Modular unmanned aerial vehicle-vehicle ad hoc network cooperation method and system

Abstract

This application discloses a modular UAV-UAV ad hoc network collaborative method and system. The method includes: dividing the total task to be executed into several subtasks, assigning a unique task continuity token to each subtask, and issuing it to the ground UAV module and the airborne UAV module in the modular UAV-UAV ad hoc network collaborative system; obtaining task execution breakpoint snapshots and task continuity token update information generated after the airborne UAV module executes the subtask with the support of the ground UAV module; when an abnormal event is detected in the airborne UAV module during task execution, constructing a candidate set of replacement actions; determining a target replacement action from the candidate set of replacement actions under preset constraints; and performing platform switching based on the target replacement action. This application improves task continuation efficiency, collaborative scheduling stability, and the continuity and completion rate of tasks to be executed in complex dynamic environments.

Description

Technical Field

[0001] This application relates to the field of unmanned system collaboration technology, and in particular to a modular unmanned aerial vehicle-unmanned vehicle self-organizing network collaboration method and system. Background Technology

[0002] With the rapid development of unmanned systems technology, unmanned ground vehicles (UGVs) and unmanned aerial vehicles (UAVs) are increasingly being used in military reconnaissance, disaster relief, and other missions. UGVs have strong load-bearing capacity and close-range precision inspection capabilities, but are limited by terrain, resulting in limited mobility and field of vision. UAVs have advantages in high mobility and wide-area coverage, but suffer from short endurance, limited payload, and weak embedded computing power.

[0003] In related technologies, air-to-ground collaborative reconnaissance composed of UGVs and UAVs often operates as two separate systems, lacking a highly integrated system. Task redistribution, fault tolerance, dwell switching, and link reconstruction are usually handled as independent processes. The two isolated systems lack an effective intelligent task allocation mechanism in actual deployment, usually relying on manual assignment, making it difficult to optimize resource scheduling based on real-time conditions. This leads to insufficient utilization of resources such as communication bandwidth, computing power, and energy. Consequently, in specific scenarios where mobile nesting stations fail, there is a lack of a task continuity maintenance mechanism. Once a UGV is damaged, loses contact, or its nesting station becomes unavailable during task execution, it will not only trigger the UAV to return and reselect its dwelling target, but also cause a chain of problems such as communication parent node switching, interrupted reconnaissance data transmission, loss of identification model context, and multiple UAVs competing for the same backup nesting position.

[0004] Therefore, for the special application scenario where UGVs assume the role of UAV mobile nesting stations, there is an urgent need for a dedicated collaborative mechanism for maintaining task continuity, so that task continuation, link reconstruction, nesting station switching and breakpoint recovery can be executed in a coordinated manner under unified constraints. Summary of the Invention

[0005] The main purpose of this application is to provide a modular UAV-autonomous vehicle self-organizing network collaborative method and system, which aims to solve the technical problem of poor task execution scheduling continuity in the modular UAV-autonomous vehicle self-organizing network collaborative scenario in related technologies.

[0006] Firstly, to achieve the above objectives, this application provides a modular UAV-unmanned vehicle self-organizing network cooperative method, the method comprising: The total tasks to be executed are divided into several sub-tasks. After assigning a unique task continuity token to each sub-task, the token is sent to the ground unmanned vehicle module and the airborne unmanned vehicle module in the modular UAV-unmanned vehicle self-organizing network collaborative system. Obtain the task execution breakpoint snapshot and task continuity token update information generated after the aerial drone module executes subtasks with the support of the ground unmanned vehicle module; When an abnormal event is detected in the airborne drone module during mission execution, a candidate set of replacement actions is constructed. Under preset constraints, the target replacement action is determined from the candidate set of replacement actions; Based on the target succession action, the platform is switched to complete the task continuity token inheritance, breakpoint snapshot synchronization, control link switching and task cursor update, so that the aerial UAV module can continue to execute the breakpoint snapshot from the interrupted position according to the updated task cursor and the inherited task.

[0007] In one embodiment, the steps of dividing the total task to be executed into several sub-tasks, assigning a unique task continuity token to each sub-task, and then issuing it to the ground unmanned vehicle module and the airborne unmanned vehicle module in the modular UAV-UAV self-organizing network cooperative system include: After receiving the total number of tasks to be executed, the quadtree grid partitioning algorithm is used to divide the total number of tasks to be executed into several sub-task units, and a corresponding task continuity token is generated for each sub-task unit. The task continuity token includes the sub-task identifier, the main execution node, the successor node, the backup resident nest, the emergency control link, the data backhaul link, and the breakpoint snapshot period. The task continuity token is issued to the ground unmanned vehicle module and the airborne unmanned vehicle module in the modular UAV-unmanned vehicle self-organizing network collaborative system, so that the airborne unmanned vehicle module can assign the airborne unmanned vehicle to execute the corresponding sub-task and update the task continuity token and generate a task execution breakpoint snapshot during the task execution process.

