimization sub-problem constructed in step S2, this embodiment uses the particle swarm optimization algorithm to update the three-dimensional deployment position of the UAV swarm within the mission area, given the matching association results and the communication and reconnaissance resource allocation results.

[0073] The particle swarm optimization algorithm is used to solve the subproblem of optimizing the deployment location of the UAVs; specifically, in the particle swarm optimization algorithm, each particle represents a three-dimensional deployment scheme of the UAV swarm, and the... The position vector of each particle is represented as: ; in, Indicates the first The first particle The three-dimensional deployment location of the drone; During the initialization phase, the positions of each drone are randomly perturbed, centered on the deployment location of the drone in the previous time slot, to generate an initial particle swarm.

[0074] To simultaneously consider both total system energy consumption and communication task service constraints, a particle fitness function is constructed based on the total system energy consumption and the default terms of service constraints: ; in, Indicates the first The first in the next particle swarm iteration One particle, Indicates the first Sub-iteration particles Total system energy consumption under the corresponding deployment scheme , , These represent the violation terms of the communication task constraints, the computation task constraints, and the detection task constraints, respectively. This is the penalty coefficient; In each particle swarm iteration, the fitness value of each particle is calculated, and the individual best position of each particle and the collective best position of the particle swarm are updated based on the fitness value. Subsequently, the particle velocity and particle position are updated based on the individual best position and the collective best position. ; ; in, Indicates the first In the second particle swarm iteration The velocity of each particle; Indicates the first The individual best position in the history of each particle. Indicates the group's historical best position; Indicates inertia weight; Represents individual learning factors. Represents the group learning factor; It is the first random number. It is the second random number.

[0075] When the particle swarm iteration reaches the preset number of deployment optimization iterations, or when the fitness change meets the preset deployment convergence condition, the UAV 3D deployment scheme corresponding to the swarm's historical best position is used as the updated UAV deployment position.

[0076] After updating the particle positions, boundary checks and constraint processing are performed on the updated particle positions to ensure that the UAV deployment positions meet the constraints of the mission area boundary, the safe distance between UAVs, and the movement distance between adjacent time slots. The fitness calculation, individual historical best position update, swarm historical best position update, velocity update, and position update processes are repeated until the particle swarm iteration count reaches the preset deployment optimization iteration count, or the particle swarm fitness change meets the preset deployment convergence condition.

[0077] When the particle swarm optimization process ends, the swarm's historical best position will be determined. The corresponding 3D deployment scheme for the drone swarm serves as the deployment location for the drones in the current round of updates: ; Thus, the optimized UAV deployment results are obtained under the current matching and association results and the communication and surveillance resource allocation results. The updated UAV deployment locations serve as the input for the next round of matching and association optimization in the subsequent alternating solution process.

[0078] Step S6: Output the final task matching association, communication and surveillance resource allocation, and UAV deployment location; Based on the updated UAV deployment location results, iteratively execute steps S3 to S5 until the preset convergence condition is met or the preset maximum number of iterations is reached, and output the joint optimization results of the final task matching association, communication and reconnaissance resource allocation, and UAV deployment location of the UAV swarm communication and reconnaissance calculation.

[0079] Initialize the drone deployment location and calculate the initial total system energy consumption; In the In the alternating iterations, the matching association optimization sub-problem is solved based on the current drone deployment location to obtain the... This match is associated with the results; Based on the current drone deployment location and the Solving the communication and detection resource allocation optimization sub-problem using the matching and association results of the first step yields the first... The results of resource allocation for secondary surveillance; Based on the The result of the second matching association and the first The results of the second reconnaissance resource allocation are used to solve the UAV deployment optimization sub-problem, resulting in the first... The updated drone deployment locations; According to the The updated drone deployment location results, the The result of the second matching association and the first The calculation of the next detection resource allocation result is as follows: The total energy consumption of the subsystem is calculated, and the change in total system energy consumption between two adjacent alternating iterations is also calculated: ; when Less than or equal to the preset convergence threshold, or When the preset maximum number of iterations is reached, the alternating iteration stops, and the final task matching and association results, communication resource allocation results, detection resource allocation results, computing resource allocation results, and UAV deployment location results are output.

