Unmanned aerial vehicle cooperative scheduling and target searching method for emergency rescue

Abstract

The invention discloses an unmanned aerial vehicle cooperative scheduling and target searching method for emergency rescue, and belongs to the technical field of unmanned aerial vehicle control, and the method comprises the steps: constructing an unmanned aerial vehicle operation model and an operation time constraint according to an unmanned aerial vehicle communication network system and a preset ground node; decomposing the optimization target of the flight task into a flight path optimization sub-target and a resource scheduling optimization sub-target to respectively obtain sub-optimization problems of a target sensing task and a communication task; and based on the operation time constraint, iteratively solving each sub-optimization problem through the unmanned aerial vehicle operation model to obtain an optimal flight strategy, and regulating and controlling a flight path and communication resources in real time through the strategy. Therefore, by implementing the method and the device, the problem that in the prior art, as efficient resource allocation and scheduling cannot be realized between communication service and perception search, the rescue target information is difficult to transmit to the background in time while the rescue target is quickly and accurately positioned can be solved.

Description

Technical Field

[0001] This application belongs to the field of unmanned aerial vehicle (UAV) control technology, specifically relating to a method for collaborative scheduling and target search of UAVs for emergency rescue. Background Technology

[0002] With the maturation of drone control technology, drones, with their core advantages such as rapid deployment, high mobility, and lack of terrain limitations, have become a key technological means to overcome the difficulties of emergency rescue in emergency scenarios such as natural disasters. Therefore, enabling drones to locate survivors or key targets scattered throughout disaster areas and transmit the location information to the backend in a timely manner can significantly improve rescue efficiency.

[0003] Existing drone dispatching solutions for emergency rescue employ codec-based machine learning methods to optimize communication quality and improve information transmission efficiency by increasing network throughput, expanding coverage, and reducing latency. However, this approach relies on the premise that the ground-based nodes with data collection capabilities are known or nearly known. While this method effectively ensures communication services, it struggles to quickly search for and locate trapped individuals in unknown locations. Therefore, it fails to meet the critical need for real-time detection and location of unknown targets in emergency search and rescue, and thus cannot solve the core challenges in actual rescue operations. Therefore, in emergency rescue, how to rapidly locate rescue targets in disaster areas using drone technology and promptly transmit the location information to the backend, breaking down the technical bottleneck of conflicting "perception and search" and "communication services," and the lack of a unified optimization goal for resource scheduling, is a pressing issue that needs to be addressed. Summary of the Invention

[0004] This application proposes a method for collaborative scheduling and target search of unmanned aerial vehicles (UAVs) for emergency rescue. It can solve the problem that existing technologies cannot achieve efficient resource allocation and scheduling between communication services and perception search, thus making it difficult to quickly and accurately locate rescue targets while timely transmitting rescue target information to the backend.

[0005] The first aspect of this application provides a method for collaborative scheduling and target search of unmanned aerial vehicles (UAVs) for emergency rescue, the method comprising:

[0006] Based on the pre-set UAV communication network system and ground nodes in the current rescue area, a UAV operation model and operation time constraints are constructed by setting the UAV's flight mission; wherein, the ground nodes are nodes with data acquisition capabilities deployed in the current rescue area.

[0007] Based on the UAV communication network system, the optimization objective of the flight mission is decomposed into a flight trajectory optimization sub-objective and a resource scheduling optimization sub-objective, respectively yielding sub-optimization problems for the target perception task and the communication task;

[0008] Based on the aforementioned runtime constraints, the optimal flight strategy is obtained by iteratively solving each of the sub-optimization problems using the UAV operation model.

[0009] The optimal flight strategy is used to adjust the flight trajectory of the drone and schedule its communication resources in real time.

[0010] The above scheme, based on the existing UAV communication network system and the current rescue area, constructs a UAV operation model and related runtime constraints during flight missions. The UAV operation model reveals the UAV's flight trajectory and communication resource allocation during communication tasks, while the runtime constraints limit the mission duration, providing a data foundation for subsequent resource scheduling and flight trajectory planning. The overall mission optimization objective is then decomposed into two sub-objectives. By solving the optimization problems of each sub-objective, efficient coordination of task resources between perception / search and communication services is achieved, avoiding mutual interference between tasks. Finally, by simultaneously solving the results of the two sub-optimization problems, an optimal flight strategy is established, achieving coordinated optimization of flight trajectory planning and information transmission. This ensures that UAVs can achieve coordinated operation of target localization and timely communication even in complex and changing rescue environments with multiple unknown locations, enhancing the adaptability and response speed of the UAV control system to complex scenarios and effectively improving rescue response efficiency.

[0011] In one possible implementation of the first aspect, based on a pre-set UAV communication network system and the current rescue area, a UAV operation model and operation time constraints are constructed by setting the UAV's flight mission, specifically as follows:

[0012] According to the preset drone working mechanism, the drone's flight mission is divided into several time intervals, and each time interval is further divided according to the duration of the flight mission; wherein, the flight mission includes target perception mission and communication mission.

[0013] Based on the current rescue area, obtain the drone's initial and final flight positions;

[0014] Based on the starting and ending flight positions, the flight trajectory of the UAV is represented by a first threshold number of time intervals, the flight trajectory is adjusted, and a UAV operation model is constructed.

[0015] The operating time constraint is constructed based on the rescue perception data fed back by the ground node and the spatial location information fed back by the UAV, combined with the remaining battery power of the UAV.

