An unmanned aerial vehicle control method for post-earthquake search and rescue and related equipment

Abstract

This application relates to the field of unmanned aerial vehicle (UAV) control technology, and provides a UAV control method and related equipment for post-earthquake search and rescue. The method includes: constructing discrete kinematic equations for ground search and rescue units based on ground motion data, and constructing state equations for target UAVs based on UAV motion data; constructing a communication objective function based on the discrete kinematic equations of all ground search and rescue units, the state equations of all target UAVs, and the physical property data of all rubble obstacles; constructing a task correlation matrix and control force expressions; solving the communication objective function based on the task correlation matrix and all control force expressions to obtain the final task correlation matrix and the final control force for each target UAV; and controlling each target UAV based on the final task correlation matrix and the final control force for each target UAV. The method of this application can improve the reliability of UAV control for post-earthquake search and rescue.

Description

Technical Field

[0001] This application relates to the field of unmanned aerial vehicle (UAV) control technology, and in particular to a UAV control method and related equipment for post-earthquake search and rescue. Background Technology

[0002] In recent years, traditional ground communication infrastructure has often suffered devastating damage in the face of sudden natural disasters (especially in the complex terrain of post-earthquake ruins). Against this backdrop of emergency rescue, unmanned aerial vehicles (UAVs), with their high mobility in three-dimensional space and on-demand deployment capabilities, have become a key force in building air-ground collaborative networks and executing wide-area search and rescue missions. Compared to traditional land-based networks constrained by fixed infrastructure, UAV collaborative networks can quickly overcome geographical and physical barriers, providing highly flexible communication and life detection coverage. However, the extreme non-line-of-sight (NLOS) obstruction and dynamic channel fading caused by ruin environments, coupled with the stringent limitations of airborne resources, make traditional "bit-level" transmission prone to errors in critical rescue information, leading to application bottlenecks. To break this deadlock, the academic community has begun to explore the introduction of new paradigms such as semantic communication into emergency search and rescue systems, and has conducted extensive forward-looking research on resource scheduling, trajectory planning, and communication mechanisms of air-ground networks. Based on different optimization objectives, current representative research results mainly focus on the following three aspects: 1) Research on minimizing system latency For time-sensitive, computationally intensive, or emergency rescue missions, reducing end-to-end data processing and transmission time is a primary concern in academia. Researchers often alleviate latency pressure through joint control of computational offloading and flight trajectory. For example, some studies have designed a multi-drone collaborative architecture that integrates task prioritization mechanisms, using deep reinforcement learning to simultaneously optimize computational allocation and UAV flight paths, significantly reducing the system's average task execution time. Other studies have proposed a latency-aware scheme based on time scheduling and 3D trajectory coordination for hybrid airborne base station scenarios, ensuring rapid delivery of high-real-time tasks. Furthermore, in disaster area building 3D reconstruction tasks, some research has effectively shortened the time spent by UAVs reconnaissance and sweeping of target areas by solving multi-viewpoint connectivity and dynamic constraints.

[0003] 2) Research on minimizing system energy consumption Given the common battery capacity bottleneck of micro-drones, extending the network uptime and operational radius of drone swarms directly depends on improving energy efficiency. In emergency network construction, some research has introduced particle swarm optimization (PSO) algorithms for integrated communication, navigation, and sensing design, achieving optimal matching of node roles and effectively curbing energy consumption during initial deployment. For large-scale maritime search and rescue, some research has constructed a path search mechanism integrating Gaussian mixture models and multi-objective optimization, eliminating unnecessary flight energy consumption while ensuring coverage of key sea areas. In low-altitude IoT services, some research has proposed a green transmission strategy that coordinates High Altitude Platforms (HAPs) and drones, maximizing the energy efficiency of the global network; while other research has designed adaptive allocation models to alleviate power starvation of ground nodes, addressing the dual needs of data collection and energy transmission. Simultaneously, some research has leveraged the reasoning capabilities of Large Language Models (LLMs) to achieve highly economical low-energy trajectory planning while ensuring compliance and safety in low-altitude flight.

[0004] 3) System performance optimization research Beyond simple time and power consumption considerations, recent research has increasingly focused on comprehensively upgrading network performance, encompassing dimensions such as data freshness (AoI, Age of Information), spatial exploration capabilities, and communication security. Some studies have constructed a multi-objective fusion planning framework integrating sensing, communication, and objects, maximizing the overall benefits of collaborative disaster relief through temporal graph convolutional networks. To improve data timeliness, some research has employed multi-agent proximal policy optimization algorithms to finely adjust swarm trajectories, successfully reducing the average weighted information age of sensor networks. In exploring unknown physical spaces, some research has proposed the Exploration Rapidly-exploring RandomTree (ERRT) algorithm, achieving an excellent balance between maximizing information gain and controlling movement costs; other research has used distance attention mechanisms to improve the robustness of multi-aircraft navigation and collision avoidance in dense obstacle courses. Regarding network security and coverage, some studies have utilized large model knowledge distillation to enhance the adaptive capabilities of multi-hop dynamic networking; others have analyzed the secure transmission capacity of air-to-ground networks from the physical layer; and still others have used LLM as an environmental feedback parsing engine to significantly optimize the efficiency of secure data collection in environments with eavesdropping threats.

[0005] However, the above methods have drawbacks such as slow computational convergence and weak cross-scenario adaptability when facing high-frequency dynamic changes in the work site. They are difficult to balance the real-time planning requirements of extremely low latency and high reliability, resulting in low reliability of UAV control for post-earthquake search and rescue. Summary of the Invention

[0006] This application provides a method and related equipment for controlling unmanned aerial vehicles (UAVs) in post-earthquake search and rescue, which can solve the problem of low reliability in UAV control for post-earthquake search and rescue.

