An unmanned aerial vehicle emergency scheduling method and system based on multi-machine nest cooperative relay

Abstract

The application provides an unmanned aerial vehicle emergency scheduling method and system of multi-machine nest cooperative relay, the method comprises the following steps: collecting multi-source sensor data including video stream, inputting the multi-source sensor data into an AI model for emergency scene type identification to obtain a corresponding type of emergency scene; a mapping relationship between corresponding departments and emergency strategies based on the emergency scene type is constructed and maintained in advance, and the corresponding emergency strategy is found based on the identified emergency scene type and the mapping relationship; a dynamic authorization pool of unmanned aerial vehicle nests of multiple departments is constructed and maintained in advance, the emergency strategy is input into a pre-trained large language model to perform task decomposition and output a multi-task list, and each task in the multi-task list is configured with a priority; a multi-unmanned aerial vehicle cooperative task conflict resolution mechanism is constructed, an unmanned aerial vehicle scheduling strategy for cooperative relay is generated according to the current unmanned aerial vehicle state, and the corresponding unmanned aerial vehicle is called from the corresponding dynamic authorization pool to execute the cooperative relay task.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) scheduling technology, and in particular to an emergency dispatching method and system for UAVs with multi-nest collaborative relay. Background Technology

[0002] Currently, the existing traditional UAV emergency dispatch technologies mainly include the following three types: (1) Single UAV task dispatch technology. This technology mainly involves operators manually selecting UAVs and task routes, and the system only provides route planning and flight monitoring functions. Typical solutions include UAV management platforms such as DJI Sikong, which support manual assignment of single UAVs to perform inspection or shooting tasks. (2) Multi-UAV static task allocation technology. Based on task requirements and UAV status, pre-planning allocation is carried out, such as batch allocation of multiple inspection tasks to multiple UAVs to optimize the total flight distance or task completion time. Some solutions introduce UAV nest (UAV airport) management to realize automatic UAV take-off, landing and charging, but task allocation is still mainly based on static planning. (3) General low-altitude flight management platform technology. Integrates functions such as airspace management, flight plan approval, and real-time monitoring to provide flight services for scenarios such as logistics and inspection. For example, some urban low-altitude management platforms support multi-source data access and flight plan optimization, but emergency dispatch still requires manual intervention.

[0003] Therefore, the existing drone emergency dispatch technologies mentioned above lack AI-driven automatic event recognition and departmental collaboration mechanisms, lack cross-departmental drone nesting authorization and dispatch mechanisms under government processes, and lack the ability to collaboratively resolve airspace conflicts among multiple departments and multiple drones. Summary of the Invention

[0004] One objective of this invention is to provide a method and system for emergency dispatching of drones using a multi-nest collaborative relay system. The method and system provide an AI scene recognition model based on multi-source data perception. This AI scene recognition model can quickly identify emergency scenarios. Specifically, this invention constructs a mapping relationship between departments and emergency strategies for each scenario. After the AI ​​scene recognition model identifies an emergency scenario, the mapping relationship enables rapid retrieval and activation of emergency strategies, thereby significantly reducing the response time of drone emergency dispatching and improving the drone's decision-making capabilities in the scenario. Under certain conditions, drones automatically execute cross-departmental collaborative relays. Therefore, this invention can solve the technical problem of uninterrupted cross-departmental drone monitoring after a sudden event in urban government emergency scenarios.

[0005] Another objective of this invention is to provide a method and system for emergency dispatching of drones using a multi-nest collaborative relay system. This method and system constructs a multi-departmental drone nest authorization pool. This invention dynamically maintains the drone nest authorization pool. Before emergency dispatch begins, an available nest in the authorization pool is automatically matched using a relay dispatching algorithm. After one-click confirmation, drone dispatching is uniformly executed for the corresponding authorized nest. Therefore, this invention's technical solution, while ensuring the compliance of cross-departmental drone flight dispatching, significantly improves the response efficiency of cross-departmental drone dispatching authorization through a normalized authorization pool mechanism, balancing compliance and flexibility, and also solving the problem of insufficient drone endurance.