[0016] The aforementioned scheme divides a complete flight mission into several time slots, each dedicated to either target perception or communication tasks. This constructs a dynamic switching strategy for the UAV's perception and communication modes, providing data support for the efficient coordination of task resources between subsequent perception search and communication services, and avoiding mutual interference between different tasks. Furthermore, by establishing runtime constraints, subsequent resource allocation is made to conform to the real-time scenario, ensuring that the UAV's remaining battery power can support the successful completion of the flight mission.

[0017] In one possible implementation of the first aspect, the flight trajectory is adjusted to construct a UAV operation model, specifically as follows:

[0018] Based on the positioning signal received by the UAV from the first ground node and combined with the UAV's current location, the UAV's perception range is adjusted; wherein the first ground node is within the adjusted perception range;

[0019] The flight trajectory is updated based on the adjusted perception range;

[0020] To meet the updated flight trajectory, the drone's flight speed is adjusted based on a preset maximum flight speed limit.

[0021] Based on the updated flight trajectory and adjusted flight speed, a drone operation model is constructed.

[0022] The above scheme determines whether the UAV can sense the ground node by receiving the positioning signal sent by the ground node. Therefore, by adjusting the UAV's sensing range, a flight trajectory that enables the UAV to sense and locate the ground node is obtained. Furthermore, by adjusting the flight trajectory, the corresponding flight speed is adjusted to obtain a UAV operation model that can characterize the UAV's real-time operation, providing a data foundation for subsequent resource allocation and trajectory planning.

[0023] In one possible implementation of the first aspect, based on the UAV communication network system, the optimization objective of the flight mission is decomposed into a flight trajectory optimization sub-objective and a resource scheduling optimization sub-objective, respectively yielding sub-optimization problems for the target perception task and the communication task, specifically:

[0024] Based on the UAV communication network system and the UAV's flight mission, an initial optimization target is constructed; wherein, the flight mission includes target perception mission and communication mission;

[0025] By evaluating the duration of the first task of completing the target perception task and the duration of the second task of completing the communication task, the initial optimization objective is decomposed into a flight trajectory optimization sub-objective and a resource scheduling optimization sub-objective, thus obtaining the sub-optimization problems of the target perception task and the communication task.

[0026] The above scheme first constructs an overall initial optimization goal, and then decomposes this overall goal into different sub-goals. Each sub-goal corresponds to a task type, thus avoiding mutual interference between tasks in subsequent resource coordination.

[0027] In one possible implementation of the first aspect, the initial optimization objective is decomposed into a flight trajectory optimization sub-objective and a resource scheduling optimization sub-objective by evaluating the first task duration for completing the target perception task and the second task duration for completing the communication task, thereby obtaining the corresponding sub-optimization problem, specifically:

[0028] Based on the UAV communication network system, the time interval from when the ground node generates a data packet, uploads the data packet to the UAV, and receives the feedback signal from the UAV is evaluated to obtain the second mission duration;

[0029] The time interval from generating the target perception task to locating the corresponding ground node is evaluated to obtain the duration of the first task;

[0030] Based on the first task duration and the second task duration, the initial optimization objective is decomposed into the flight trajectory optimization sub-objective and the resource scheduling optimization sub-objective;

[0031] Combining the aforementioned flight trajectory optimization sub-objective with preset perception scheduling constraints, a sub-optimization problem for the target perception task is constructed;

[0032] By combining the resource scheduling optimization sub-objective and the preset communication scheduling constraints, a sub-optimization problem for the communication task is constructed.

[0033] The above scheme provides a scientific basis for multi-task scheduling of UAVs by evaluating the duration of target perception tasks and communication tasks, thereby prioritizing the allocation of limited resources to the most urgent tasks and effectively improving rescue response efficiency.

[0034] One possible implementation of the first aspect also includes:

[0035] Based on the distance between the UAV's current location and the located ground node, determine the channel quality between the UAV and the located ground node;

[0036] Based on the current location of the UAV, determine the UAV's perception range and search progress for the unlocated ground nodes;

[0037] The sub-optimization problem of the communication task is adjusted using the channel quality; the sub-optimization problem of the target perception task is adjusted using the sensing range and search progress.

[0038] The above scheme also considers the impact of other factors on target perception and communication tasks, thereby making reasonable adjustments to the sub-optimization problem and providing a scientific decision-making basis for subsequent UAV multi-task scheduling.

[0039] In one possible implementation of the first aspect, based on the runtime constraint, the optimal flight strategy is obtained by iteratively solving each of the sub-optimization problems using the UAV operation model. Specifically:

[0040] Using the particle swarm optimization algorithm, the optimal UAV position is searched through the UAV operation model, and the sub-optimization problem of the target perception task is solved iteratively to obtain the optimal UAV flight trajectory.

[0041] The Lagrange dual decomposition method is used to decouple the sub-optimization problems of the communication task and obtain the optimal resource allocation strategy.

[0042] Based on the optimal UAV flight trajectory and the optimal resource allocation strategy, construct the optimal flight strategy.

[0043] The above scheme achieves coordinated optimization of flight trajectory planning and data communication by iteratively solving different sub-optimization problems, effectively alleviating the problems of task conflict and resource coordination in UAV emergency rescue, and improving the adaptability and response speed of UAV control in complex dynamic environments.

[0044] In one possible implementation of the first aspect, the optimal UAV position is searched using the UAV operation model, and the sub-optimization problem of the target perception task is solved iteratively to obtain the optimal UAV flight trajectory, specifically as follows:

[0045] When the UAV receives the positioning signal transmitted by the ground node, it triggers the perception mode; in the perception mode, the UAV is controlled using a preset flight trajectory.

[0046] When the UAV does not receive the positioning signal transmitted by the ground node, based on the running time constraint, the current optimal UAV position is searched through the UAV running model in each iteration until the preset maximum number of iterations is met or the optimal solution of the particle swarm has been found, and the final optimal UAV position is output to obtain the optimal UAV flight trajectory.