[0007] In a first aspect, embodiments of this application provide a drone control method for post-earthquake search and rescue, the drone control method comprising: Acquire physical property data of multiple rubble obstacles in the target post-earthquake area, ground movement data of multiple ground search and rescue units, and drone movement data of multiple target drones; Based on the ground motion data of each ground search and rescue unit, the discrete kinematic equations of each ground search and rescue unit are constructed, and based on the drone motion data of each target drone, the state equations of each target drone are constructed. Based on the discrete kinematic equations of all ground search and rescue units, the state equations of all target UAVs, and the physical property data of all rubble obstacles, a communication objective function is constructed; the communication objective function is used to describe the relationship between ground search and rescue units and target UAVs. Construct a task association matrix between ground search and rescue units and target UAVs, and construct control force expressions for all target UAVs; the elements in the task association matrix represent the association relationships between ground search and rescue units and target UAVs. The communication objective function is solved based on the task association matrix and all control force expressions to obtain the final task association matrix and the final control force of each target UAV. Each target UAV is controlled based on the final mission correlation matrix and the final control capability of each target UAV.

[0008] Secondly, embodiments of this application provide a drone control device for post-earthquake search and rescue, comprising: The acquisition module is used to acquire physical attribute data of multiple rubble obstacles in the target post-earthquake area, ground movement data of multiple ground search and rescue units, and drone movement data of multiple target drones; The first construction module is used to construct the discrete kinematic equations of each ground search and rescue unit based on the ground motion data of each ground search and rescue unit, and to construct the state equations of each target UAV based on the UAV motion data of each target UAV. The second construction module is used to construct a communication objective function based on the discrete kinematic equations of all ground search and rescue units, the state equations of all target UAVs, and the physical property data of all rubble obstacles; the communication objective function is used to describe the relationship between ground search and rescue units and target UAVs; The construction module is used to construct the task association matrix between ground search and rescue units and target UAVs, and to construct the control force expressions for all target UAVs; the elements in the task association matrix are the association relationships between ground search and rescue units and target UAVs. The solver module is used to solve the communication objective function based on the task association matrix and all control force expressions, so as to obtain the final task association matrix and the final control force of each target UAV. The control module is used to control each target UAV based on the final task association matrix and the final control capability of each target UAV.

[0009] Thirdly, embodiments of this application provide a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the aforementioned UAV control method for post-earthquake search and rescue.

[0010] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned unmanned aerial vehicle (UAV) control method for post-earthquake search and rescue.

[0011] The above-mentioned solution in this application has the following beneficial effects: In the embodiments of this application, discrete kinematic equations for each ground search and rescue unit are constructed based on ground motion data of each unit, and state equations for each target UAV are constructed based on UAV motion data of each target UAV. Then, a communication objective function is constructed based on the discrete kinematic equations of all ground search and rescue units, the state equations of all target UAVs, and the physical property data of all rubble obstacles. Next, a task association matrix between the ground search and rescue units and the target UAVs is constructed, and control force expressions for all target UAVs are constructed. The communication objective function is then solved based on the task association matrix and all control force expressions to obtain the final task association matrix and the final control force of each target UAV. Finally, each target UAV is controlled based on the final task association matrix and the final control force of each target UAV. Specifically, a communication objective function is constructed based on data from the rubble and obstacles in the target post-earthquake area, data from ground search and rescue units, and data from the target UAV. This ensures that the communication objective function fully considers the relevant data of each participant in the target post-earthquake area, improving the practicality and accuracy of the communication objective function. The solution objective is defined as the task-related matrix and the control force of the UAV, thereby decomposing the solution objective, improving the solution speed of the communication objective function, meeting the requirements for low-latency real-time control of the UAV, and thus improving the reliability of UAV control for post-earthquake search and rescue.

[0012] Other beneficial effects of this application will be described in detail in the following detailed description section. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art 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 based on these drawings without creative effort.

[0014] Figure 1 A flowchart of an unmanned aerial vehicle (UAV) control method for post-earthquake search and rescue provided as an embodiment of this application; Figure 2 A simulation diagram illustrating UAV control for post-earthquake search and rescue, provided as an embodiment of this application; Figure 3 A schematic diagram of the structure of an unmanned aerial vehicle (UAV) control device for post-earthquake search and rescue provided in an embodiment of this application; Figure 4 This is a schematic diagram of the structure of a terminal device provided in an embodiment of this application. Detailed Implementation

[0015] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0016] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0017] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0018] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0019] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0020] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0021] To address the low reliability of existing UAV control methods for post-earthquake search and rescue, this application provides a UAV control method for post-earthquake search and rescue. This method constructs discrete kinematic equations for each ground search and rescue unit based on its ground motion data, and constructs state equations for each target UAV based on its motion data. Then, based on the discrete kinematic equations of all ground search and rescue units, the state equations of all target UAVs, and the physical property data of all rubble obstacles, a communication objective function is constructed. Next, a task correlation matrix between the ground search and rescue units and the target UAVs is constructed, along with control force expressions for all target UAVs. The communication objective function is then solved based on the task correlation matrix and all control force expressions to obtain the final task correlation matrix and the final control force for each target UAV. Finally, each target UAV is controlled based on the final task correlation matrix and the final control force of each target UAV. Specifically, a communication objective function is constructed based on data from the rubble and obstacles in the target post-earthquake area, data from ground search and rescue units, and data from the target UAV. This ensures that the communication objective function fully considers the relevant data of each participant in the target post-earthquake area, improving the practicality and accuracy of the communication objective function. The solution objective is defined as the task-related matrix and the control force of the UAV, thereby decomposing the solution objective, improving the solution speed of the communication objective function, meeting the requirements for low-latency real-time control of the UAV, and thus improving the reliability of UAV control for post-earthquake search and rescue.

[0022] The following is an exemplary description of the UAV control method for post-earthquake search and rescue provided in this application.