[0006] Another objective of this invention is to provide a method and system for emergency dispatching of unmanned aerial vehicles (UAVs) in a multi-nest collaborative relay. This method and system constructs a monitoring layer, a coordination layer, and a delivery layer at different altitudes. A dynamic priority coordination strategy is built based on different altitude layers and event types. The altitude layers of the monitoring layer, coordination layer, and delivery layer are isolated from each other. Therefore, this invention can effectively improve the success rate of conflict resolution in multi-department, multi-UAV collaborative operations in the same domain, and can significantly reduce the possibility of collisions during UAV collaborative operations. Furthermore, this invention achieves a high degree of automation throughout the entire process, from event identification, department matching, contingency plan decomposition, nest scheduling, relay coordination to conflict resolution. Only one-click authorization confirmation from departmental on-duty personnel is required, significantly reducing the reliance on the experience of professional personnel for emergency dispatching.

[0007] To achieve at least one of the above-mentioned objectives, the present invention further provides a method for emergency dispatching of unmanned aerial vehicles (UAVs) through multi-nest cooperative relay, the method comprising: Collect multi-source sensor data, including video streams, and input the multi-source sensor data into an AI model to identify emergency scenario types and obtain the corresponding types of emergency scenarios. A mapping relationship between departments and emergency strategies based on emergency scenario types is pre-built and maintained, and the corresponding emergency strategies are found based on the identified emergency scenario types and mapping relationships; A dynamic authorization pool for drone nests of multiple departments is pre-built and maintained. The emergency strategy is input into a pre-trained large language model to decompose tasks and output a multi-task list. Priority is configured for each task in the multi-task list. A spatially isolated and layered multi-UAV collaborative task conflict resolution mechanism is constructed. Based on the current UAV status, a UAV scheduling strategy for collaborative relay is generated, and the corresponding UAV is called from the corresponding dynamic authorization pool to execute the collaborative relay task.

[0008] According to a preferred embodiment of the present invention, the emergency scenario type identification method includes: acquiring visible light images and infrared light images of the target area; inputting the visible light images into a target detection model for target identification to obtain target information; inputting the infrared light images into a thermal imaging detection model for thermal anomaly identification to obtain thermal anomaly information; performing fusion analysis based on the identified target information and thermal anomalies; and generating the emergency scenario type, subtype, and risk level within the target area.

[0009] According to another preferred embodiment of the present invention, the task decomposition method of the emergency strategy includes: converting the unstructured emergency strategy text found by mapping into a structured JSON format multi-task list by calling the structured extraction interface of a pre-trained large language model, wherein the task types of the multi-task list include main detection tasks, collaborative tasks and conditional tasks, wherein the priority of the main detection task is higher than that of the collaborative task, the priority of the collaborative task is higher than that of the conditional task, and the triggering conditions between different task types are configured.

[0010] According to another preferred embodiment of the present invention, the method for constructing the UAV nest dynamic authorization pool includes: configuring normal authorization and temporary authorization according to the attributes of each department, wherein the normal authorization constructs a callable normal authorization electronic certificate, the normal authorization electronic certificate includes nest ID, authorizing department, authorized event type range, validity period, and platform public key fingerprint, and UAV scheduling is performed through HTTPS two-way certificate authentication; the temporary authorization is achieved by the requester generating a temporary authorization request and sending the temporary authorization request to the corresponding department, and after confirming the temporary authorization request, generating a temporary token and returning it to the requester, the requester using the temporary token to generate temporary UAV scheduling for the corresponding nest.

[0011] According to another preferred embodiment of the present invention, the UAV scheduling method includes: using a genetic algorithm to select UAVs that meet the corresponding task requirements from authorized UAV nests, wherein the decision variable of the genetic algorithm is a task-nest allocation matrix, the objective function is to minimize the main monitoring response time, minimize the total completion time and maximize the sum of weighted products of power redundancy, and the constraints include power constraints, authorization state constraints and single nest single task constraints; and solving to output the task-nest-UAV mapping, the estimated takeoff time and the planned route.