[0080] By integrating the aforementioned key technologies, an energy consumption perception deployment optimization method for UAV swarm communication, reconnaissance, and computing collaborative services is constructed. This method jointly optimizes the matching and correlation relationships between UAVs and user communication, reconnaissance, and computing tasks, communication, reconnaissance, and computing resources, and the three-dimensional deployment location of UAVs in scenarios where communication, reconnaissance, and computing tasks coexist. This enables the UAV swarm system to achieve a coordinated balance between mission service quality, airborne resource utilization efficiency, flight safety constraints, and system energy consumption, thereby promoting the efficient application of UAV swarms in complex areas and full-cycle emergency scenarios.

[0081] Simulation parameters: To verify the effectiveness of the method of the present invention, in an optional embodiment, the task area is set to a horizontal range of [missing information]. The operating altitude range of the drone is A three-dimensional rectangular area is defined and uniformly divided into 30×30 grids. Ground users within the area can request one or more services among communication, detection, and computation. The amount of communication data a communication user needs to receive in a time slot is set to 0.2 Mbit to 1 Mbit, and the amount of task data a computation user needs to unload in a time slot is set to 1 Mbit to 1.5 Mbit. A detection task is triggered when the environmental uncertainty of a grid near a detection user exceeds a preset trigger threshold, and triggering of additional detection tasks stops when the environmental uncertainty of the corresponding grid decreases below a preset release threshold. The maximum flight distance of the UAV in each time slot is set to 30m, and the safe distance between any two UAVs is set to no less than 5m. The available total downlink bandwidth and total uplink bandwidth are both set to 20MHz, the maximum downlink transmit power of the UAV is set to 2W, the user's uplink transmit power is set to 0.5W, and the noise power spectral density is set to... W / Hz, carrier frequency set to 2GHz, onboard CPU frequency set to 10GHz, and drone sensor field of view set to 90°.

[0082] Simulation results comparison: like Figure 5As shown, within 60 time slots, when the proportion of multi-task users is 0.30 and 0.60 respectively, the total system energy consumption of the multi-service UAV solution with integrated communication, detection, and computing capabilities is approximately 3800J and 4500J, respectively, both lower than the approximately 5200J and 6200J corresponding to the single-service UAV solution with only a single service capability. This result indicates that in dynamic multi-task scenarios where communication, detection, and computing tasks coexist, this invention can reduce redundant UAV deployments and cross-service location scheduling by reusing multi-service service capabilities, thereby reducing the total system energy consumption.

[0083] like Figure 6 As shown, the average energy consumption per UAV for EAD, RA-EAD, ERA-EAD, and FRA-EAD are 345.19J, 378.27J, 407.41J, and 357.39J, respectively. Here, EAD represents the complete energy consumption awareness deployment optimization method proposed in this invention; RA-EAD represents the method that uses a random association scheme to replace the matching association optimization module; ERA-EAD represents the method that uses an average resource allocation scheme to replace the communication and reconnaissance resource allocation optimization module; and FRA-EAD represents the method that fixes the resource allocation results during the deployment optimization phase. The comparison results show that the method of this invention has the lowest average energy consumption, indicating that matching association optimization, communication and reconnaissance resource allocation optimization, and the resource coordination mechanism during the deployment phase all contribute to reducing the energy consumption of UAV swarm operation. This verifies that the method proposed in this invention can achieve better low-energy operation performance in dynamic multi-task complex scenarios.

[0084] Example 2: Based on the same inventive concept, the present invention also provides a computer-readable storage medium storing a computer program thereon, wherein the computer program, when executed by a processor, implements the steps of the method of Embodiment 1. The storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0085] Example 3: Based on the same inventive concept, 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, wherein the processor executes the computer program to implement the steps of the method of embodiment 1.