[0047] The above scheme distinguishes ground nodes into known and unknown ones. For unknown ground nodes, target perception is prioritized while communicating to achieve rapid node positioning. For known nodes, control is performed according to a preset flight trajectory, and resources are prioritized for communication tasks to achieve efficient resource allocation.

[0048] In one possible implementation of the first aspect, the Lagrange dual decomposition method is used to decouple the sub-optimization problems of the communication task, thereby obtaining the optimal resource allocation strategy, specifically:

[0049] By using the Lagrange dual decomposition method, the sub-optimization problem of the communication task is decoupled into several dual sub-problems, each of which corresponds to a communication resource scheduling problem. The communication resource scheduling problem includes the data upload scheduling of the ground node, the computing resource scheduling of the UAV, and the transmission power allocation of the ground node.

[0050] Based on the aforementioned runtime constraints, each dual subproblem is solved to obtain the optimal solution for each dual subproblem;

[0051] By integrating the optimal solutions to each dual subproblem, an optimal resource allocation strategy is constructed.

[0052] In one possible implementation of the first aspect, an optimal flight strategy is used to control the UAV's flight trajectory and schedule its communication resources in real time, specifically:

[0053] Using the optimal flight strategy, the flight trajectory of the UAV is adjusted in real time based on the communication between the UAV and the ground node;

[0054] The optimal flight strategy is used to schedule the communication resources of the UAV by adjusting the transmission power and data upload rate of the ground node and adjusting the computing resource scheduling of the UAV.

[0055] The above solution employs an optimal flight strategy, adjusting the drone's flight trajectory to ensure it can receive positioning signals from ground nodes within its perception range, thus enabling rapid location of the rescue target. Furthermore, resource scheduling limits communication duration, ensuring the drone can promptly relay rescue information to the backend, facilitating rapid rescue operations. Attached Figure Description

[0056] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0057] Figure 1 This is a schematic diagram illustrating the specific process of a method for collaborative scheduling and target search of unmanned aerial vehicles (UAVs) for emergency rescue, provided in an embodiment of this application.

[0058] Figure 2 This is a flight mission execution flowchart of a method for collaborative scheduling and target search of unmanned aerial vehicles (UAVs) for emergency rescue, provided in one embodiment of this application.

[0059] Figure 3 This is a schematic diagram of the working mechanism of a UAV collaborative scheduling and target search method for emergency rescue provided in an embodiment of this application;

[0060] Figure 4 This is a schematic diagram of flight trajectory adjustment in the perception mode of a method for collaborative scheduling and target search of unmanned aerial vehicles (UAVs) for emergency rescue provided in an embodiment of this application. Detailed Implementation

[0061] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0062] It should be understood that the step numbers used in the text are for ease of description only and are not intended to limit the order in which the steps are performed.

[0063] First Embodiment

[0064] In emergency rescue scenarios, the flight mission of a drone not only involves timely transmission of rescue target information to the backend, but also requires precise location of ground nodes to receive feedback on the rescue target information. Therefore, in this case, the flight mission can be divided into target perception and communication tasks. This application's embodiment is the first to incorporate the timeliness of the target perception task into the optimization objective. By allocating resources and optimizing flight trajectories, it alleviates the conflicts and resource coordination problems between the target perception and communication tasks, enabling stable and efficient operation even in complex environments. It quickly and accurately locates the rescue target while promptly transmitting the rescue target information to the backend, ensuring the smooth execution of the rescue mission.

[0065] like Figure 1 As shown, to address the problem in existing technologies where efficient resource allocation and scheduling between communication services and perception search are impossible, making it difficult to quickly and accurately locate rescue targets while promptly transmitting rescue target information to the backend, the first embodiment of this application provides a detailed flowchart of a UAV collaborative scheduling and target search method for emergency rescue. This embodiment of the UAV collaborative scheduling and target search method for emergency rescue includes steps S1 to S4, detailed below:

[0066] Step S1: Based on the preset UAV communication network system and the ground nodes in the current rescue area, construct the UAV operation model and operation time constraints by setting the UAV flight mission.

[0067] To better understand how drones perform flight missions, Figure 2 The flight mission execution flowchart provided in this application embodiment illustrates an emergency rescue process based on the integrated sensing, computing, and communication capabilities of a rotary-wing UAV. Dashed lines with arrows represent the UAV's flight trajectory, and circles constructed with dashed lines represent the UAV's sensing coverage area. First, a UAV, also known as a UAV, is deployed within the rescue area. Equipped with sensing devices, it is responsible for searching and locating K ground nodes within the area. The UAV dynamically executes its search mission even when the node locations are completely unknown, and interacts with the located ground nodes in real time via a data upload link to collect regional environmental information, achieving comprehensive perception of the rescue site situation. The collected data is then transmitted to the backend, i.e., the system control center. The UAV is equipped with a mobile edge computing server to support comprehensive analysis of the data collected from the ground nodes. The UAV takes off from a designated starting point, performing the ground node search mission (i.e., target perception mission), while simultaneously providing emergency communication and computing resources (i.e., communication mission) to the discovered ground nodes. If two or more ground nodes simultaneously fall within the UAV's communication coverage area and both have data upload missions, it is assumed that the ground nodes communicate with the UAV using FDMA.

[0068] Ground nodes are deployed within the emergency rescue area and possess data collection capabilities. These include smart devices carried by trapped personnel and environmental sensors. When the location of ground nodes is unknown, drones are used to perform target perception tasks for localization. The data upload task refers to the ground nodes uploading the collected perception data to the drone's edge computing server via the drone's communication link. This enables real-time data transmission and centralized processing, helping to improve rescue response efficiency and data processing capabilities.