[0012] According to another preferred embodiment of the present invention, the UAV collaborative relay task includes: triggering the collaborative relay task based on the battery value of the UAV currently performing the main detection task and a preset battery threshold; querying authorized UAV nests and obtaining candidate UAVs that meet the task requirements; using the weighted sum of the distance between the candidate UAV and the current location, the battery percentage of the candidate UAV, and the power matching degree of the candidate UAV as the candidate UAV scheduling comprehensive score; and selecting the candidate UAV with the highest scheduling comprehensive score to perform the UAV collaborative relay task.

[0013] According to another preferred embodiment of the present invention, the method of the multi-UAV cooperative task conflict resolution mechanism includes: constructing mutually isolated spatial layers based on different task types, wherein the spatial layer includes a monitoring layer, a coordination layer and a delivery layer, each scheduled candidate UAV performs a continuous flight task of the corresponding spatial layer according to its own task type, UAVs in the same spatial layer perform temporal coordination according to the task dependency relationship DAG, and low-priority task UAVs perform hovering or horizontal offset actions of a certain distance when encountering high-priority UAVs.

[0014] To achieve at least one of the above-mentioned objectives, the present invention further provides a multi-nest collaborative relay drone emergency dispatch system, wherein the system executes the above-mentioned multi-nest collaborative relay drone emergency dispatch method.

[0015] The present invention further provides a computer-readable storage medium storing a computer program, which is executed by a processor to implement the above-described method for emergency dispatching of unmanned aerial vehicles (UAVs) through multi-nest collaborative relay.

[0016] This further provides a multi-nest collaborative relay method for emergency dispatching of unmanned aerial vehicles (UAVs), the core of which includes: The system monitors the status of the drone performing the primary monitoring task in real time. When the status parameters meet the preset relay trigger conditions, it queries available drones in the surrounding authorized drone nests. Based on the location, battery level, and functional compatibility of the available drones, it calculates a comprehensive score for the replacement source and selects the available drone with the highest score as the replacement drone. It synchronizes a task context data packet to the replacement drone, which includes the current monitoring point coordinates, target tracking feature vector and predicted trajectory, and marked point information. It controls the replacement drone to take over the primary monitoring task at the handover point, and the original drone returns to base, achieving continuous and uninterrupted emergency site monitoring.

[0017] In one preferred embodiment of the present invention, the method includes: the authorized drone nest includes a normally authorized drone nest and a temporarily authorized drone nest; the normally authorized drone nest directly calls the corresponding drone according to a pre-set digital certificate, and the temporarily authorized drone nest obtains a time-limited scheduling token through one-click confirmation by the corresponding department.

[0018] The beneficial effects of the above technical solution are as follows: 1. Improved response speed: From AI scene recognition to the arrival of the first drone, the entire process is controlled within 10 seconds (normal authorized drone nest) to within 60 seconds (including temporary authorization confirmation), which is 30-90 times faster than the existing manual scheduling mode (5-15 minutes).

[0019] 2. Breakthrough in continuous monitoring capabilities: Through multi-nest relay scheduling and task context synchronization, the main monitoring task can be executed continuously for several hours with a monitoring interruption time of less than 1 second, achieving a qualitative breakthrough compared to the single-machine endurance (≤30 minutes) and the gap for machine replacement and relocation (5-10 minutes) of existing technologies.

[0020] 3. Cross-departmental scheduling is compliant and efficient: the normalized authorization hub can be automatically called in 0 seconds, and the temporary authorization hub can be authorized in an average of 45 seconds with one-click confirmation, with a maximum of no more than 5 minutes, which solves the long-standing contradiction between "emergency efficiency" and "departmental authority" in government scenarios.