[0086] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A low-energy joint optimization method for UAV swarms performing reconnaissance and surveillance tasks, characterized in that, Includes the following steps: Step S1: Based on the UAV swarm status information, user communication and reconnaissance task requirements information, and system resource constraint information, construct a low-energy joint optimization model with the goal of minimizing the total energy consumption of the UAV swarm system. Step S2: The low-energy joint optimization model is divided into three sub-problems: communication and reconnaissance task matching and correlation optimization, communication and reconnaissance resource allocation optimization, and UAV deployment location optimization. The sub-problems are solved alternately in the order of communication and reconnaissance task matching and correlation, communication and reconnaissance resource allocation, and UAV deployment location. Step S3: Given the deployment location of the UAV, solve the sub-problem of matching and association optimization of the communication and reconnaissance task to obtain the matching and association results between the UAV and the user's communication and reconnaissance task. Step S4: Given the deployment location of the UAV and the matching association result, solve the sub-problem of resource allocation optimization for communication and reconnaissance computing to obtain the resource allocation result of communication and reconnaissance computing. Step S5: Given the matching association result and the resource allocation result, solve the UAV deployment location optimization sub-problem to obtain the updated UAV deployment location result; Step S6: Based on the updated UAV deployment location results, iteratively execute steps S3 to S5 until the preset convergence condition is met or the preset maximum number of iterations is reached, and output the joint optimization results of the final task matching association, communication resource allocation and UAV deployment location of the UAV swarm communication calculation.

2. The low-energy joint optimization method for UAV swarms performing communication and reconnaissance tasks according to claim 1, characterized in that... Step S1 specifically includes the following steps: The drone swarm status information includes the current location, service capacity, remaining energy, maximum number of services, and available resource budget for each drone; among which, to distinguish the service functions of different drones, the service capacity vector of drone m is obtained: ； in, , , Both are binary variables, representing the drone respectively. Does it possess communication service capabilities, detection service capabilities, and computing service capabilities? When the corresponding variable takes a value of 1, it indicates that the drone... It possesses the corresponding service capabilities; when the corresponding variable takes a value of 0, it indicates that the drone... Lacking the corresponding service capabilities; based on the aforementioned UAV service capability vector, the UAV set is divided into a communication UAV set. Detection drone collection and computing drone collection ; Represents communication tasks. Represents a reconnaissance mission. Represents a computational task; The communication, detection, and computation tasks are decomposed into specific user sub-tasks, including each user's location in the current time slot, service request type, and corresponding task parameters; to describe whether a user requests a communication task, detection task, or computation task, the service request vector of user n in time slot t is obtained: ； in, , , Both are binary variables, representing users respectively. In the time slot Whether to request communication, detection, and computation tasks; when the corresponding variable is 1, it indicates that the user... Request the corresponding task; when the corresponding variable is 0, it indicates that the user... No corresponding task is requested; assuming the drone ensemble is... The user set is The time slot set is ; The user's communication and detection task requirements include communication task requirements, detection task requirements, and calculation task requirements. The communication task requirement information is represented by a communication task tuple as follows: ； in, Indicates user In the time slot The amount of communication task data to be received. Indicates user In the time slot The maximum allowed completion time for a communication task. Indicates user In the time slot The minimum access signal-to-interference-plus-noise ratio (SINR) threshold required to complete the communication task request; The detection task requirement information is represented by a detection task tuple as follows: ； in, Indicates user In the time slot The corresponding detection target area, Indicates user In the time slot The amount of baseline information requested in the detection task required to complete the detection task. Indicates user In the time slot The maximum allowed completion time for a detection mission; The computation task requirement information is represented by a computation task tuple as follows: ； in, Indicates user In the time slot The amount of data to be processed in the computational tasks. Indicates user In the time slot The computational intensity required per unit bit of data Indicates user In the time slot The maximum allowable completion time for the computation task; Based on the aforementioned UAV swarm status information, user communication and surveillance task requirements, and system resource constraints, a low-energy joint optimization model is constructed with the objective of minimizing the total energy consumption of the UAV swarm system; wherein, the UAVs In the time slot The total energy consumption of the system is expressed as: ； in, Represents flight status; Represents the communication status; Indicates drone In the time slot The energy consumption during flight and hovering. Indicates drone In the time slot The energy consumption of performing communication tasks; Indicates drone In the time slot The energy consumption for performing computational tasks. Indicates drone In the time slot The energy consumption for performing detection tasks; The low-energy joint optimization model is expressed as follows: ； Among them, the optimization variables are ; This represents the set of drone deployment locations. Indicates in time slot The set of matching related variables, Indicates in time slot The set of downlink bandwidth allocation variables, Indicates in time slot The set of allocation variables for downlink transmit power. Indicates in time slot The set of uplink bandwidth allocation variables, Indicates in time slot The set of CPU frequency allocation variables for computational tasks. Indicates in time slot The set of CPU frequency allocation variables for detection tasks; Represents the downlink; Represents the uplink; The system resource constraint information includes constraints applicable to the low-energy joint optimization model. These constraints include unique task association constraints, maximum number of UAV services constraints, resource budget constraints, safe distance constraints between UAVs, task area boundary constraints, and adjacent time slot travel distance constraints.