[0069] The flight mission proposed in this application includes target perception mission and communication mission, indicating that the UAV not only needs to search for and locate ground nodes, but also needs to quickly feed back the rescue perception data uploaded by the ground nodes to the background to provide support for subsequent emergency rescue.

[0070] Since target perception and communication tasks are clearly distinct, they conflict in resource scheduling. To fundamentally resolve the mutual interference between target perception and communication tasks, this application introduces a time-division multiplexing-based UAV working mechanism to improve emergency rescue efficiency, specifically as follows: Figure 3 As shown. In Figure 3In this approach, the UAV's flight mission is divided into T time slots, each with a length of τ. Each time slot is then further divided into "sensing sub-time slots" or "communication / computation sub-time slots," determining the duration of target sensing and communication tasks within each time slot. For example, in... Figure 3 In the second time slot, the time allocated to the target perception task is β·τ, and the time allocated to the communication task is (1-β)·τ. Here, a weighting coefficient β is introduced to dynamically adjust the time ratio of the two tasks, achieving flexible allocation of time resources for the target perception and communication tasks. This design effectively avoids resource competition between the two types of tasks, laying a solid foundation for the implementation of subsequent collaborative scheduling schemes.

[0071] Based on the current rescue area and the aforementioned drone operating mechanism, the drone's initial flight position and final flight position are obtained. Then, based on the initial flight position and final flight position, the drone's flight trajectory is represented using a first threshold number of time intervals.

[0072] For example, assuming the drone flies at a constant altitude H, its flight trajectory is approximated by a two-dimensional discrete sequence {U(1), U(2), ..., U(T)} with T time intervals. Let U(t) represent the three-dimensional coordinate position of the UAV during time interval t, where U(1) is the initial flight position of the UAV, and U(T) is the final position of the UAV after completing its flight mission; uav (t) and y uav (t) represents the horizontal and vertical coordinates of the UAV in time slot t, respectively.

[0073] In addition, this application embodiment also assumes that the channel between the UAV and the ground node is a LoS channel.

[0074] For UAV positioning of ground nodes, it is assumed that the ground nodes periodically transmit positioning signals, including the node ID, to the UAV to facilitate identification and search. During the process of receiving the positioning signals and completing positioning, the UAV needs to continuously observe the same ground node multiple times to improve positioning accuracy. It is set that the UAV needs to continuously observe the same ground node M times, with an observation time slot of M-1. That is, except for the initial signal reception, the UAV needs to continuously receive the positioning signal of the ground node for M-1 time slots.

[0075] Optionally, M is a system design variable that can be adjusted according to task requirements and positioning accuracy requirements.

[0076] Therefore, after receiving a positioning signal from a ground node, the UAV needs to adjust its sensing range based on its current location so that the ground node falls within the adjusted sensing range. In this way, the UAV can continuously observe the ground node and achieve accurate positioning of it.

[0077] As an improvement to the above solution, the embodiments of this application are designed as follows: Figure 4 The flight trajectory adjustment method is shown below. Assuming the UAV does not receive a positioning signal from the ground node at point a1, but receives the signal for the first time at point a2 in the next time slot, then the path extension line R between points a1 and a2 is calculated using the current UAV position as the starting point. s With the center at point / 2, construct a circle with a radius of r = R. s The UAV will fly along a circular trajectory of 2 / 2 for the next M-1 consecutive time slots, thereby ensuring that the ground node is within the UAV's perception range R as much as possible. s Within, multiple effective observations were completed.

[0078] Therefore, the flight trajectory can be updated based on the adjusted sensing range using the above-described flight trajectory adjustment method.

[0079] To meet the updated flight trajectory requirements, the drone's flight speed also needs to be adjusted in conjunction with the preset maximum flight speed limit.

[0080] For example, based on Figure 4 The demonstrated method for adjusting the flight trajectory involves adjusting the drone's flight speed to... This allows the UAV to complete a full orbital flight within M-1 consecutive time slots, ensuring sufficient positioning signal measurement samples are obtained. After the flight, the UAV combines the collected M positioning signal data, and this invention uses a Bayesian estimation algorithm to recursively update the three-dimensional position of the ground node. This algorithm, based on prior position probability distribution and observation data, uses a Bayesian filtering method to fuse measurement information from multiple signal samplings, progressively optimizing the position estimation accuracy and effectively reducing the impact of measurement errors and environmental interference. It also uses a binary index function L... k (t) represents node s k In the positioning state of time slot t, if L k (t)=1, then s k Location has been successfully established; otherwise, it has not.

[0081] Therefore, based on the above-mentioned flight trajectory adjustment scheme and flight speed adjustment scheme, an unmanned aerial vehicle (UAV) operation model is constructed with the updated flight trajectory and the adjusted flight speed to characterize the state of the UAV when performing flight missions.

[0082] Furthermore, in addition to uploading the collected rescue perception data to the drones, the ground nodes, as intelligent terminals with environmental perception and computing capabilities, are also responsible for collecting data in real time and generating corresponding task packages based on the collected content. These task packages cover important information such as environmental analysis and survivor detection in disaster relief scenarios. The generated task packages are first stored in the data cache of each ground node. This buffering mechanism aims to balance the task generation rate and upload capacity, preventing task package loss and delays, and ensuring the continuity and integrity of the task packages.