[0021] 4. Enhanced airspace collaboration security: Through three layers of high isolation and dynamic priority adjustment of the monitoring layer, collaboration layer and delivery layer, the success rate of resolving conflicts in multi-department and multi-UAV operations in the same domain is >99%, avoiding the risk of collisions caused by manual coordination.

[0022] 5. Reduced reliance on manual intervention: From event identification, department matching, contingency plan decomposition, machine nesting scheduling, relay collaboration to conflict resolution, the entire process has an automation rate of >95%, requiring only one-click authorization confirmation from department on-duty personnel, which greatly reduces the reliance on the experience of professional personnel for emergency dispatch. Attached Figure Description

[0023] Figure 1 The diagram shows a flowchart of an emergency dispatch method for unmanned aerial vehicles (UAVs) with multi-nest collaborative relay according to the present invention. Detailed Implementation

[0024] The following description is intended to disclose the present invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art. The basic principles of the invention defined in the following description can be applied to other embodiments, modifications, improvements, equivalents, and other technical solutions that do not depart from the spirit and scope of the invention.

[0025] It is understood that the term "a" should be understood as "at least one" or "one or more," that is, in one embodiment, the number of an element can be one, while in another embodiment, the number of the element can be multiple, and the term "a" should not be understood as a limitation on the number.

[0026] Please combine Figure 1 This invention discloses an emergency dispatch method and system for UAVs with nested relay coordination, wherein the method mainly includes: S01. Collect multi-source sensor data including video streams, input the multi-source sensor data into an AI model to identify emergency scenario types, and obtain the corresponding type of emergency scenario. S02. Pre-build and maintain a mapping relationship between departments and emergency strategies based on emergency scenario types, and find the corresponding emergency strategies based on the identified emergency scenario types and mapping relationships; S03. Pre-build and maintain a dynamic authorization pool for drone nests of multiple departments, input the emergency strategy into a pre-trained large language model to decompose tasks and output a multi-task list, and configure priority for each task in the multi-task list. S04. Construct a multi-UAV collaborative task conflict resolution mechanism with spatial isolation and layering. Generate a UAV scheduling strategy for collaborative relay based on the current UAV status, and call the corresponding UAV from the corresponding dynamic authorization pool to execute the collaborative relay task.

[0027] Specifically, the method and system described in this invention operate on an urban low-altitude emergency sensing and dispatching system platform. This platform communicates with multiple drone systems deployed in various locations within the monitored area. The platform is deployed on a government cloud or private data center and communicates with each drone location via a 4G / 5G private network or the government extranet. Each drone location belongs to a different government department (such as fire, emergency response, public security, township / street offices, etc.) and accesses the system through a unified platform, but the dispatching authority of each drone location is controlled by departmental authorization policies. The number of drone locations is configured according to the actual situation of the monitored area, typically 3-5 covering key townships and the core area of ​​county towns. A wider range of emergency response is achieved through the core relay mechanism of this invention.

[0028] In this invention, drone nests in different departments are connected through a platform-authorized communication mechanism. Each drone is equipped with multiple sensors, including but not limited to visible light cameras and infrared thermal imagers. The visible light camera has a resolution of 4K and a frame rate of 30fps. The infrared thermal imager has a temperature measurement range of -20℃ to 550℃ and a thermal sensitivity of <50mK. The sensors also include an onboard edge computing unit with a computing power configuration of 4TOPs.

[0029] The emergency scene recognition described in this invention employs multiple models for identification. In a preferred embodiment, an improved YOLOv8 model is used to identify the video stream captured by the visible light camera frame by frame to obtain target information. The improved YOLOv8 model is used as input to the visible light image and outputs bounding boxes for target information such as smoke and crowds. A ResNet-50 model is used to extract and classify features from the infrared thermal image, which can identify thermal anomalies including, but not limited to, fires. Furthermore, this invention performs a dual-path fusion judgment based on the aforementioned target information and thermal anomaly information to obtain the final scene classification result. The fusion judgment method can employ, but is not limited to, existing DS evidence theory to determine the scene classification. The scene classification result includes scene type, subtype, and risk level. It should be noted that both the AI ​​model and DS evidence theory have existing technologies, and this invention does not improve upon them. Therefore, the principles of the AI ​​model and DS evidence theory will not be elaborated upon further in this invention.