[0008] In one embodiment, the step of determining a target successor action from the candidate set of successor actions under the constraints of backup nesting site resources and emergency control link resources includes: Determine at least one candidate successor action that satisfies the preset constraints from the candidate successor action set; Calculate the task execution context inheritance metric and action switching cost for each candidate successor action; Based on context inheritance metrics and action switching costs, the target replacement action is determined from each candidate replacement action.

[0009] In one embodiment, the calculation steps for the task execution context inheritance metric include: The task execution context inheritance metric is calculated based on the inheritance degree of the coverage graph, the inheritance degree of the target tracking, the inheritance degree of the data not returned, the version matching degree, and the interface compatibility degree. The calculation formula is as follows: in, Indicates that candidate ground-based unmanned vehicles are inheriting snapshots The state of the overlay map that can be restored later; This indicates the target tracking state that the candidate ground unmanned vehicle can recover after inheriting the snapshot; This represents the set of data indexes that are expected to be lost or need to be retransmitted during the inheritance process; This represents the interface compatibility determination function between the aerial drone i and the candidate ground unmanned vehicle j; This indicates the model or task rule version identifier corresponding to the current snapshot of aerial drone i; Indicates candidate ground unmanned vehicles Supported model or task rule version identifier; to For normalized weights.

[0010] In one embodiment, the action switching cost is defined by the following formula: in, This indicates that the coverage area needs to be redone. Indicates the planned coverage area. Indicates the continuity retention rate. Indicates recovery delay. Indicates the restoration of the time delay budget, This indicates the number of switching oscillations within the statistics window. This indicates the upper limit of the allowed number of oscillations. to For normalized weights.

[0011] In one embodiment, the steps of performing platform switching based on the target succession action include: Based on the target takeover action, a preparation request is initiated, requesting the target ground unmanned vehicle's pre-occupied nesting site and link lease, and receiving a snapshot of the task execution breakpoint; If the target ground-based unmanned vehicle sends back a confirmation message, and the task execution breakpoint snapshot verification is successful, then the task continuity token is updated and issued to the ground-based unmanned vehicle and the aerial drone that inherits the task, so that the aerial drone that inherits the task can continue the task with the support of the ground-based unmanned vehicle, and the following steps are executed: Continue executing unfinished tasks, perform supplementary scanning for missing task coverage areas, perform partial reconstruction for lost tracking status, and perform retransmission for unconfirmed data.

[0012] Secondly, to achieve the above objectives, this application further provides a modular UAV-unmanned vehicle self-organizing network cooperative system, the system comprising: The aerial drone module includes several aerial drones; The ground unmanned vehicle module includes several ground unmanned vehicles, each of which includes a drone nest for carrying aerial drones; The scheduling center module is used to execute the steps of the modular UAV-autonomous vehicle self-organizing network cooperative method described above; The aerial drone module, the ground unmanned vehicle module, and the dispatch center module communicate and connect with each other.

[0013] One or more technical solutions proposed in this application have at least the following technical effects: This application enables unified binding of subtasks, main execution nodes, successor nodes, backup nests, emergency links, and breakpoint snapshots through task continuity tokens. In the event of an anomaly, it makes joint decisions based on task context inheritance and switching costs, and completes a smooth platform succession through an atomic switching process of preparation, submission, and rollback. This avoids task interruptions, repeated reconnaissance, and context loss caused by UAVs losing connection, malfunctioning, or the nesting site becoming unavailable. At the same time, this application can suppress resource conflicts and switching oscillations when multiple UAVs contend for backup resources, reduce recovery latency and redo coverage, and improve task continuation efficiency, collaborative scheduling stability, and the continuity and completion rate of pending tasks in complex dynamic environments. Attached Figure Description

[0014] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating the modular UAV-unmanned vehicle self-organizing network collaborative method in the embodiments of this application.

[0017] Figure 2 This is a schematic diagram of the system framework in an embodiment of this application. Detailed Implementation

[0018] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0019] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0020] The main solution of this application embodiment is as follows: After receiving the total tasks to be executed, the scheduling center divides them into several sub-tasks and generates a unique task continuity token for each sub-task, which is then issued to the ground unmanned vehicle module and the aerial drone module. During the task execution process, the scheduling center collects task execution breakpoint snapshots and token update information formed by the drone with the support of the ground unmanned vehicle. When an abnormal event is detected, a candidate set of replacement actions is constructed, which includes the target ground unmanned vehicle, the backup nest, the emergency control link, and the snapshot version to be inherited. Under the constraints of backup nest resources and link resources, the task context inheritance degree and the action switching cost are jointly solved to determine the optimal target replacement action. Then, an atomic switching process of preparation, submission, and rollback is executed to complete the task continuity token inheritance, breakpoint snapshot synchronization, control link switching, and task cursor update. This allows the drone to continue executing the remaining sub-tasks from the interrupted position based on the updated task cursor and the inherited breakpoint snapshot.

[0021] Specifically, this application provides a modular UAV-Vehicle self-organizing network cooperative method, referring to... Figure 1 The modular UAV-autonomous vehicle self-organizing network collaborative method includes steps S10~S50: Step S10: Divide the total tasks to be executed into several sub-tasks, assign a unique task continuity token to each sub-task, and then send it to the ground unmanned vehicle module and the airborne unmanned vehicle module in the modular UAV-unmanned vehicle self-organizing network collaborative system.