[0083] This application embodiment distinguishes task packages with different business requirements based on the size of the task package generated by the ground node and the number of CPU cycles required. If node s k If the cache in time slot t is empty, then both the task packet size and the number of CPU cycles are 0. Ground node s k When the cache is not empty, use the binary index function a. k (t) represents node s k Does it upload its mission packet to the drone in time slot t? If a k (t)=1, then s k Determine the task package to upload; if a k (t) = 0 indicates that the node is in a waiting state and will not upload the task package.

[0084] The drone completed the task of monitoring the ground node. k After the search and location are completed, and the node is within the communication coverage area R of the drone... c Within the system, ground nodes communicate with the UAV equipment and upload their mission packets to the UAV. The horizontal coverage radius of the UAV is expressed as R. c =H·tan(θ), where θ is the azimuth angle of the UAV and H is the flight altitude of the UAV.

[0085] In addition, since the power supply to ground nodes is basically limited, in this embodiment of the application, the ground nodes in the UAV communication network system do not perform local calculations, but upload the task packets to the MEC server on the UAV through the constructed communication channel to ensure the continuous operation of the ground nodes.

[0086] Finally, rescue perception data is obtained based on the task package uploaded by the ground node; the two-dimensional spatial position information of the drone is obtained from the drone, and combined with the drone's remaining battery power, an operating time constraint is constructed. The specific expression of the operating time constraint is as follows:

[0087]

[0088] In the formula, T represents the total flight time of the UAV, and E... max For the remaining battery power of the drone, F maxP represents the total number of CPU cycles for the drone. move c represents the propulsion power consumption of the drone during horizontal movement, and c is the effective switched capacitor.

[0089] Step S2: Based on the UAV communication network system, the optimization objective of the flight mission is decomposed into a flight trajectory optimization sub-objective and a resource scheduling optimization sub-objective, respectively obtaining the sub-optimization problems of the target perception task and the communication task.

[0090] First, based on the UAV communication network system, an initial optimization objective is constructed. In fact, this initial optimization objective is a multi-objective optimization problem that combines communication and perception performance, and includes an initial multi-objective function and several scheduling constraints.

[0091] The initial multi-objective function describes the minimization problem of the average weighted AoT and average weighted AoPT of all ground nodes within T time intervals (i.e., the total flight mission duration). AoT (Age of Task) measures the time interval from when a ground node generates a data packet, uploads the data packet (i.e., the task packet) to the UAV, and receives the UAV's feedback signal; it represents the time required to complete the communication task. AoT can be used to evaluate the service efficiency and responsiveness of UAV mobile edge computing. AoPT measures the time interval from generating the target perception task to locating the corresponding ground node; it represents the time required to complete the target perception task. AoPT can be used to evaluate the timeliness of UAV search and localization.

[0092] The expression for the initial multi-objective function is as follows:

[0093]

[0094] In the formula, w k and v k For ground node s respectively k The weighted coefficients for the time taken to complete the communication task and the time taken to complete the target perception task, where U(t) is the flight trajectory of the UAV in time slot t, and a k (t) represents the task packet upload scheduling strategy for the UAV in time slot t. To assign UAVs to ground nodes s in time slot t k CPU cycle count For ground node s k The transmit power in time slot t, where K is the total number of ground nodes.

[0095] The scheduling constraint is specifically expressed as follows:

[0096] l t,t+1 ≤v max ·τ,1<<t<<T-1; (1)

[0097]

[0098]

[0099] E uav ≤E max (14)

[0100]

[0101] In the above formulas (1) to (16), formula (1) represents the flight trajectory constraint of the UAV, l t,t+1 Let b be the flight distance of the UAV in the t-th time slot; in equation (2) k (t) represents the indicator variable for the UAV entering the perception mode, i.e., whether the UAV performs the target perception task, 1 indicates execution, and 0 indicates non-execution; Equation (3) represents the speed constraint for the UAV to enter the perception mode; L in Equation (4) k (t) indicates whether the k-th ground node has completed positioning, a k (t) indicates that any task packet is either uploaded or waiting to be transmitted in each time slot; Equation (5) is the transmit power limit constraint for ground equipment, P max At maximum transmission power, For ground node s k The transmit power in the t-th time slot; Equation (6) represents the average transmission power constraint of the ground node. The average transmission power limit for each node is given; Equation (7) indicates that the UAV can only schedule ground equipment within the coverage area of ​​a located node. Let Z be the square of the horizontal distance between the UAV and the ground node in the t-th time slot; Equation (8) indicates that the number of ground nodes scheduled cannot exceed the total number of sub-channels of the UAV's FMDA communication resources, where Z is the total number of sub-channels; Equation (9) represents the time required to upload the task packet. It consists of two parts: task data transmission time. and task calculation time Equation (10) represents the allocated computing resource constraint, where represents the number of CPU cycles allocated to the scheduled ground nodes. The total CPU resources of the drone's MEC server cannot exceed F. max Equation (11) indicates that the number of CPU cycles allocated to the scheduled ground node cannot exceed the total CPU resources of the UAV MEC server; Equation (12) indicates that the communication task m k (t) is the indicator function for whether the data queue can be successfully executed; Equation (13) represents the constraint on the stability of the data queue. Represents node s k Whether to generate an emergency mission package at the start of the t-th time slot; Equation (14) represents the UAV energy consumption constraint, Euav E represents the current energy consumption of drones. max The maximum energy consumption that the UAV can use; Equation (15) is the evolution formula of AoT for each ground node, where This represents the generation time of the task packet transmitted in the t-th time slot, m. k (t) represents the scheduling node s k Whether the communication task can be completed within the t-th time slot; Equation (16) represents the AoPT evolution constraint for each ground node. This indicates the generation time of the target perception task.