[0023] like Figure 1As shown, the UAV control method for post-earthquake search and rescue provided in this application includes the following steps: Step 11: Obtain physical property data of multiple rubble obstacles in the target post-earthquake area, ground movement data of multiple ground search and rescue units, and drone movement data of multiple target drones.

[0024] The aforementioned target post-earthquake area refers to the area requiring post-earthquake search and rescue, such as an earthquake-stricken area. The aforementioned ground search and rescue units refer to ground-based units participating in the search and rescue mission (such as search dogs, which carry cameras and other equipment to provide real-time footage to rescue personnel). The aforementioned target drone refers to an aerial drone performing rescue missions, providing communication to the ground-based search and rescue units. The aforementioned rubble obstacles refer to collapsed buildings and other ruins. Physical attribute data includes the apex, coordinates, and height of the rubble obstacles. The aforementioned ground motion data includes the position and speed of ground search and rescue units at multiple historical moments. The aforementioned drone motion data includes the position, speed, and battery level of the target drone at multiple historical moments. The last historical moment is the current moment.

[0025] In some embodiments of this application, physical property data of rubble obstacles can be obtained through devices such as positioning devices and rangefinders, and ground movement data of ground search and rescue units and drone movement data of target drones can be obtained through devices such as sensors.

[0026] For example, random distribution within a region A set of static ruins and obstacles is represented as ,ruins , Described by a set of geometric property parameters: ; in, Represents ruins The projection coordinates of the geometric center on the horizontal plane, This represents the set of vertices of the polygon that encloses the irregular outline of the ruins. A closed polygonal region formed by connecting all vertices in sequence. This represents the height of the accumulated ruins. Combined with the vertex set. With height The ruins constitute a multi-faceted prism obstruction in three-dimensional space, affecting the existence of line-of-sight communication links.

[0027] At the center of the ruins With the origin as the starting point, the first... The polar coordinate radius of each vertex Following a random distribution, forming a sawtooth boundary: ; in, , The average baseline radius of the ruins, The radius perturbation factor follows the interval... Uniform distribution on It characterizes the degree of jaggedness at the edge of the ruins and establishes the boundary of the restricted area for physical movement.

[0028] Distributed in the gaps between the ruins Search and rescue dogs wearing only communication modules are represented as a set. Search and rescue dogs exhibit biological movement characteristics of "inertial maintenance" (continuous direction over a short period) and "random exploration" (intermittent turning). Their trajectories are characterized using a restricted probability orientation model. For any search and rescue dog... At discrete time The state is defined as: ; in, This represents the environmental semantic source data (RGB image) collected by search and rescue dogs. The number of channels, image height, and image width are mapped to channel symbols for transmission via Deep JSCC. For search and rescue dogs Location coordinates, This is the normalized motion direction vector.

[0029] The air rescue cluster is composed of Composed of rotary-wing unmanned aerial vehicles, the set is represented as For any drone , its in The instantaneous state at time t is defined as a triple: ; in, and Indicates drone Position and velocity, To normalize the remaining battery power, an adaptive gain weight for semantic communication is determined.

[0030] Step 12: Based on the ground motion data of each ground search and rescue unit, construct the discrete kinematic equations of each ground search and rescue unit, and based on the drone motion data of each target drone, construct the state equations of each target drone.

[0031] The discrete kinematic equations described above are used to describe the motion of ground search and rescue units, while the state equations described above are used to describe the motion and battery status of the target UAV.

[0032] Specifically, the discrete kinematic equations are: ; ; in, Indicates the first Ground search and rescue units are at all times Location coordinates, Indicates the first Ground search and rescue units are at all times Candidate position coordinates, , This represents the set of numbers for ground search and rescue units. Indicates the safety margin. Indicates the boundary length of the target post-earthquake region. Represents the first in the ground motion data Ground search and rescue units are at all times Location coordinates, This indicates the average movement speed of ground search and rescue units. Indicates time interval, Indicates the first Ground search and rescue units are at all times Direction vector: ; in, Indicates the first Ground search and rescue units are at all times directional vector, This represents the Gaussian perturbation vector. , Let represent the covariance matrix of a Gaussian distribution, where the noise in each dimension is independent and the variance is 1. Represents a uniformly distributed random vector. , Indicates the turning probability threshold. .

[0033] The state equation is: ; ; in, Indicates the first The target drone at time speed, , This represents the set of IDs for the target drone. Indicating the first [unmanned aerial vehicle] motion data The target drone at time speed, Indicates the air resistance damping coefficient. , Indicates the first The target drone at time Control, Indicates the first The target drone at time Location coordinates, Indicating the first [unmanned aerial vehicle] motion data The target drone at time Location coordinates, Indicates time interval, Indicates the first The target drone at time The amount of electricity, Indicating the first [unmanned aerial vehicle] motion data The target drone at time The amount of electricity, This indicates the power dissipation of the bottom rotor. This represents the instantaneous additional mechanical power of the target drone. The battery internal resistance decay coefficient represents the rate of decrease in battery power, which increases exponentially as the remaining charge decreases. This indicates the total energy of the battery in Joules.

[0034] It should be noted that, in the underlying flight control, to eliminate trajectory oscillations and achieve dynamic obstacle avoidance, this method decouples the UAV thrust vector into horizontal damping control and vertical smart fly-over control. In the horizontal direction, to suppress high-frequency oscillations caused by the potential field force, a velocity-based proportional-derivative (PD) braking damping force is introduced. In the vertical direction, the system introduces a realistic gravity field and dynamically sets the target cruising altitude based on the height of the ruins ahead. Vertical upward rotor thrust Generated by the closed-loop PD controller: ; By vectoring the horizontal potential force with the vertical rotor thrust, the UAV can adaptively choose to bypass or ascend to fly over obstacles.