[0022] Step S20: Obtain the task execution breakpoint snapshot and task continuity token update information generated after the aerial drone module executes the sub-task with the support of the ground unmanned vehicle module.

[0023] Step S30: When an abnormal event is detected in the airborne UAV module during mission execution, a candidate set of replacement actions is constructed.

[0024] Step S40: Determine the target replacement action from the candidate set of replacement actions under preset constraints.

[0025] Step S50: Based on the target succession action, perform platform switching to complete task continuity token inheritance, breakpoint snapshot synchronization, control link switching, and task cursor update, so that the aerial UAV module can execute breakpoint snapshots according to the updated task cursor and the inherited task, and continue to execute sub-tasks from the interrupted position.

[0026] This embodiment mainly provides a method for maintaining task continuity, which is applied to a collaborative system consisting of at least one ground unmanned vehicle serving as a mobile home station, at least one aerial drone performing sub-tasks, and at least one scheduling node.

[0027] In this embodiment, the ground unmanned vehicle is used to provide services such as parking location, energy supply access, control access and data transmission access for the aerial drone. During the execution of the mission, the aerial drone is allowed to switch ownership between different ground unmanned vehicles. The backup parking location and emergency control link in the system are limited shared resources.

[0028] Among them, the abnormal events triggered by the aerial drone module during mission execution mainly include the loss of contact between the aerial drone performing the mission and its associated ground unmanned vehicle, the failure of the associated ground unmanned vehicle, the unavailability of the associated ground unmanned vehicle's nesting station, the aerial drone module's remaining flight time being lower than the safety threshold, or the mission time limit risk exceeding the warning threshold.

[0029] In this embodiment, a modular UAV-Vehicle self-organizing network collaboration method is implemented. Within any scheduling cycle of the scheduling center, the scheduling center first receives the operational status data of the ground Vehicle module, such as the location status, charging capacity, link capacity, health status, remaining resources, and interface compatibility information of each ground Vehicle; and the operational status data of the airborne UAV module (such as the UAV's pose, speed, remaining battery power, current subtask identifier, task cursor (the task cursor can be used to represent the current task's track segment number, coverage grid coordinates, target tracking status number, and pending data queue index, etc.), unreturned data index, current snapshot version, and payload type). The system includes control link quality and health, link status information (such as bandwidth, latency, packet loss rate, congestion level and availability of each candidate link), communication link data (such as bandwidth, latency, packet loss rate, congestion level and availability of each candidate link), and task status information (such as subtask priority, remaining coverage area, target tracking status and deadline). Based on this information, task scheduling is then implemented, and target scheduling actions of the ground unmanned vehicle module and the aerial unmanned vehicle module are output, such as target replacement ground unmanned vehicle identifier, target backup nest identifier, emergency control link identifier, snapshot version to be inherited, updated task cursor and continuation control instructions, etc.

[0030] Specifically, the scheduling center module first divides the total tasks to be executed into several sub-tasks, and generates a unique task continuity token for each sub-task before issuing it to the ground-based unmanned vehicle module and the aerial drone module. The task continuity token is used to uniformly bind task context information such as the primary execution node, the successor node, the backup residing location, the emergency control link, and breakpoint snapshots. This ensures that the task execution state is no longer dependent on a single ground-based unmanned vehicle platform, but rather forms a unified, transferable, and inheritable task state object, thus providing a foundation for task continuation in subsequent abnormal situations. Compared to the independent processing methods of task scheduling, link switching, and breakpoint recovery in related technologies, this embodiment can establish a unified continuity constraint relationship at the task creation stage, improving the scheduling consistency and collaborative stability of the system in the face of dynamic task environments.

[0031] During mission execution, the aerial UAV module, supported by the ground-based unmanned vehicle module in parking, power supply, communication, and control, executes corresponding sub-tasks and periodically generates mission execution breakpoint snapshots and mission continuity token update information. The mission execution breakpoint snapshot records mission context information such as the current mission cursor, target tracking status, index of unreturned data, and changes in coverage area, enabling the system to save the mission execution status in real time. When link fluctuations, mission switching, or platform anomalies occur, the scheduling center module can quickly restore the mission context based on the latest breakpoint snapshot, avoiding mission re-execution due to lost mission status, thereby reducing redundant reconnaissance range and mission recovery latency, and improving the system's mission continuation efficiency.

[0032] When an abnormal event is detected in the aerial UAV module during mission execution, the dispatch center module constructs a candidate set of replacement actions and determines the target replacement action from this set under preset constraints. These preset constraints may include constraints on backup residing space resources, emergency control link resources, recovery latency, and interface compatibility. By comprehensively analyzing the status of the ground UAV module, communication link status, and mission status information, the dispatch center module jointly evaluates the task execution context inheritance capability and action switching cost of each candidate replacement action. This avoids the link conflicts, task redoing, and context loss problems caused by traditional solutions that only select the nearest replacement platform. The system can prioritize target replacement actions with higher context inheritance integrity and lower recovery costs, improving the ability to maintain mission continuity in complex dynamic environments.