[0102] In this embodiment, the initial optimization objective, which includes an initial multi-objective function, is decomposed into a flight trajectory optimization sub-objective and a resource scheduling optimization sub-objective. The resource scheduling optimization sub-objective is used to determine whether to allow ground nodes to upload task packages to the UAV, and how much computing resources to allocate to the UAV and ground nodes.

[0103] Specifically, based on the preset optimization framework, and combining the task queue status (i.e., task packets waiting to be uploaded or being uploaded) generated by all ground nodes in the current time interval, the first task duration for completing the target perception task (implemented through AoPT technology) and the second task duration for completing the communication task (implemented through AoT technology) are evaluated, the initial optimization objective is decomposed into flight trajectory optimization sub-objective and resource scheduling optimization sub-objective.

[0104] Because the location of the UAV is highly coupled with task scheduling and resource allocation, the original initial optimization objective is decomposed into two sub-objectives for the purposes of this application embodiment, thereby obtaining the corresponding sub-optimization problem. This enables the priority allocation of limited resources to the most urgent tasks, effectively improving the efficiency of rescue response, and providing a foundation for real-time optimization of subsequent task scheduling and resource allocation.

[0105] The sub-optimization problem of the target perception task includes an objective function for the flight trajectory optimization sub-objective and corresponding perception scheduling constraints. The objective function for the flight trajectory optimization sub-objective is specifically expressed as follows:

[0106]

[0107] For the sub-optimization problem of the communication task, there is an objective function based on the resource scheduling optimization sub-objective and corresponding communication scheduling constraints. The specific expression of the objective function of the resource scheduling optimization sub-objective is as follows:

[0108]

[0109] The perception scheduling constraint and the communication scheduling constraint are both derived from the scheduling constraints in equations (1) to (16) above.

[0110] Furthermore, since the distance between the UAV's position and the located ground nodes determines the channel quality for those nodes, and this channel quality directly affects scheduling decisions and resource allocation optimization, the sub-optimization problem of the communication task can be adjusted based on the channel quality. Additionally, the UAV's current position also affects the perception range and search progress for unlocated ground nodes; therefore, the sub-optimization problem of the target perception task can be adjusted based on the perception range and search progress.

[0111] Step S3: Based on the running time constraint, the optimal flight strategy is obtained by iteratively solving each of the sub-optimization problems using the UAV operation model.

[0112] Equations (2) and (3) in step S2 are the speed and trajectory constraints of the UAV in perception mode. When the UAV receives the positioning signal transmitted by the ground node, it triggers perception mode. In perception mode, the UAV is controlled by a preset flight trajectory and flies along a pre-calculated circular trajectory for the subsequent M-1 time slots. At this time, there is no need to optimize the flight trajectory in real time, that is, there is no need to solve the sub-optimization problem of the target perception task. The flight speed is adjusted according to equation (3) to ensure that one circle of flight is completed within M time slots. Of course, the UAV also needs to take into account the communication task while perceiving the ground node.

[0113] When the UAV does not receive a positioning signal transmitted by the ground node, the particle swarm optimization algorithm is used to search for the optimal UAV position through the UAV operation model based on the running time constraint. The sub-optimization problem of the target perception task is solved iteratively to obtain the optimal UAV flight trajectory.

[0114] Specifically, when b in equation (2) k When (t) = 0, the position U(t) of the UAV at the current time t is affected by its position U(t-1) at the previous time t-1. The UAV's position U(t) can only be within a radius v centered on U(t-1). max The drone is located within the circle. Therefore, this embodiment uses a particle swarm optimization algorithm to search for the current optimal drone position through the drone operation model during each iteration until the preset maximum number of iterations is met or the optimal solution of the particle swarm has been found. Then, the final optimal drone position is output, and the optimal drone flight trajectory is obtained.

[0115] This application uses the Lagrange dual decomposition method to decouple the sub-optimization problems of the communication task, resulting in several dual sub-problems, each corresponding to a communication resource scheduling problem. The communication resource scheduling problem includes data upload scheduling of the ground nodes, computational resource scheduling of the UAV, and transmission power allocation of the ground nodes.

[0116] By solving the corresponding dual subproblems, the optimal solution can be obtained for each communication resource scheduling problem, namely, the optimal data upload scheduling of ground nodes, the computing resource scheduling of UAVs, and the transmission power allocation of the ground nodes.

[0117] Due to the aforementioned communication scheduling constraints, the UAV can only schedule ground nodes within its perception range, and because the task package uploads the decision variable a... k (t) (used to determine whether to upload the task package to the drone) is a binary variable, therefore, the embodiments of this application list the dual subproblems under different conditions below.

[0118] If the k-th ground node is not scheduled in the current time slot, i.e., a k When (t) = 0, the dual subproblem can be restated as:

[0119] min:V·w k ·(A k (t)+τ);

[0120] Clearly, the objective function of this problem is a constant. Therefore, when a k When (t) = 0, no optimization is needed.

[0121] If the kth ground node is scheduled, i.e., a k (t) = 1, and the dual subproblem can be restated as follows:

[0122]

[0123] It can be observed It is negative; in order to minimize the objective function, m... k (t) is set to 1. Therefore, for the dual subproblem in this case, the optimal feasibility test is first performed, which can be described as the following problem:

[0124]

[0125] Clearly, the optimal solution to this dual problem lies in and At the boundary value, i.e., F max and P max If the optimal value of the objective function is M(F) max ,P max If the value is less than (1-β)τ, it indicates that the dual subproblem has an optimal solution. Otherwise, the optimal solution to the dual subproblem is not found. and All are 0, and the optimal value of the corresponding objective function is also 0.