[0035] The above The power dissipation of the bottom rotor is taken into account for the induced power. The instantaneous additional mechanical power of the drone consists of two parts: climb penalty and high-speed cruise loss. Its calculation logic is as follows: When vertical velocity At that time, a surge penalty term based on the work done against gravity is introduced. ,in This represents the ramp-up power gain coefficient. It is a vertical thrust.

[0036] When the horizontal speed is too high, a nonlinear air resistance dissipation term is introduced. ,in This is the air resistance loss coefficient. This is the horizontal velocity vector. This full-vector hard-core energy consumption model accurately quantifies the energy cost during the three-dimensional maneuver, providing a physically significant feedback base for subsequent adaptive potential field gain control.

[0037] Step 13: Construct a communication objective function based on the discrete kinematic equations of all ground search and rescue units, the state equations of all target UAVs, and the physical property data of all rubble obstacles.

[0038] The aforementioned communication objective function is used to describe the relationship between ground search and rescue units and the target UAV.

[0039] Specifically, the communication objective function is: ; ; ; ; ; ; ; ; ; in, Indicates the optimization objective. Indicates the last moment. Indicates at time The following association matching decision, For indicator functions, For semantic performance mapping function, For the first The target drone and the first Ground search and rescue units are at all times The instantaneous physical signal-to-noise ratio is below. This is the semantic confidence threshold. Indicates the first constraint. Indicates the second constraint. Indicates the maximum flight speed. Indicates a third constraint. Indicates the first The target drone at time Location coordinates, Indicates a safe distance. For the fourth constraint, Represents the horizontal axis coordinate. Represents the vertical axis coordinate. Represents the vertical axis coordinates. Indicates obstacles caused by ruins The closed polygon region is formed by connecting the vertex data in the physical attribute data in sequence. Indicates ruins and obstacles The height of the rubble pile in the physical attribute data, Indicates the number of obstacles in the ruins. For the fifth constraint, Indicates the boundary of the feasible flight area. For the sixth constraint, This is the seventh constraint. This indicates the maximum number of service nodes. This indicates the eighth constraint. This indicates the safe return-to-home battery level threshold.

[0040] Constraint C1 provides a discrete kinematic model for UAV position updates, establishing the mapping relationship between current velocity and next-moment position; Constraint C2 specifies the maximum flight speed limit for UAVs, satisfying the physical maneuver constraints of rotorcraft; Constraint C3 specifies the safe collision avoidance distance between UAVs in a swarm, preventing mid-air collisions during dense operations; Constraint C4 defines a three-dimensional no-fly zone based on the geometric features of the ruins, prohibiting UAVs from entering the columnar space enclosed by the horizontal contour and pile height of the ruins, constituting a non-convex hard constraint condition in trajectory planning; Constraint C5 limits the boundaries of the feasible flight area for UAVs; Constraint C6 specifies the uniqueness of task allocation, meaning that at any given time, a ground search and rescue dog can only be semantically covered by a maximum of one UAV; Constraint C7 sets the maximum upper limit for the number of service nodes for a single UAV, achieving load balancing at the swarm level; Constraint C8 specifies the energy constraints for UAVs, requiring their normalized remaining battery power to be higher than the safe return threshold during mission execution.

[0041] It should be noted that the physical signal-to-noise ratio at the receiving end... Determined by transmit power, path loss, and shadowing fading: ; in, The transmission power (dBm) of the search and rescue dog. This is the path loss index. This represents the power spectral density of Gaussian white noise. This is a line-of-sight occlusion indication function. When the link between the drone and the search dog passes through the rubble... hour, Penetration loss is .

[0042] Environmental semantic source RGB image data collected by search and rescue dogs, represented by the following dimensions: To characterize semantic communication performance at the system level, classification accuracy is used as a semantic usability metric to measure whether the receiver can correctly understand the transmitted environmental semantic information (such as target category or environmental state). Compared to traditional bit-level communication metrics, it can more directly reflect the actual value of information in rescue missions.

[0043] To establish the link between physical channels and semantic understanding, a Deep Joint Source-Channel Coding (Deep JSCC) network was pre-trained based on the CIFAR-10 dataset. Through extensive offline simulations, the mapping relationship between signal-to-noise ratio (SNR) and semantic recognition accuracy was fitted to a sigmoid function. ; Indicates drone For search and rescue dogs Confidence in semantic understanding of sent information and These are the fitting parameters.

[0044] The model shows that in the low signal-to-noise ratio region, semantic accuracy increases rapidly with the improvement of SNR, while in the high signal-to-noise ratio region, the marginal effect decreases.

[0045] Based on semantic performance mapping, a valid semantic connection at the system layer is defined. This definition clearly distinguishes between two different levels of communication states: "physical connectivity" and "semantic validity." Even at the physical level, if the communication link is not completely interrupted, but the semantic confidence level is below a preset threshold, the transmitted information may still not provide reliable support for rescue decisions and therefore should not be considered valid system communication. By introducing a semantic confidence judgment mechanism, the system can automatically filter out low-value or high-uncertainty information transmissions at the communication and task levels, thereby improving the credibility and validity of the transmitted information.

[0046] Construct the semantic communication topology graph of the system Vertex set edge set The connection relationship in the is defined as: ; in, The semantic confidence threshold is when At that time, the drone was determined Successfully established a system for search and rescue dogs Effective semantic coverage. The topology map changes dynamically over time, reflecting the impact of UAV trajectory, environmental occlusion, and channel conditions on semantic communication coverage. The system's optimization objective is to maximize the number of effective semantic connections in the network and improve overall rescue information acquisition efficiency by jointly optimizing the association between UAV trajectory and task, under energy-constrained and complex environmental conditions.

[0047] Based on the established 3D ruin environment model and air-ground collaborative semantic communication framework, the multi-agent collaborative problem in post-earthquake rescue is formally expressed as a standard mathematical optimization model. Under the constraints of limited airborne energy and complex unstructured environment, the effective semantic connection number is maximized by jointly optimizing the task association matching matrix and flight trajectory of the UAV swarm. The decision variables are defined as follows: Task association matrix: Represents discrete task assignment variables.