[0030] For a specific example, taking a mountainous township in the northern part of a county as an example, during routine drone inspections, AI identified abnormal smoke and temperature in an industrial park. The system then integrated these findings to output a decision categorized as {Accident / Disaster, Fire, Level III}. The system's control platform stored the results in a time-series database and sent a trigger signal to the scheduling decision module. This AI-based automated emergency scenario identification can effectively reduce response time in emergency situations and minimize human intervention.

[0031] It is worth mentioning that the system platform described in this invention constructs and maintains a preset "scenario-department-emergency strategy" mapping table, which is stored in a PostgreSQL relational database. The emergency strategy is efficiently queried using the PostgreSQL relational database. For example, based on the scenario type "accident disaster" and subtype "fire" identified by the AI ​​model, the mapping table is queried to determine that the leading unit is the fire department, and the cooperating units are public security, medical, etc., triggering the emergency plan number XF-001.

[0032] Furthermore, this invention requires task decomposition of the emergency strategy to obtain different task types. The specific decomposition method includes: the platform calls a structured extraction interface based on a pre-trained large language model (such as ChatGLM-6B), and converts the unstructured emergency strategy text imported into the system into a JSON format task list through a prompting process. Taking the XF-001 fire emergency plan as an example, the emergency strategy text is dynamically decomposed into: T1 main monitoring task, which includes: continuous thermal imaging monitoring, fire department, estimated 45 minutes; T2 collaborative task, which includes: aerial evacuation announcement, public security department, estimated 20 minutes, triggered 5 minutes after T1 starts; T3 conditional task, which includes: nighttime lighting guarantee, fire department, triggered when ambient light < 50 lux; T4 collaborative task, which includes: emergency supplies delivery, medical department, estimated 15 minutes, triggered after T2 is completed; the system platform performs JSON schema validation on the LLM output, and stores it in a Redis task queue after passing the validation. The above-mentioned multi-task decomposition method, which utilizes large language modules, satisfies the rationality and scientific nature of the tasks, while also reducing the need for manual intervention in task configuration.

[0033] This invention establishes an authorization pool for drone nests across different departments based on a system platform. This authorization pool is used for dynamic authorization of drone nests in different departments. The construction method of the authorization pool includes: pre-constructing normalized authorization and temporary authorization. Normalized authorization is generally for fire-fighting drone nests, but in other feasible scenarios of this invention, drone nests with public interest coordination functions can also be configured for normalized authorization. Temporary authorization can be for public security drone nests, medical drone nests, etc. For example, fire-fighting drone nests A and B are defined as normalized authorizations. The system platform directly calls them using a pre-set national cryptographic SM2 digital certificate. The certificate includes the nest ID, authorizing department, authorized event type range, validity period, and platform public key fingerprint. The call is performed through HTTPS two-way certificate authentication, with the entire process taking less than 500ms. Public security drone nest C and medical drone nest D are temporary authorizations. When the requester sends a temporary authorization request in JWT format to the department's emergency duty terminal through the system platform described in this invention, the duty personnel confirm with one click, and the platform obtains a temporary token for dispatch. At this time, the requester can use the temporary token to execute drone dispatch for the corresponding department's drone nest. The above-mentioned authorization method, which combines routine and temporary authorizations to create an authorization pool, enables drone relay dispatch to meet both compliance and efficiency requirements.