[0126] When the optimal value of the objective function is known to be less than (1-β)τ, that is, mk If (t) = 1, then we need to further determine the optimal solution. Therefore, we obtain a new subproblem:

[0127]

[0128] Let λ τ For the Lagrange multipliers associated with the aforementioned new subproblem, a portion of their Lagrange functions can be expressed as:

[0129]

[0130] And the corresponding Lagrange dual function G(λ) τ The following formula is given:

[0131]

[0132] Therefore, the duality problem is given as follows:

[0133]

[0134] On the other hand, the dual function G(λ) can be observed. τ The two constraints are respectively for the variables and They are independent and decoupled. Therefore, we can further decompose the above partial Lagrangian function into two subproblems: the computational resource scheduling subproblem and the transmit power allocation subproblem.

[0135] The computational resource scheduling subproblem can be expressed as:

[0136]

[0137] Clearly, the computational resource scheduling subproblem is a standard optimization problem with a convex objective function and linear constraints. Solving it yields the optimal computational resource scheduling solution:

[0138]

[0139] The transmit power allocation problem can be expressed as:

[0140]

[0141] The objective function of the transmit power allocation subproblem exhibits a trend of first decreasing and then increasing within its domain. In this case, an efficient one-dimensional line search technique, such as the golden section search, can be used to solve the transmit power allocation subproblem and obtain the optimal transmit power.

[0142] Once the optimal transmit power and optimal computational resource scheduling solution are obtained, the optimal data upload scheduling decision can be determined. Finally, by integrating the optimal transmit power, optimal computational resource scheduling solution, and optimal data upload scheduling decision, the optimal resource allocation strategy is obtained.

[0143] Based on the optimal UAV flight trajectory and the optimal resource allocation strategy, construct the optimal flight strategy.

[0144] This application embodiment achieves coordinated optimization of flight trajectory planning and data transmission through decomposition and iterative solution, ensuring that the system maintains stable and efficient operation in complex and changing rescue environments.

[0145] Step S4: Use the optimal flight strategy to adjust the drone's flight trajectory and schedule its communication resources in real time.

[0146] The optimal flight strategy includes the UAV's optimal flight trajectory, optimal transmission power, optimal computational resource scheduling solution, and optimal data upload scheduling decision. Therefore, using the optimal flight strategy, the UAV's flight trajectory is adjusted in real time based on the communication status between the UAV and the ground node; the optimal flight strategy also adjusts the ground node's transmission power and data upload rate, and regulates the UAV's computational resource scheduling, thereby managing the UAV's communication resources.

[0147] The above-mentioned adjustment plan effectively alleviated the problems of mission conflict and resource coordination in drone emergency rescue, and achieved significant improvements in mission timeliness and resource utilization, greatly enhancing the search efficiency and positioning accuracy of drones in unknown environments.

[0148] Implementing the embodiments of this application has the following beneficial effects:

[0149] This application embodiment constructs a UAV operation model and related runtime constraints based on existing UAV communication network systems and the current rescue area. The UAV operation model reveals the UAV's flight trajectory and communication resource allocation during communication tasks, while the runtime constraints limit the mission duration, providing a data foundation for subsequent resource scheduling and flight trajectory planning. The overall flight mission optimization objective is then decomposed into two sub-objectives. By solving the optimization problems of each sub-objective, efficient coordination of task resources between perception / search and communication services is achieved, avoiding mutual interference between tasks. Finally, by simultaneously solving the results of the two sub-optimization problems, an optimal flight strategy is established, achieving coordinated optimization of flight trajectory planning and information transmission. This ensures that the UAV can achieve coordinated operation of target positioning and timely communication even in complex and changing rescue environments with multiple unknown locations, enhancing the adaptability and response speed of the UAV control system to complex scenarios and effectively improving rescue response efficiency.

[0150] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above descriptions are merely specific embodiments of this application and are not intended to limit the scope of protection of this application. In particular, it should be noted that any modifications, equivalent substitutions, or improvements made by those skilled in the art within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for collaborative scheduling and target search of unmanned aerial vehicles (UAVs) for emergency rescue, characterized in that, include: Based on the pre-set UAV communication network system and ground nodes in the current rescue area, a UAV operation model and operation time constraints are constructed by setting the UAV's flight mission; wherein, the ground nodes are nodes with data acquisition capabilities deployed in the current rescue area. Based on the UAV communication network system, the optimization objective of the flight mission is decomposed into a flight trajectory optimization sub-objective and a resource scheduling optimization sub-objective, respectively yielding sub-optimization problems for the target perception task and the communication task; Based on the aforementioned runtime constraints, the optimal flight strategy is obtained by iteratively solving each of the sub-optimization problems using the UAV operation model. The optimal flight strategy is used to adjust the flight trajectory of the drone and schedule its communication resources in real time.

2. The method for collaborative scheduling and target search of unmanned aerial vehicles (UAVs) for emergency rescue according to claim 1, characterized in that, The process involves setting up a drone operation model and time constraints based on a pre-defined drone communication network system and the current rescue area, specifically by defining the drone's flight mission. According to the preset drone working mechanism, the drone's flight mission is divided into several time intervals, and each time interval is further divided according to the duration of the flight mission; wherein, the flight mission includes target perception mission and communication mission. Based on the current rescue area, obtain the drone's initial and final flight positions; Based on the starting and ending flight positions, the flight trajectory of the UAV is represented by a first threshold number of time intervals, the flight trajectory is adjusted, and a UAV operation model is constructed. The operating time constraint is constructed based on the rescue perception data fed back by the ground node and the spatial location information fed back by the UAV, combined with the remaining battery power of the UAV.