[0048] in, : Indicates the time. drones Assigned to be responsible for receiving search and rescue dogs Semantic data, and track its trajectory. : indicates no association.

[0049] Flight trajectory collection: This represents the sequence of three-dimensional spatial coordinates for all drones.

[0050] The system's objective function is defined as the sum of all successfully established valid semantic links across the entire time domain. Unlike traditional communication that aims to maximize throughput, the objective function introduces a semantic confidence judgment mechanism determined by the Deep JSCC performance curve. Only when the UAV logically associates with the target and physically moves to a position where the signal-to-noise ratio meets the preset semantic threshold is the link deemed "valid," contributing precise survivor coordinates to the command center.

[0051] Step 14: Construct the mission association matrix between the ground search and rescue unit and the target UAV, and construct the control force expressions for all target UAVs.

[0052] The elements in the above task correlation matrix represent the relationships between ground search and rescue units and the target UAV. The above control force expression describes the control force of the target UAV.

[0053] In some embodiments of this application, the steps of constructing the task association matrix between the ground search and rescue unit and the target UAV, and constructing the control force expressions for all target UAVs, include: The first step, based on the principle of minimum potential energy, is to determine the optimal associated drone for each ground search and rescue unit from all target drones.

[0054] Specifically, through the formula: ; Determine the optimal associated drones for ground search and rescue units .

[0055] in, This represents the load balancing weighting factor. This represents the parameter indicating the potential energy growth rate. Indicates the first The target drone at time Number of ground search and rescue units providing services.

[0056] It should be noted that the definition For drones, Assemble the ground search and rescue dogs. For any drone... With search and rescue dogs Its generalized matching potential energy The definition is as follows: ; The first term characterizes the physical reachability and large-scale channel fading potential. The second term... For drones The cumulative number of search and rescue dogs in service at the previous decision-making moment. This is the load balancing weighting factor. The load penalty function is nonlinear. To build a strong repulsion barrier under high load conditions on the UAV, it is designed to grow exponentially. .

[0057] The second step is to assign a value of 1 to the association matching decision between the ground search and rescue unit and the optimal associated UAV for each ground search and rescue unit, and assign a value of 0 to the association matching decision between the ground search and rescue unit and other target UAVs.

[0058] That is, it is assumed that there is a correlation between the ground search and rescue unit and the optimal associated UAV, but no correlation between them and other target UAVs.

[0059] The third step is to integrate all the association matching decisions into a matrix to obtain the task association matrix.

[0060] The fourth step is to define the fundamental potential field force for each target drone.

[0061] Specifically, the fundamental potential field force is: ; in, Represents the force in the fundamental potential field. This represents gradient operation. This represents the Euclidean distance potential function. Represents the repulsive potential energy function. Represents the inter-machine collision avoidance potential energy function: ; ; ; in, Indicates the first The Euclidean distance potential energy of the target drone The energy gain function, For semantic transmission error term, Indicates the first The repulsive potential energy of the target drone This represents the gain coefficient of the repulsive potential field. Indicates the first The Euclidean distance between the target drone and the nearest surface of the rubble obstacle. Represents a collection of ruins and obstacles. Indicates the radius of influence of the ruins or obstacles. Indicates the first The target drone and its current distance from the surface of the rubble obstacle. Indicates the first Inter-drone collision avoidance potential energy of a target drone This represents the gain coefficient of the inter-machine collision avoidance potential field. Indicates the first The set of neighboring target drones of a target drone. Indicates the first The target drone and the first The Euclidean distance between the target drones This indicates the safe collision distance threshold between drones. When the distance between two drones is less than this value, a strong repulsive force is generated to avoid a collision.

[0062] It should be noted that, The semantic transmission error term, combined with the nonlinear mapping characteristics of Deep JSCC, characterizes the semantic communication quality at the current location. This is an energy-adaptive gain function. Based on the negative gradient method ( ), which can lead to the conclusion that it applies to drones Total semantic gravity vector: ; The gravitational force always points towards the search and rescue dog target, and its amplitude is determined by both the physical distance and the semantic quality adaptive gain. Unlike the traditional potential field method, the semantic gain coefficient... This gives the physical control layer the ability to perceive the communication quality of the application layer: at the same distance, if a poor channel environment leads to increased semantic transmission errors, This will significantly enhance the "semantic pull" that drives the drone closer to the target to improve the signal-to-noise ratio; conversely, if the semantic connection is clear enough, the gravitational pull will remain at a basic level, preventing the drone from making unnecessary excessive maneuvers near the target.

[0063] Energy gain function Designed as a variant of the Sigmoid function to achieve task urgency adjustment under low power conditions: ; in, This refers to the total battery capacity of the drone. To prevent numerical stability constants where the denominator is zero. When the remaining electricity... When sufficient, the system maintains the base gain. To achieve smooth cruise; when the battery is depleted and approaches the critical value, the gain increases nonlinearly, giving the drone greater acceleration authority at the dynamic level, ensuring that high-value targets are prioritized for acquisition before the mission window closes.

[0064] The fifth step is to construct the control force expression for each target UAV based on the force field of each basic potential field.

[0065] Specifically, the expression for control force is: ; in, Indicates the controllability of the target drone. This represents the surface normal vector of the rubble obstacle. This indicates the transpose operation. Indicates the tangential sliding guidance coefficient. Indicates perpendicular to The tangential unit vector.

[0066] Step 15: Solve the communication objective function based on the task association matrix and all control force expressions to obtain the final task association matrix and the final control force of each target UAV.

[0067] The elements in the aforementioned final task correlation matrix represent the correlation between the ground search and rescue unit and the target UAV at the next moment from the current moment, and the aforementioned final control force represents the control force of the target UAV at the next moment from the current moment.