[0033] After determining the target replacement action, the scheduling center module performs a platform switch based on the target replacement action to complete task continuity token inheritance, breakpoint snapshot synchronization, control link switching, and task cursor updates. After the platform switch is completed, the aerial UAV module uses the updated task cursor and the inherited task to execute the breakpoint snapshot and continue executing the remaining sub-tasks from the interrupted position without having to re-execute the entire reconnaissance process. This approach not only reduces the amount of task redoing and overlapping coverage areas but also reduces the risk of task interruption caused by inconsistent link switching, nest resource conflicts, or asynchronous task states during the switchover process. This improves the task completion rate, task recovery stability, and overall collaborative efficiency of the modular UAV-Vehicle self-organizing network collaborative system in complex environments.

[0034] In one feasible implementation, step S10 further includes steps A10 to A20: Step A10: After receiving the total number of tasks to be executed, the total task area to be executed is divided into several sub-task units using a quadtree grid partitioning algorithm, and a corresponding task continuity token is generated for each sub-task unit. The task continuity token includes the sub-task identifier, the main execution node, the successor node, the backup resident nest, the emergency control link, the data backhaul link, and the breakpoint snapshot period. Step A20: The task continuity token is sent to the ground unmanned vehicle module and the airborne unmanned vehicle module in the modular UAV-unmanned vehicle self-organizing network collaborative system, so that the airborne unmanned vehicle module assigns the airborne unmanned vehicle to execute the corresponding sub-task and updates the task continuity token and generates a task execution breakpoint snapshot during the task execution process.

[0035] Step S40 also includes steps B10 to B30: Step B10: Determine at least one candidate successor action that satisfies the preset constraints from the candidate successor action set.

[0036] Step B20: Calculate the task execution context inheritance metric and action switching cost for each candidate successor action.

[0037] Step B30: Based on the context inheritance metric and action switching cost, determine the target replacement action from among the candidate replacement actions.

[0038] The calculation steps for the task execution context inheritance metric include: The task execution context inheritance metric is calculated based on the inheritance degree of the coverage graph, the inheritance degree of the target tracking, the inheritance degree of the data not returned, the version matching degree, and the interface compatibility degree. The calculation formula is as follows: in, Indicates that candidate ground-based unmanned vehicles are inheriting snapshots The state of the overlay map that can be restored later; This indicates the target tracking state that the candidate ground unmanned vehicle can recover after inheriting the snapshot; This represents the set of data indexes that are expected to be lost or need to be retransmitted during the inheritance process; This represents the interface compatibility determination function between the aerial drone i and the candidate ground unmanned vehicle j; This indicates the model or task rule version identifier corresponding to the current snapshot of aerial drone i; Indicates candidate ground unmanned vehicles Supported model or task rule version identifier; to For normalized weights.

[0039] The cost of the action switching is defined by the following formula: in, This indicates that the coverage area needs to be redone. Indicates the planned coverage area. Indicates the continuity retention rate. Indicates recovery delay. Indicates the restoration of the time delay budget, This indicates the number of switching oscillations within the statistics window. This indicates the upper limit of the allowed number of oscillations. to For normalized weights.

[0040] Furthermore, step S50 includes steps C10 to C20: Step C10: Based on the target takeover action, initiate a preparation request, request the pre-occupied nesting site and link lease of the target ground unmanned vehicle, and receive a snapshot of the task execution breakpoint; Step C20: If the target ground unmanned vehicle sends back a confirmation message and the task execution breakpoint snapshot verification is successful, then update the task continuity token and issue it to the ground unmanned vehicle and the aerial drone that inherit the task, so that the aerial drone that inherits the task can continue the task with the support of the ground unmanned vehicle, and execute the following steps: Continue executing unfinished tasks, perform supplementary scanning for missing task coverage areas, perform partial reconstruction for lost tracking status, and perform retransmission for unconfirmed data.

[0041] Specifically, after receiving the total number of tasks to be executed, the scheduling center module uses a quadtree grid partitioning algorithm to divide the reconnaissance area into several sub-task units and generates a corresponding task continuity token for each sub-task unit. The task continuity token is used to bind task context information such as sub-task identifier, main execution node, successor node, backup residing location, emergency control link, data backhaul link, and breakpoint snapshot period, and is then distributed to the ground unmanned vehicle module and the aerial drone module. By establishing a unified task state constraint relationship during the task creation phase, the task execution state can be inherited and migrated between different platforms, avoiding the problem of task scheduling, link switching, and breakpoint recovery being isolated from each other in related technologies. This improves the system's ability to maintain task continuity and the stability of collaborative scheduling in complex dynamic environments.

[0042] During mission execution, the aerial UAV module, supported by the ground-based unmanned vehicle module in parking, power supply, control, and communication, executes corresponding sub-tasks and continuously updates the mission continuity token and generates mission execution breakpoint snapshots. These snapshots record mission context information such as the mission cursor, target tracking status, coverage area changes, and indexes of unreturned data, enabling the scheduling center module to monitor the mission execution status in real time. In the event of platform disconnection, link anomalies, or mission switching, the system can quickly restore the mission context based on the latest breakpoint snapshot, reducing mission rework and redundant reconnaissance, thereby lowering mission recovery latency and improving mission continuation efficiency.