3. The method for collaborative scheduling and target search of unmanned aerial vehicles (UAVs) for emergency rescue according to claim 2, characterized in that, The adjustment of the flight trajectory and the construction of the UAV operation model specifically involves: Based on the positioning signal received by the UAV from the first ground node and combined with the UAV's current location, the UAV's perception range is adjusted; wherein the first ground node is within the adjusted perception range; The flight trajectory is updated based on the adjusted perception range; To meet the updated flight trajectory, the drone's flight speed is adjusted based on a preset maximum flight speed limit. Based on the updated flight trajectory and adjusted flight speed, a drone operation model is constructed.

4. The method for collaborative scheduling and target search of unmanned aerial vehicles (UAVs) for emergency rescue according to claim 1, characterized in that, The optimization objective of the flight mission is decomposed into a flight trajectory optimization sub-objective and a resource scheduling optimization sub-objective based on the UAV communication network system, resulting in sub-optimization problems for the target perception task and the communication task, respectively. Based on the UAV communication network system and the UAV's flight mission, an initial optimization target is constructed; wherein, the flight mission includes target perception mission and communication mission; By evaluating the duration of the first task of completing the target perception task and the duration of the second task of completing the communication task, the initial optimization objective is decomposed into a flight trajectory optimization sub-objective and a resource scheduling optimization sub-objective, thus obtaining the sub-optimization problems of the target perception task and the communication task.

5. The method for collaborative scheduling and target search of unmanned aerial vehicles (UAVs) for emergency rescue according to claim 4, characterized in that, The initial optimization objective is decomposed into flight trajectory optimization sub-objective and resource scheduling optimization sub-objective by evaluating the first task duration for completing the target perception task and the second task duration for completing the communication task, resulting in corresponding sub-optimization problems: Based on the UAV communication network system, the time interval from when the ground node generates a data packet, uploads the data packet to the UAV, and receives the feedback signal from the UAV is evaluated to obtain the second mission duration; The time interval from generating the target perception task to locating the corresponding ground node is evaluated to obtain the duration of the first task; Based on the first task duration and the second task duration, the initial optimization objective is decomposed into the flight trajectory optimization sub-objective and the resource scheduling optimization sub-objective; By combining the aforementioned flight trajectory optimization sub-objective and the preset perception scheduling constraints, a sub-optimization problem for the target perception task is constructed; By combining the resource scheduling optimization sub-objective and the preset communication scheduling constraints, a sub-optimization problem for the communication task is constructed.

6. The method for collaborative scheduling and target search of unmanned aerial vehicles (UAVs) for emergency rescue according to claim 5, characterized in that, Also includes: Based on the distance between the UAV's current location and the located ground node, determine the channel quality between the UAV and the located ground node; Based on the current location of the UAV, determine the UAV's perception range and search progress for the unlocated ground nodes; The sub-optimization problem of the communication task is adjusted using the channel quality; the sub-optimization problem of the target perception task is adjusted using the sensing range and search progress.

7. The method for collaborative scheduling and target search of unmanned aerial vehicles (UAVs) for emergency rescue according to claim 1, characterized in that, Based on the runtime constraint, the optimal flight strategy is obtained by iteratively solving each of the sub-optimization problems using the UAV operation model, specifically as follows: Using the particle swarm optimization algorithm, the optimal UAV position is searched through the UAV operation model, and the sub-optimization problem of the target perception task is solved iteratively to obtain the optimal UAV flight trajectory. The Lagrange dual decomposition method is used to decouple the sub-optimization problems of the communication task and obtain the optimal resource allocation strategy. Based on the optimal UAV flight trajectory and the optimal resource allocation strategy, construct the optimal flight strategy.

8. The method for collaborative scheduling and target search of unmanned aerial vehicles (UAVs) for emergency rescue according to claim 7, characterized in that, The process of searching for the optimal UAV position using the UAV operation model and iteratively solving the sub-optimization problem of the target perception task to obtain the optimal UAV flight trajectory is as follows: When the UAV receives the positioning signal transmitted by the ground node, it triggers the perception mode; in the perception mode, the UAV is controlled using a preset flight trajectory. When the UAV does not receive the positioning signal transmitted by the ground node, based on the running time constraint, the current optimal UAV position is searched through the UAV running model in each iteration until the preset maximum number of iterations is met or the optimal solution of the particle swarm has been found, and the final optimal UAV position is output to obtain the optimal UAV flight trajectory.

9. The method for collaborative scheduling and target search of unmanned aerial vehicles (UAVs) for emergency rescue according to claim 7, characterized in that, The Lagrange dual decomposition method is used to decouple the sub-optimization problems of the communication task, thereby obtaining the optimal resource allocation strategy, specifically as follows: By using the Lagrange dual decomposition method, the sub-optimization problem of the communication task is decoupled into several dual sub-problems, each of which corresponds to a communication resource scheduling problem. The communication resource scheduling problem includes the data upload scheduling of the ground node, the computing resource scheduling of the UAV, and the transmission power allocation of the ground node. Based on the aforementioned runtime constraints, each dual subproblem is solved to obtain the optimal solution for each dual subproblem; By integrating the optimal solutions to each dual subproblem, an optimal resource allocation strategy is constructed.

10. The method for collaborative scheduling and target search of unmanned aerial vehicles (UAVs) for emergency rescue according to claim 1, characterized in that, The use of optimal flight strategies to adjust the drone's flight trajectory and schedule its communication resources in real time specifically includes: Using the optimal flight strategy, the flight trajectory of the UAV is adjusted in real time based on the communication between the UAV and the ground node; The optimal flight strategy is used to schedule the communication resources of the UAV by adjusting the transmission power and data upload rate of the ground node and adjusting the computing resource scheduling of the UAV.

Images