[0068] For example, tools such as MATLAB fmincon can be used to solve the communication objective function. With the goal of maximizing the communication objective function, the task correlation matrix and the control force of the target UAV can be iteratively updated, and the final task correlation matrix and the final control force of each target UAV can be output.

[0069] Step 16: Control each target UAV according to the final task association matrix and the final control capability of each target UAV.

[0070] Specifically, in the next moment after the current moment, the final control force will be input into the control system of the target UAV and the target UAV will be controlled to move. At the same time, according to the final mission association matrix, communication connections will be established between the ground search and rescue units that are related to the target UAV.

[0071] For example, after controlling the target drone through the above process, consider the drone... The motion follows a second-order dynamical model, and its extended state vector is defined as follows: ,in and Let represent the position and velocity vectors, respectively. To prove the asymptotic stability of the system, the following composite energy function is constructed as a Lyapunov candidate function. : ; in, For the quality of drones. The semantic gravitational potential field is located at the target position of the search and rescue dog. It is continuously differentiable and attains a global minimum. ; As the potential energy for collision avoidance between drones, when the distance between drones approaches the safety threshold. Timely satisfaction To ensure cluster security; The repulsive potential field of the obstacle is located at the boundary of the obstacle. Satisfying Obviously, It satisfies positive definiteness and is radially unbounded, i.e. This ensures the boundedness of the system state.

[0072] right Request regarding time The total derivative of the system is obtained and expanded along the system trajectory: ; Substituting into the second-order dynamic equation .in, The control force output by the algorithm, This is a positive definite symmetric damping matrix, representing air resistance or active velocity damping in the underlying flight control system. Simplified, we get: ; Scenario 1: Free Space Mode When the drone is not within the influence range of obstacles or the region dominated by repulsive fields, the control force is generated entirely by the negative gradient of the potential field, i.e. Substituting into the above equation, the potential field gradient term and the control force term cancel each other out: ,because It is a positive definite matrix, therefore It is always true (semi-negative).

[0073] This indicates that the system is continuously dissipating energy. Considering the target of the search and rescue dogs... The continuous motion characteristics cause the potential field minimum point to drift over time, and the convergence tolerance introduced in engineering implementation... (Set to 1.0m in the simulation), the system state will no longer converge to It is not an absolutely static point, but rather eventually converges and is constrained to a point where... Within the bounded region centered on the term. Outside this region, the damping dissipation term... The system dynamics are always controlled, ensuring robust tracking of the trajectory for moving targets, which is completely consistent with the simulation results.

[0074] Scenario 2: Sliding Mode Avoidance Correction: When the drone approaches the ruins and the expected force is directed towards the interior of the obstacle (i.e. When this occurs, the algorithm activates sliding mode correction. Define the tangent plane projection operator. The corrected control force is decomposed into: ; At this point, the energy derivative is: ; ; ; The above equation clearly reveals the energy competition mechanism in the sliding mode obstacle avoidance process: First item : This is damped dissipation, and its amplitude is proportional to the square of the velocity. ), constantly remove system energy; The second item: the tangential guiding energy actively injected into the algorithm, the magnitude of which is linearly related to the velocity. ).

[0075] Although It is not always negative, but because the growth rate of the quadratic dissipation term with increasing speed is much higher than that of the linear injection term, when the drone speed exceeds a certain threshold, the dissipation term will inevitably dominate the energy change (i.e., This characteristic ensures the Lagrange stability of the system during obstacle avoidance, meaning that the UAV's trajectory is always constrained within a bounded range and will not diverge due to the introduction of tangential forces.

[0076] The time and space complexity of an algorithm are key metrics for evaluating its deployment on resource-constrained airborne embedded platforms. The Sem-APF algorithm employs a hierarchical decoupled architecture, effectively avoiding the combinatorial explosion problem common in traditional centralized global optimization algorithms. The asymptotic time complexity of its single-step decision-making is analyzed below.

[0077] First, in each decision cycle, the system needs to construct a generalized cost matrix. For drones and Search and rescue dogs only, calculation is required. The process involves sub-Euclidean distance and load potential energy. It involves intensive matrix operations but has no recursion or backtracking. Its time complexity is O(n log n). .

[0078] Secondly, at the physical motion control layer, the computational overhead mainly comes from the superposition and updating of potential forces such as semantic attraction, inter-drone repulsion, and obstacle avoidance correction. For any single drone in the cluster, its state update requires traversing the associated set of search and rescue dogs (number of dogs). ), neighboring drones within communication range (number) ) and the set of obstacles in the environment (number) The semantic attraction, inter-drone repulsion, and sliding mode obstacle avoidance corrections are calculated separately. Therefore, the computational cost for a single UAV is... Extend to the entire system. The total computational complexity of a single step at the physical layer for a drone is... .

[0079] Based on the above two levels of analysis, in the total task time step Within this timeframe, the overall time complexity of the Sem-APF algorithm is O(n). This result demonstrates that the computational complexity of the algorithm increases exponentially with system size, exhibiting good scalability. In contrast, traditional methods for solving such mixed-integer nonlinear programming (MINLP) problems suffer from significant computational bottlenecks: the branch-and-bound method has a worst-case complexity of O(n log n). The population size increases exponentially; while the genetic algorithm, although a heuristic algorithm, typically requires a number of iterations to converge that are much larger than the product of the population size and the time step. This results in insufficient real-time performance.

[0080] In summary, the algorithm in this application successfully reduces the NP-hard joint optimization problem to a heuristic solution process with polynomial complexity.