[0034] It is worth mentioning that in this invention, the UAV's functions match the task requirements, the battery power meets the condition of "flying + executing + returning" ≤ rated battery power × 80%, and the communication signal strength > -85dBm. The NSGA-II multi-objective genetic algorithm is used to solve the initial scheduling scheme: the decision variable of the genetic algorithm is the task-nest allocation matrix, and the objective function is the sum of the weighted products of minimizing the main monitoring response time, minimizing the total completion time, and maximizing the battery redundancy. The parameters corresponding to the above objective function are normalized parameters, and the constraints include, but are not limited to, battery constraints, authorized state constraints, and single-nest single-task constraints. The algorithm parameters of the genetic algorithm include: population size 100, 500 iterations, crossover probability 0.8, and mutation probability 0.1. The genetic algorithm is used to solve for the output task-nest-UAV mapping, estimated takeoff time, and planned route.

[0035] Furthermore, this invention provides a method for multi-nest relay scheduling and airspace conflict resolution, the method comprising: During the execution of the main monitoring task, the system platform receives a status data packet of U1 drone every 10 seconds. U1 represents the drone currently performing the corresponding type of task. When the battery level drops below the 20% threshold, a relay task scheduling is triggered: Authorized drone nests within a 50km radius of U1 are queried. A comprehensive score is calculated using Score = 0.5 / (distance + 0.1) + 0.3 × battery percentage + 0.2 × functional compatibility. The fire-fighting drone nest B with the highest comprehensive score is selected as the replacement source. Assume that fire-fighting drone nest B is 3.5km from U1's current location, the thermal imaging drone is available, and the battery level is 90%.

[0036] The system platform sends a pre-takeoff command to U1 / U2, which is selected through the genetic algorithm. U2 takes off in advance and flies along the route planned by the A* algorithm to the handover point. During the handover, U1 and U2 synchronize a task context data packet (approximately 900 bytes), including: task ID, execution duration, current monitoring point coordinates, gimbal pitch angle, azimuth angle, zoom magnification, target thermal imaging 128-dimensional feature vector, and target predicted trajectory. The target predicted trajectory includes six sets of coordinate points for the next 30 seconds and a list of marked points. Environmental parameters are also acquired, including wind speed, visibility, and temperature. After loading the context, U2 adjusts its gimbal alignment. When the cosine similarity of the feature vector is >0.85, target lock is confirmed, and U2 takes over. U1 returns to base, and monitoring is seamlessly switched. This collaborative relay and context synchronization method effectively improves the continuous capability of scene monitoring.

[0037] When multiple departments and multiple drones respond simultaneously, the airspace conflict resolution module performs the following: Default altitude layer isolation, where the altitude layers include: monitoring layer (120-150m), coordination layer (80-110m), and delivery layer (60-80m), with a vertical spacing ≥10m; in case of conflict within the same layer, dynamic priority adjustment is performed according to "main monitoring task > personnel evacuation task > material delivery task," with lower-priority drones hovering or horizontally offset by 30-50m; fine-tuning of altitude or replanning of flight paths is performed as needed. Simultaneously, time-series coordination is performed according to the task dependency DAG, such as automatically triggering T4 after T2 is completed. The above emergency dispatch can be repeated according to a specified logical order.

[0038] Furthermore, after the emergency response is completed, the criteria for determining the end of the emergency response can be: the commanding personnel click "Event Resolved" or the AI-recognized scene features disappear for 10 minutes; the system platform sends a return-to-home command to all drones, and the drones automatically land and recharge. The platform releases all temporary authorization tokens, marks them as "expired," and notifies the department terminals. Data archiving of the entire task process includes: video clips, flight trajectories, and scheduling logs. The scheduling logs contain JSON-formatted text and records of authorization, relay, and conflict resolution. An automatic response report is generated, which includes response time, duration of continuous monitoring, number of relays, and departmental collaboration efficiency indicators.