[0043] When the aerial drone module triggers an abnormal event during mission execution, the scheduling center module constructs a candidate set of replacement actions and selects those that meet preset constraints. Subsequently, the system evaluates each candidate replacement action based on mission execution context inheritance metrics and action switching costs. The mission execution context inheritance metrics comprehensively consider coverage graph inheritance, target tracking inheritance, unreturned data inheritance, version matching, and interface compatibility to measure the candidate ground drone's ability to recover from the current mission context. The action switching cost, combined with mission redo volume, recovery latency, and switching oscillations, evaluates the overall recovery overhead caused by platform switching. By simultaneously considering mission context inheritance capabilities and action switching costs, the system avoids the link conflicts, mission redo, and context loss problems caused by traditional solutions that select replacement platforms solely based on geometric distance. This allows target replacement actions to reduce recovery overhead and system oscillation risks while ensuring mission continuity.

[0044] After determining the target takeover action, the dispatch center module initiates a preparation request based on the target takeover action, requesting the target ground unmanned vehicle to pre-occupy a nesting site and link lease, and receiving a snapshot of the task execution breakpoint. When the target ground unmanned vehicle sends back a confirmation message and the task execution breakpoint snapshot is successfully verified, the system updates the task continuity token and issues it to the ground unmanned vehicle and the aerial drone inheriting the task, enabling the aerial drone inheriting the task to continue executing the unfinished task with the support of the new ground unmanned vehicle. If any step fails, a rollback is performed, releasing the pre-occupied resources and re-entering the candidate action or emergency mode. In this way, the aerial drone does not need to re-execute the entire reconnaissance process, but continues to execute the remaining tasks from the interruption position based on the updated task cursor and the inherited task execution breakpoint snapshot, and only performs supplementary scanning on the missing task coverage area, partial reconstruction on the lost tracking status, and supplementary transmission of unconfirmed data, thereby effectively reducing the duplicate coverage area, reducing task recovery latency, and improving the overall task completion rate and task continuation stability of the modular drone-unmanned vehicle self-organizing network collaborative system.

[0045] Furthermore, to ensure real-time performance, in a preferred embodiment, the status acquisition cycle is configured as follows: to The breakpoint snapshot period is configured as follows: to The local joint solution cycle configuration is as follows to Snapshot synchronization budget configured as to One round-trip delay window; the number of candidate replacement actions is preferably limited to [number]. Within a certain number, so that airborne or vehicle-mounted compensation calculations remain within a given computing power budget.

[0046] For airborne drones with limited onboard computing power, it is preferable for the airborne drones to only perform candidate feasibility pre-screening and snapshot packaging, while the ground unmanned vehicles or scheduling nodes complete the joint solution; for cases where the central link is unreachable, it is preferable for the airborne drones and neighboring ground unmanned vehicles to perform decentralized local solutions and atomic switching.

[0047] When no candidate replacement action satisfies the feasibility constraint, the system enters the emergency handling branch: If the remaining battery power is higher than the emergency hovering threshold, the drone in the air will enter a timed waiting period and repeatedly initiate candidate searches.

[0048] If the remaining battery power is lower than the emergency hovering threshold, the landing will be carried out according to the preset emergency landing point or safe forced landing strategy. At the same time, the most recent keyframe pointer and the index of the unreturned data will be sent to the scheduling node or the healthy ground unmanned vehicle through the available link.

[0049] If the snapshot version verification fails, it will fall back to the most recent common keyframe and perform a patch scan only on the missing differential parts.

[0050] If the optimal cost of local joint solution exceeds the continuity budget, a global replanning for the relevant task cluster is triggered.

[0051] In the above method, the universal nesting station, charging device, dual-link communication device and standardized task interface on the ground unmanned vehicle serve as the execution environment for implementing the software method. The aerial target recognition model, path planning model or communication anti-interference model can all be accessed as optional modules, but they themselves do not constitute a necessary limitation of the task continuity maintenance method, unless their output is written into the task continuity token or breakpoint snapshot and directly participates in the candidate succession action solution.

[0052] To facilitate understanding of this method embodiment, a specific implementation example is provided below for further detailed explanation: In this example, including More than one unmanned ground vehicle that can serve as a mobile home station , More than 100 aerial drones to perform the tasks to be carried out ,as well as Each scheduling node. Release information such as nest location status, health, resource lease table, and available link status to the public; The system publishes the current location, remaining energy, task cursor, snapshot version number, and index of data not yet returned; the scheduling node maintains the task continuity token table, keyframe table, differential snapshot table, and resource lease table. The system time synchronization error is preferably no greater than [value missing]. This ensures consistent judgment between the nest time window and the link time window.