[0081] In some embodiments of this application, a system as described above is constructed using Python. Figure 2 A simulation environment containing irregular obstacles (gray area) and a dynamically moving search and rescue dog (triangle) was used to verify the actual performance of the algorithm. In the figure, UAV0Path is the trajectory of UAV0 with an initial altitude H of 70.4m, UAV1Path is the trajectory of UAV1 with an initial altitude H of 87.7m, and UAV2Path is the trajectory of UAV2 with an initial altitude H of 82.5m. The horizontal axis X and vertical axis Y are the geographic coordinate axes from the top of the region. As shown in the figure, when facing extreme physical obstruction, the UAV (circle) relies on the semantic artificial potential field and sliding mode obstacle avoidance control strategy proposed in this application to not only achieve safe obstacle avoidance throughout the process, but also successfully overcome the fatal defect of traditional potential field methods that are prone to "local deadlock" at the edge of ruins, and complete smooth bypassing and breakthrough close to the edge of the obstacle. Finally, with extremely low solution latency, the UAV, in coordination with the trajectory of the dynamic search and rescue dog, ensured the continuous coverage of high-confidence semantic information, proving the high reliability and excellent dynamic tracking capability of the scheme in complex post-disaster harsh environments.

[0082] It is worth mentioning that a communication objective function is constructed based on data from the rubble and obstacles in the target post-earthquake area, data from ground search and rescue units, and data from the target UAV. This allows the communication objective function to fully consider the relevant data of each participant in the target post-earthquake area, improving the practicality and accuracy of the communication objective function. The solution objective is defined as the task-related matrix and the control force of the UAV, realizing the decomposition of the solution objective, improving the solution speed of the communication objective function, meeting the requirements for low-latency real-time control of the UAV, and thus improving the reliability of UAV control for post-earthquake search and rescue.

[0083] The following is an exemplary description of the unmanned aerial vehicle (UAV) control device for post-earthquake search and rescue provided in this application.

[0084] like Figure 3 As shown, this application provides a drone control device for post-earthquake search and rescue, the drone control device 300 for post-earthquake search and rescue includes: The acquisition module 301 is used to acquire physical attribute data of multiple ruins and obstacles in the target post-earthquake area, ground movement data of multiple ground search and rescue units, and drone movement data of multiple target drones; The first construction module 302 is used to construct the discrete kinematic equations of each ground search and rescue unit based on the ground motion data of each ground search and rescue unit, and to construct the state equations of each target UAV based on the UAV motion data of each target UAV. The second construction module 303 is used to construct a communication objective function based on the discrete kinematic equations of all ground search and rescue units, the state equations of all target UAVs, and the physical property data of all rubble obstacles; the communication objective function is used to describe the relationship between ground search and rescue units and target UAVs; Module 304 is used to construct the task association matrix between the ground search and rescue unit and the target UAV, and to construct the control force expressions for all target UAVs; the elements in the task association matrix are the association relationships between the ground search and rescue unit and the target UAV. The solver module 305 is used to solve the communication objective function based on the task association matrix and all control force expressions to obtain the final task association matrix and the final control force of each target UAV. The control module 306 is used to control each target UAV based on the final task association matrix and the final control force of each target UAV.

[0085] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0086] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0087] like Figure 4As shown, an embodiment of this application provides a terminal device, wherein the terminal device D10 of this embodiment includes: at least one processor D100 ( Figure 4 The diagram shows only one processor, a memory D101, and a computer program D102 stored in the memory D101 and executable on the at least one processor D100, wherein the processor D100 executes the computer program D102 to implement the steps in any of the above method embodiments.

[0088] Specifically, when the processor D100 executes the computer program D102, it constructs the discrete kinematic equations of each ground search and rescue unit based on the ground motion data of each unit, and constructs the state equations of each target UAV based on the UAV motion data of each target UAV. Then, based on the discrete kinematic equations of all ground search and rescue units, the state equations of all target UAVs, and the physical property data of all rubble obstacles, it constructs a communication objective function. Then, it constructs the task association matrix between the ground search and rescue units and the target UAVs, and constructs the control force expressions of all target UAVs. Then, it solves the communication objective function based on the task association matrix and all control force expressions to obtain the final task association matrix and the final control force of each target UAV. Finally, it controls each target UAV based on the final task association matrix and the final control force of each target UAV. Specifically, a communication objective function is constructed based on data from the rubble and obstacles in the target post-earthquake area, data from ground search and rescue units, and data from the target UAV. This ensures that the communication objective function fully considers the relevant data of each participant in the target post-earthquake area, improving the practicality and accuracy of the communication objective function. The solution objective is defined as the task-related matrix and the control force of the UAV, thereby decomposing the solution objective, improving the solution speed of the communication objective function, meeting the requirements for low-latency real-time control of the UAV, and thus improving the reliability of UAV control for post-earthquake search and rescue.

[0089] The processor D100 can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0090] In some embodiments, the memory D101 may be an internal storage unit of the terminal device D10, such as a hard disk or memory of the terminal device D10. In other embodiments, the memory D101 may be an external storage device of the terminal device D10, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the terminal device D10. Furthermore, the memory D101 may include both internal and external storage units of the terminal device D10. The memory D101 is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory D101 can also be used to temporarily store data that has been output or will be output.

[0091] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the various method embodiments above.

[0092] This application provides a computer program product that, when run on a terminal device, enables the terminal device to implement the steps described in the various method embodiments above.

[0093] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying the computer program code to a drone control method / terminal device for post-earthquake search and rescue, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0094] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0095] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0096] The above description is the preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this invention, and these improvements and modifications should also be considered within the scope of protection of this invention.

Claims

1. A method for controlling unmanned aerial vehicles (UAVs) for post-earthquake search and rescue, characterized in that, include: Acquire physical property data of multiple rubble obstacles in the target post-earthquake area, ground movement data of multiple ground search and rescue units, and drone movement data of multiple target drones; Based on the ground motion data of each ground search and rescue unit, a discrete kinematic equation is constructed for each ground search and rescue unit, and based on the drone motion data of each target drone, a state equation is constructed for each target drone. Based on the discrete kinematic equations of all ground search and rescue units, the state equations of all target UAVs, and the physical property data of all rubble obstacles, a communication objective function is constructed. The communication objective function is used to describe the association between the ground search and rescue unit and the target UAV; Construct a task association matrix between ground search and rescue units and target UAVs, and construct control force expressions for all target UAVs; The elements in the task association matrix represent the association relationships between ground search and rescue units and target drones; The communication objective function is solved based on the task association matrix and all control force expressions to obtain the final task association matrix and the final control force of each target UAV. Each target UAV is controlled based on the final task association matrix and the final control capability of each target UAV.