[0039] It should be noted that the conflict resolution method in this invention includes: constructing a three-layer altitude layer mechanism for airspace conflict resolution. To simplify airspace conflict resolution and avoid overlapping flight paths of drones from multiple departments, the emergency airspace is divided into three altitude layers according to task type: a monitoring layer (120-150m), where the primary monitoring task has the highest priority (P1); a coordination layer (80-110m), used for functions such as announcements, lighting, and communication relay, with a priority second only to the primary monitoring task (P2); and a delivery layer (60-80m), mainly used for material delivery, with the lowest priority (P3). The drone, using an onboard barometer (accuracy ±0.5m) combined with RTK altitude correction, maintains flight within ±5m of the center of the assigned altitude layer, with a vertical spacing ≥10m.

[0040] Same-layer conflict detection: The platform maintains the occupancy status of a 100m×100m×10m 3D grid in real time, updating every 5 seconds. Two UAVs within the same altitude layer are marked as conflicting if their horizontal distance is less than 200m and their flight paths intersect (intersection angle > 30°) or if their distance in the same direction is less than 100m.

[0041] The conflict resolution strategy is executed in priority order: 1. Priority avoidance: lower priority drones hover or shift horizontally by 30-50m; 2. Sequential waiting: drones of the same priority are served on a "first-come, first-served" basis, with later-arriving drones hovering 100m before the conflict point; 3. Altitude fine-tuning: ±5m fine-tuning is allowed in emergencies (platform reporting required); 4. Flight path replanning: if the above strategies fail, the A* algorithm will reroute, with an upper limit of 120% of the original flight path. Simultaneously, sequential coordination is performed based on task dependency DAGs, such as automatically triggering material delivery tasks after the completion of a broadcast evacuation mission to avoid airspace congestion. This conflict resolution scheme effectively avoids the possibility of collisions between drones from different departments coordinating their broadcasts.

[0042] The processes described in the flowcharts above, as disclosed in the embodiments of this invention, can be implemented as computer software programs. The embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication component, and / or installed from a removable medium. When the computer program is executed by a central processing unit (CPU), the methods of this application are not limited to the aforementioned functions. It should be noted that the computer-readable medium described above in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. Computer-readable storage media can be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wire segments, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to: wireless segments, wire segments, optical fibers, RF, etc., or any suitable combination thereof.

[0043] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0044] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are merely examples and do not limit the present invention. The purpose of the present invention has been fully and effectively achieved. The functions and structural principles of the present invention have been shown and explained in the embodiments. Without departing from the stated principles, the implementation of the present invention may have any variations or modifications.

Claims

1. A method for emergency dispatching of unmanned aerial vehicles (UAVs) through multi-nest collaborative relay, characterized in that, The method includes: Collect multi-source sensor data, including video streams, and input the multi-source sensor data into an AI model to identify emergency scenario types and obtain the corresponding types of emergency scenarios. A mapping relationship between departments and emergency strategies based on emergency scenario types is pre-built and maintained, and the corresponding emergency strategies are found based on the identified emergency scenario types and mapping relationships; A dynamic authorization pool for drone nests of multiple departments is pre-built and maintained. The emergency strategy is input into a pre-trained large language model to decompose tasks and output a multi-task list. Priority is configured for each task in the multi-task list. A spatially isolated and layered multi-UAV collaborative task conflict resolution mechanism is constructed. Based on the current UAV status, a UAV scheduling strategy for collaborative relay is generated, and the corresponding UAV is called from the corresponding dynamic authorization pool to execute the collaborative relay task.

2. The emergency dispatch method for multi-nest cooperative relay of unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The emergency scenario type identification method includes: acquiring visible light images and infrared light images of the target area; inputting the visible light images into a target detection model for target identification to obtain target information; inputting the infrared light images into a thermal imaging detection model for thermal anomaly identification to obtain thermal anomaly information; performing fusion analysis based on the identified target information and thermal anomalies; and generating the emergency scenario type, subtype, and risk level within the target area.