3. The low-energy joint optimization method for UAV swarms performing reconnaissance tasks according to claim 1, characterized in that... Step S2 specifically includes: The low-energy joint optimization model is divided into matching and correlation optimization sub-problems, communication and reconnaissance computing resource allocation optimization sub-problems, and UAV deployment optimization sub-problems according to the type of optimization variables. At a given drone deployment location Under the condition of matching associated variables To construct the matching association optimization subproblem for the optimization variables. : ； in Let P1 represent the total energy consumption of UAV m in time slot t. Subproblem P1 is to obtain the minimum total energy consumption of M UAVs in all time slots. The constraints of the communication, reconnaissance, and computation task matching and association optimization subproblem include unique association constraints for communication tasks, unique association constraints for computation tasks, unique association constraints for reconnaissance tasks, and constraints on the maximum number of times the UAV can serve. At a given drone deployment location and matching associated variables Under the given conditions, the communication and surveillance computing resource allocation optimization sub-problem is constructed using the communication and surveillance computing resource allocation variable as the optimization variable. : ； The constraints of the communication and reconnaissance computing resource allocation optimization sub-problem include downlink bandwidth budget constraints, downlink transmit power budget constraints, uplink bandwidth budget constraints, airborne CPU frequency budget constraints, and communication and reconnaissance computing task completion constraints. Given matching associated variables Under the conditions of resource allocation results and reconnaissance, the deployment location of the UAV is considered. To optimize the variables, construct the UAV deployment optimization sub-problem. : ； The constraints of the UAV deployment optimization sub-problem include the safe distance constraints between UAVs, the boundary constraints of the mission area, the travel distance constraints between adjacent time slots, and the completion constraints of the communication and reconnaissance task. The matching and correlation optimization sub-problem, the communication and reconnaissance computing resource allocation optimization sub-problem, and the UAV deployment optimization sub-problem are solved alternately in that order.

4. The low-energy joint optimization method for UAV swarms performing reconnaissance tasks according to claim 1, characterized in that, Step S3 specifically includes: Based on the current drone deployment location and user location, calculate the drone... With users Spatial distance between And construct matching edge weights using the negative values ​​of the spatial distance: ； in, Indicates the task type. This represents the weight values ​​of the different edges formed by different users and drones during the matching process; Based on the maximum number of services per drone Expanding the drone nodes to generate a virtual drone node set transforms the capacity-constrained one-to-many matching problem into a one-to-one maximum weight matching problem. The objective function of the one-to-one maximum weight matching problem is: ； in, For task type Unmanned aerial vehicle With users In the time slot Related variables; Indicates drone Matching users Weight value at time For executable tasks A collection of drones For having a task The set of users who require it; An auction matching algorithm is used to solve the one-to-one maximum weight matching problem. In the auction matching algorithm, unmatched user tasks select bidding nodes based on their net income from each virtual drone node. Virtual drone nodes update node prices and establish or replace matching relationships based on bidding until the user task matching association is completed or a preset termination condition is met. The matching association result is constructed based on the final matching relationship.