[0053] In this embodiment, the software architecture includes the following functional modules: a status acquisition and normalization module, used to acquire and normalize data from... , The system includes: a link node status collection module that converts the status into a unified status vector; a task continuity token management module for creating, updating, and inheriting task continuity tokens for subtasks; a snapshot versioning module for generating keyframes, differential snapshots, and maintaining snapshot version pointers; a candidate replacement action generation module for constructing quaternary replacement actions after an anomaly is triggered; a continuity evaluation module for calculating continuity preservation rate and atomic switching cost; and a local joint solution module for handling multiple... Nest and link allocation under concurrent contention; atomic switching execution module for performing preparation, commit, and rollback; breakpoint resumption module for resuming execution state and controlling execution based on task cursor. Continue to complete the remaining tasks.

[0054] In a preferred interface implementation, the heartbeat and loss detection interfaces Used to transmit liveness and health information; Task token interface Used to transmit task continuity token update messages; snapshot interface Used for transmitting keyframe pointers, differential snapshots, and checksums; Nest lease interface Used for transmitting messages about reserving and releasing spare nesting sites; external control interface Used to issue control commands related to position, speed, attitude, or task cursor.

[0055] If adopted or compatible The implementation, Mappable to Keep-alive messages, and Mappable to , and related messages about external settings; if using Control, external control keep-alive flow should continuously meet platform requirements; if adopted For control-type themes, it is preferable to use , and Semantic, sensing and telemetry summarization preferred use Semantics, snapshots and token synchronization topics are preferred. Semantics; if adopted Then snapshots and tokens can be obtained. Or higher semantic transmission. The above mapping is only a preferred implementation and does not limit the present invention.

[0056] The software flow of this embodiment is as follows.

[0057] Step 1: Task Receiving and Subtask Generation After receiving the tasks to be executed, the scheduling node divides the task area into several atomic task units and generates a corresponding task continuity token for each atomic task unit. .

[0058] The partitioning method can be a grid, partition, corridor, or target cluster, as long as it can provide a unique location for the cursor in subsequent tasks.

[0059] Task continuity token During creation, the main execution node, candidate successor node set, backup nest time window, emergency control link set, initial keyframe pointer, and continuity budget are written simultaneously.

[0060] Step 2: Normal Execution and Snapshot Generation In its jurisdiction Subtasks are executed under the provided takeoff and landing and control support, and in a periodic manner. Generate a snapshot of the breakpoint; Alternatively, when completing a critical flight segment, identifying a key target, experiencing link fluctuations, or about to enter a high-risk area, a differential snapshot can be generated through the snapshot versioning module.

[0061] Differential snapshots are preferably made to include only the changed regions and changed states to reduce synchronization overhead.

[0062] Scheduling node or master After receiving the differential snapshot, update the task continuity token. The snapshot pointer and task cursor in the file.

[0063] Step 3: Exception Triggering and Candidate Generation When the belonging is detected When the heartbeat times out, health level falls below the threshold, the nest station becomes unavailable, remaining flight time falls below the threshold, or the risk of mission overdue exceeds the threshold. Alternatively, the scheduling node may initiate a decision to continue execution.

[0064] The system starts from neighborhood health A set of candidate replacement actions is constructed. Each candidate replacement action is simultaneously bound to a target. The specific nest location, specific emergency control link, and snapshot version to be inherited should be provided, rather than just a single candidate. Logo.

[0065] Step 4: Continuous evaluation and motion solving The system first bases its score on a health threshold. Continuity retention rate threshold Recovery delay threshold Remaining energy threshold Minimum bandwidth threshold Maximum delay threshold Feasibility pruning is performed based on interface compatibility conditions.

[0066] Subsequently, the system calculates the continuity retention rate for each candidate replacement action. and the cost of switching atoms .

[0067] If multiple aircraft are in the same time period If a switch is required, the system will jointly consider the conflicts caused by shared nesting sites and shared links, and output the action combination with the minimum total cost.

[0068] Step 5: Atomic Switch Execution The system initiates a preparation request for the target action. Request target Reserve nesting sites and link leases, and receive differential snapshot metadata.

[0069] If the target Send confirmation message If the snapshot verification is successful, then the commit operation will be performed. Atomic updates are performed on the main execution node, task cursor, snapshot pointer, and link binding.

[0070] If any condition is not met, a rollback operation will be performed. Release the pre-reserved resources and switch to the next candidate successor action. Upon receiving Fly to the target It completes guided landing, link switching, and continuation of operations.

[0071] Step 6: Resuming execution from breakpoints Stay at the target Afterwards, instead of re-downloading the entire task, it continues to execute the unfinished parts based on the inherited task cursor; it only performs supplementary scanning on the missing coverage areas, only performs partial reconstruction on the lost tracking status, and only performs supplementary transmission on the unconfirmed data.

[0072] If snapshot inheritance Below the full inheritance state but still above the threshold Then, the execution will proceed in the order of "continuing the execution first, followed by partial compensation".

[0073] In a specific example, The original main platform was When an anomaly occurs, the candidate and The straight-line distance is Candidate and The straight-line distance is If the nearest distance strategy is adopted, it will be selected first. ;but The backup nest has a conflict within the predicted arrival window, and its control interface version is incompatible with the current snapshot version, which makes the differential snapshot unable to be directly inherited, resulting in a large expected recovery delay and an increase in task redoing. Although the distance is greater, its nesting sites are pre-reserved, the interfaces are compatible, the current emergency control link can be directly accessed, and the snapshot retention rate is higher. Therefore, the atomic switching cost can be obtained by calculating according to the cost function described in this invention, and the system ultimately selects... Therefore, in the scenario of this invention, a more distant replacement platform can achieve a shorter effective recovery time and less task redoing, indicating that the optimality of this invention is not based on geometric distance, but on task continuity.