3. The emergency dispatch method for multi-nest cooperative relay of unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The task decomposition method of the emergency strategy includes: by calling the structured extraction interface of the pre-trained large language model, converting the unstructured emergency strategy text found by mapping into a structured JSON format multi-task list through prompt word engineering, wherein the task types of the multi-task list include main detection tasks, collaborative tasks and conditional tasks, wherein the priority of the main detection task is higher than that of the collaborative task, the priority of the collaborative task is higher than that of the conditional task, and the triggering conditions between different task types are configured.

4. The emergency dispatch method for multi-nest cooperative relay of unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The method for constructing a dynamic authorization pool for drone nests includes: configuring normal authorization and temporary authorization based on the attributes of each department; wherein the normal authorization constructs a callable normal authorization electronic certificate, which includes nest ID, authorizing department, authorized event type range, validity period, and platform public key fingerprint, and uses HTTPS two-way certificate authentication during drone scheduling; the temporary authorization involves the requester generating a temporary authorization request and sending the temporary authorization request to the corresponding department, and after confirming the temporary authorization request, generating a temporary token and returning it to the requester, who then uses the temporary token to generate temporary drone scheduling for the corresponding nest.

5. The emergency dispatch method for multi-nest cooperative relay of unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The UAV scheduling method includes: using a genetic algorithm to select UAVs that meet the corresponding task requirements from authorized UAV nests, wherein the decision variable of the genetic algorithm is the task-nest allocation matrix, the objective function is to minimize the main monitoring response time, minimize the total completion time and maximize the sum of the weighted products of power redundancy, and the constraints include power constraints, authorization status constraints and single nest single task constraints; and solving to output the task-nest-UAV mapping, the estimated takeoff time and the planned route.

6. The emergency dispatch method for multi-nest cooperative relay of unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The drone collaborative relay task includes: triggering the collaborative relay task based on the drone's own battery value and a preset battery threshold, querying the authorized drone nests, obtaining candidate drones that meet the task requirements, using the weighted sum of the distance between the candidate drone and the current location, the candidate drone's battery percentage, and the candidate drone's power matching degree as the candidate drone's scheduling comprehensive score, and selecting the candidate drone with the highest scheduling comprehensive score to execute the drone collaborative relay task.

7. The method for emergency dispatching of unmanned aerial vehicles (UAVs) with multi-nest cooperative relay as described in claim 1, characterized in that, The method of the multi-UAV collaborative task conflict resolution mechanism includes: constructing mutually isolated spatial layers based on different task types, wherein the spatial layer includes a monitoring layer, a collaboration layer and a delivery layer, each scheduled candidate UAV performs a continuous flight task in the corresponding spatial layer according to its own task type, UAVs in the same spatial layer perform temporal collaboration according to the task dependency relationship DAG, and low-priority task UAVs perform hovering or horizontal offset actions of a certain distance when encountering high-priority UAVs.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that is executed by a processor to implement a multi-nest cooperative relay unmanned aerial vehicle (UAV) emergency dispatch method according to any one of claims 1-7.

9. A method for emergency dispatching of unmanned aerial vehicles (UAVs) through multi-nest collaborative relay, characterized in that, The method includes: real-time monitoring of the status of the drone performing the main monitoring task; when the status parameters meet the preset relay trigger conditions, querying available drones in the surrounding authorized drone nests; calculating a comprehensive score for the replacement source based on the location, battery level, and functional compatibility of the available drones, and selecting the available drone with the highest score as the replacement drone; synchronizing a task context data packet to the replacement drone, the data packet including the current monitoring point coordinates, target tracking feature vector and predicted trajectory, and marked point information; controlling the replacement drone to take over the main monitoring task at the handover point, and the original drone returning to base, thus achieving continuous and uninterrupted emergency site monitoring.

10. A method for emergency dispatching of unmanned aerial vehicles (UAVs) through multi-nest collaborative relay, characterized in that, The method includes: the authorized drone nests include normally authorized drone nests and temporarily authorized drone nests; the normally authorized drone nests directly call the corresponding drones according to the pre-set digital certificates, and the temporarily authorized drone nests obtain time-limited scheduling tokens through one-click confirmation by the corresponding departments.

Images