5. The low-energy joint optimization method for UAV swarms performing reconnaissance tasks according to claim 1, characterized in that: Step S4 specifically includes: Let users {n} be a large set, which includes users of communication tasks. I Users of detection tasks j and users of computing tasks k ; For communication tasks, given the known deployment locations of drones and the matching association results, determine the drones... Assigned to communication users in communication tasks downlink bandwidth and downlink transmit power The downlink transmission rate is calculated according to the following formula: ； in, Indicates drone With users The equivalent channel power gain between Indicates the noise power spectral density; Under the conditions of satisfying the minimum access signal-to-interference-plus-noise ratio threshold and the maximum completion time of the communication task, determine the minimum feasible downlink bandwidth and minimum feasible downlink transmit power of the communication task: ； ； For detection missions, determine the drone Detection users assigned to detection tasks CPU frequency detection And determine the minimum feasible detection CPU frequency based on the detection and processing latency constraints: ； in, Indicates the strength of the detection calculation. Indicates drone When performing a detection task, detect the user The amount of detection data generated; For computational tasks, determine the drone. Assigned to users in computing tasks Uplink bandwidth And calculate CPU frequency The uplink unloading rate is calculated using the following formula: ； in, Indicates computing user Uplink transmit power; Given uplink bandwidth Under the given conditions, determine the minimum feasible CPU frequency based on the total computational task latency constraint: ； The computational resource allocation problem is transformed into a single-variable optimization problem concerning uplink bandwidth. The uplink bandwidth allocation result is determined by Lagrange dual update and one-dimensional search, and the computational CPU frequency allocation result is recovered based on the minimum feasible computational CPU frequency. The minimum feasible downlink bandwidth, minimum feasible downlink transmit power, minimum feasible detection CPU frequency, uplink bandwidth allocation result, and calculated CPU frequency allocation result are used as the communication detection computing resource allocation result.

6. The low-energy joint optimization method for UAV swarms performing reconnaissance tasks according to claim 1, characterized in that: Step S5 specifically includes: The particle swarm optimization algorithm is used to solve the subproblem of optimizing the deployment location of the UAV; where each particle represents a three-dimensional deployment scheme of the UAV, and the i-th particle represents the t-th particle. The position vector of each particle is represented as: ； in, Indicates the first The first particle The three-dimensional deployment location of the drone; Using the drone deployment location of the previous time slot as the center, the particle position is initialized with random perturbation. Construct a particle fitness function based on total system energy consumption and business constraint default terms: ； in, Indicates the first Sub-iteration particles Total system energy consumption under the corresponding deployment scheme , , These represent the violation terms of the communication task constraints, the computation task constraints, and the detection task constraints, respectively. This is the penalty coefficient; Update the individual and group historical best positions based on the fitness values ​​of each particle, and update the particle velocity and position according to the following formula: ； ； in, Indicates the first In the nth iteration The speed of each particle Indicates the first The individual best position in the history of each particle. This indicates the group's historical best position. Indicates inertia weight, It is an individual learning factor. It is a group learning factor. It is the first random number. It is the second random number; When the particle swarm iteration reaches the preset number of deployment optimization iterations, or when the fitness change meets the preset deployment convergence condition, the UAV 3D deployment scheme corresponding to the swarm's historical best position is used as the updated UAV deployment position.

7. The low-energy joint optimization method for UAV swarms performing reconnaissance tasks according to claim 1, characterized in that: Step S6 specifically includes: Initialize the drone deployment location and calculate the initial total system energy consumption; In the In the alternating iterations, the matching association optimization sub-problem is solved based on the current drone deployment location to obtain the... This match is associated with the results; Based on the current drone deployment location and the Solving the communication and detection resource allocation optimization sub-problem using the matching and association results of the first step yields the first... The results of resource allocation for secondary surveillance; Based on the The result of the second matching association and the first The results of the second reconnaissance resource allocation are used to solve the UAV deployment optimization sub-problem, resulting in the first... The updated drone deployment locations; According to the The updated drone deployment location results, the The result of the second matching association and the first The calculation of the next detection resource allocation result is as follows: The total energy consumption of the subsystem is calculated, and the change in total system energy consumption between two adjacent alternating iterations is also calculated: ； when Less than or equal to the preset convergence threshold, or When the preset maximum number of iterations is reached, the alternating iteration stops, and the final task matching and association results, communication resource allocation results, detection resource allocation results, computing resource allocation results, and UAV deployment location results are output.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

Images