[0074] In another example, and Simultaneously apply for the same The same backup nesting window. The system treats the two succession actions as mutually exclusive actions, adds them to the local conflict set, and performs a joint solution based on their respective continuity budgets, recovery latency, and task priorities. If If the mission has more remaining critical objectives and a tighter recovery budget, then nest slots should be prioritized for allocation. ; Then choose the second-best but feasible option. Alternatively, it can enter a short waiting period and solve the problem again. This method avoids multiple [problems / issues]. The failure to continue execution due to simultaneous competition for the same resource.

[0075] Under a preferred real-time configuration, the publication cycle for status messages and keep-alive messages can be configured as follows: The upper limit of snapshot differential synchronization should preferably not exceed the size of the most recent keyframe. to The optimal solution for local candidate actions is found in Completed within [timeframe], atomic switching control decision optimization is preferred within [timeframe]. Issued internally.

[0076] If an external flight control interface is used, the keep-alive control messages should continuously meet the minimum transmission requirements of the interface platform; if based on... In the message bus, control messages prioritize reliability, loss identifiability, and settable time limits, while sensor summaries prioritize real-time freshness.

[0077] In this embodiment, the computation location of any module can be deployed on the scheduling node. or Above. For cases where the central link is reachable, it is preferable for the scheduling node to maintain the global resource lease and complete the joint action solution; for cases where the central link is temporarily unreachable but neighboring links are reachable, it is preferable for... and target Decentralized switching is performed based on locally cached task continuity tokens and snapshot pointers.

[0078] This application also provides a modular UAV-unmanned vehicle self-organizing network collaborative system; please refer to [reference needed]. Figure 2 The modular UAV-unmanned vehicle self-organizing network collaborative system includes: The aerial drone module includes several aerial drones; The ground unmanned vehicle module includes several ground unmanned vehicles, each of which includes a drone nest for carrying the aerial drone; The scheduling center module is used to execute the steps of the modular UAV-autonomous vehicle self-organizing network cooperative method described above; The aerial drone module, the ground unmanned vehicle module, and the dispatch center module are interconnected.

[0079] For example, the system includes several ground-based unmanned vehicles (UGVs), several aerial unmanned aerial vehicles (UAVs), and a central dispatch center. Each UGV is equipped with a universal UAV pod for parking and supporting 1-2 UAVs, enabling precise take-off and landing, efficient charging, and high-speed communication. The dispatch center maintains data connectivity with all UGVs and UAVs via a 5G+LoRa / Mesh dual-link wireless communication network.

[0080] The entire system adopts a heterogeneous architecture that combines distributed collaboration with centralized scheduling: UGVs perform tasks such as patrol, close-range fine reconnaissance, payload support, and communication relay on the ground, while also serving as high-precision mobile bases, charging stations, and communication relay stations for UAVs; UAVs perform tasks such as wide-area reconnaissance, continuous target surveillance, and cognitive radio spectrum sensing networking in the air, completing target identification through lightweight federated learning, and achieving integrated air-ground collaboration with UGVs in terms of payload complementarity, communication coordination, and target designation.

[0081] Each UGV consists of a tracked chassis mobile platform, an integrated universal unmanned aerial vehicle (UAV) pod, a standardized mission payload interface, a tri-mode communication and high-performance computing module, and a fault diagnosis module. The UGV chassis possesses autonomous navigation and obstacle avoidance capabilities based on the fusion of lidar and vision, and can maneuver to designated locations or patrol routes according to instructions from the dispatch center.

[0082] The universal drone nest installed on the UGV is an integrated unit that combines a vision + laser dual-modal guided landing system, an electromagnetic induction wireless charging device, and a 5G + LoRa / Mesh dual communication base station. The guided landing system uses high-definition cameras and lidar to assist the UAV in completing precise landings in complex environments such as low light and fog; the charging device integrates overvoltage and overcurrent protection modules, automatically triggering charging after the UAV lands; the communication base station supports automatic protocol switching, used to realize data exchange between the UAV and UGV and communication relay to the dispatch center.

[0083] The UGV features a modular task payload interface with standardized physical and data interfaces. It has a pre-installed quick-release clip rail mount and a 5-pin aviation power supply / Ethernet / RS485 universal data interface on the vehicle body, which can quickly mount different functional modules without tools to meet the diverse needs of the tasks to be performed.

[0084] The UAV in this example is a quadcopter vertical takeoff and landing unmanned aerial vehicle with a standardized physical + data dual-interface modular design. It is equipped with an embedded computing module, an autonomous flight control system, a cognitive radio tri-mode communication module, and a fault management module, and is suitable for reconnaissance in complex terrain.