2. The UAV control method according to claim 1, characterized in that, The discrete kinematic equations are: in, Indicates the first Ground search and rescue units are at all times Location coordinates, Indicates the first Ground search and rescue units are at all times Candidate position coordinates, , This represents the set of numbers for ground search and rescue units. Indicates the safety margin. Indicates the boundary length of the target post-earthquake region. The first term in the ground motion data refers to... Ground search and rescue units are at all times Location coordinates, This indicates the average movement speed of ground search and rescue units. Indicates time interval, Indicates the first Ground search and rescue units are at all times Direction vector: in, Indicates the first Ground search and rescue units are at all times directional vector, This represents the Gaussian perturbation vector. , The covariance matrix of the Gaussian distribution is represented. Represents a uniformly distributed random vector. , Indicates the turning probability threshold. .

3. The UAV control method according to claim 2, characterized in that, The state equation is: in, Indicates the first The target drone at time speed, , This represents the set of IDs for the target drone. Indicating the first [unmanned aerial vehicle] motion data The target drone at time speed, Indicates the air resistance damping coefficient. , Indicates the first The target drone at time Control, Indicates the first The target drone at time Location coordinates, The first one mentioned in the drone motion data The target drone at time Location coordinates, Indicates time interval, Indicates the first The target drone at time The amount of electricity, Indicating the first [unmanned aerial vehicle] motion data The target drone at time The amount of electricity, This indicates the power dissipation of the bottom rotor. This represents the instantaneous additional mechanical power of the target drone. The battery internal resistance decay coefficient represents the rate of decrease in battery power, which increases exponentially as the remaining charge decreases. This indicates the total energy of the battery in Joules.

4. The UAV control method according to claim 3, characterized in that, The communication objective function is: in, Indicates the optimization objective. Indicates the last moment. Indicates at time The following association matching decision, For indicator functions, For semantic performance mapping function, For the first The target drone and the first Ground search and rescue units are at all times The instantaneous physical signal-to-noise ratio is below. This is the semantic confidence threshold. Indicates the first constraint. Indicates the second constraint. Indicates the maximum flight speed. Indicates a third constraint. Indicates the first The target drone at time Location coordinates, Indicates a safe distance. For the fourth constraint, Represents the horizontal axis coordinate. Represents the vertical axis coordinate. Represents the vertical axis coordinates. Indicates obstacles caused by ruins The closed polygon region is formed by connecting the vertex data in the physical attribute data in sequence. Indicates ruins and obstacles The height of the rubble pile in the physical attribute data, Indicates the number of obstacles in the ruins. For the fifth constraint, Indicates the boundary of the feasible flight area. For the sixth constraint, This is the seventh constraint. This indicates the maximum number of service nodes. This indicates the eighth constraint. This indicates the safe return-to-home battery level threshold.

5. The UAV control method according to claim 4, characterized in that, The construction of the mission association matrix between ground search and rescue units and target UAVs includes: Based on the principle of minimum potential energy, the optimal associated UAV for each ground search and rescue unit is determined from all target UAVs. For each ground search and rescue unit, the association matching decision between the ground search and rescue unit and the optimal associated UAV is assigned a value of 1, and the association matching decision between the ground search and rescue unit and other target UAVs is assigned a value of 0. All association matching decisions are integrated into a matrix to obtain the task association matrix.

6. The UAV control method according to claim 5, characterized in that, The method for determining the optimal associated drone for each ground search and rescue unit from all target drones based on the principle of minimum potential energy includes: Through the formula: Determine the optimal associated drones for ground search and rescue units ; in, This represents the load balancing weighting factor. This represents the parameter indicating the potential energy growth rate. Indicates the first The target drone at time Number of ground search and rescue units providing services.

7. The UAV control method according to claim 6, characterized in that, The construction of control force expressions for all target UAVs includes: Define the fundamental potential field force for each target UAV; The control force expression for each target UAV is constructed based on the force field of each fundamental potential field.

8. The UAV control method according to claim 7, characterized in that, The force in the fundamental potential field is: in, Represents the force in the fundamental potential field. This represents gradient operation. This represents the Euclidean distance potential function. Represents the repulsive potential energy function. Represents the inter-machine collision avoidance potential energy function: in, Indicates the first The Euclidean distance potential energy of the target drone The energy gain function, For semantic transmission error term, Indicates the first The repulsive potential energy of the target drone This represents the gain coefficient of the repulsive potential field. Indicates the first The Euclidean distance between the target drone and the nearest surface of the rubble obstacle. Represents a collection of ruins and obstacles. Indicates the radius of influence of the ruins or obstacles. Indicates the first The target drone and its current distance from the surface of the rubble obstacle. Indicates the first Inter-drone collision avoidance potential energy of a target drone This represents the gain coefficient of the inter-machine collision avoidance potential field. Indicates the first The set of neighboring target drones of a target drone. Indicates the first The target drone and the first The Euclidean distance between the target drones This indicates the safe collision distance threshold between drones; The expression for the control force is: in, Indicates the controllability of the target drone. This represents the surface normal vector of the rubble obstacle. This indicates the transpose operation. Indicates the tangential sliding guidance coefficient. Indicates perpendicular to The tangential unit vector.

9. A terminal 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 UAV control method for post-earthquake search and rescue as described in any one of claims 1 to 8.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the UAV control method for post-earthquake search and rescue as described in any one of claims 1 to 8.

Images