[0085] The UAV is equipped with a modular mission payload interface that conforms to the same standard as UGVs, allowing for tool-less and rapid replacement of airborne sensors or equipment. The flight control system automatically adapts to new payloads through a standard driver library, enabling deployment without complex modifications. Typical payload configurations include: visible light / infrared imaging module (4K / 30 frames, infrared thermometry -10℃~800℃), laser ranging / radar module (range ±0.1m, scanning frequency 10Hz), electronic reconnaissance module (spectrum sensing 1MHz-6GHz), and communication relay module (LoRa / Mesh, 10 hops), etc.

[0086] The UAV is equipped with an autonomous flight control system that combines lidar and vision fusion, and a LoRa+Mesh+5G tri-mode communication module. It can receive commands from the dispatch center or the UGV to take off autonomously, fly along designated routes, and land with centimeter-level precision. During missions, the UAV can automatically take off from the UGV's general-purpose drone nest, complete reconnaissance, and then autonomously land back to the nest for high-speed data upload and battery charging. During flight, the UAV maintains real-time communication with the UGV and the dispatch center via a cognitive radio link, transmitting reconnaissance images, monitoring data, and target identification results for mission evaluation, and receiving new commands.

[0087] Each UAV also features intelligent fault management and cross-vehicle dwelling capabilities, with built-in fault detection sensors for real-time monitoring of power, communication, and load status. When it detects a loss of contact with its assigned UGV, receives a cross-vehicle landing command, or meets the level 3 triggering conditions, the UAV obtains the location and status information of nearby UGV nesting stations through an ad hoc network. Based on the principle of minimizing continuity loss, it comprehensively evaluates the dwelling distance, backup nesting station conflict risk, link reconstruction overhead, task redoing volume, and breakpoint snapshot inheritance integrity, and selects a target UGV. After completing a two-way health handshake, it automatically flies to the target UGV and achieves a centimeter-level safe landing. During cross-vehicle dwelling, the task continuity token inheritance, breakpoint snapshot transmission, and emergency control link switching are completed simultaneously, thereby achieving seamless task recovery and continuous execution under cross-platform dwelling conditions.

[0088] When data is exchanged between modules, the data stream formats include: a state stream for transmitting pose, health, and link quality; a task stream for transmitting subtasks and continuity tokens; a snapshot stream for transmitting keyframes, incremental snapshots, and indexes of data not yet returned; and a switchover stream for transmitting pre-emption, acknowledgment, commit, and rollback commands. These data streams are logically independent but work collaboratively under the same task continuity token constraint.

[0089] The modular UAV-Vehicle self-organizing network cooperative system provided in this application, employing the modular UAV-Vehicle self-organizing network cooperative method in the above embodiments, can solve the technical problem of poor task execution scheduling continuity in modular UAV-Vehicle self-organizing network cooperative scenarios in related technologies. Compared with related technologies, the beneficial effects of the modular UAV-Vehicle self-organizing network cooperative device provided in this application are the same as those of the modular UAV-Vehicle self-organizing network cooperative method provided in the above embodiments, and other technical features in the modular UAV-Vehicle self-organizing network cooperative device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0090] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0091] The above are only some embodiments of this application and do not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A modular UAV-unmanned vehicle self-organizing network cooperative method, characterized in that, The method includes: The total tasks to be executed are divided into several sub-tasks. After assigning a unique task continuity token to each sub-task, the token is sent to the ground unmanned vehicle module and the airborne unmanned vehicle module in the modular UAV-unmanned vehicle self-organizing network collaborative system. Obtain the task execution breakpoint snapshot and task continuity token update information generated after the aerial drone module executes subtasks with the support of the ground unmanned vehicle module; When an abnormal event is detected during the execution of a mission by the aerial drone module, a candidate set of replacement actions is constructed. Under preset constraints, the target replacement action is determined from the candidate set of replacement actions; Based on the target handover action, a platform switch is performed to complete the task continuity token inheritance, breakpoint snapshot synchronization, control link switching, and task cursor update. This enables the aerial UAV module to execute the breakpoint snapshot based on the updated task cursor and the inherited task, and continue executing the sub-task from the interrupted position.

2. The modular UAV-UGV self-organizing network coordination method of claim 1, wherein, The steps of dividing the total tasks to be executed into several sub-tasks, assigning a unique task continuity token to each sub-task, and then issuing it to the ground unmanned vehicle module and the airborne unmanned vehicle module in the modular UAV-UAV self-organizing network collaborative system include: After receiving the total number of tasks to be executed, the quadtree grid partitioning algorithm is used to divide the total number of tasks to be executed into several sub-task units, and a corresponding task continuity token is generated for each sub-task unit. The task continuity token includes a sub-task identifier, a main execution node, a successor node, a backup resident nest, an emergency control link, a data backhaul link, and a breakpoint snapshot period. The task continuity token is sent to the ground unmanned vehicle module and the airborne unmanned vehicle module in the modular UAV-unmanned vehicle self-organizing network collaborative system, so that the airborne unmanned vehicle module assigns the airborne unmanned vehicle to execute the corresponding sub-task and updates the task continuity token and generates a task execution breakpoint snapshot during the task execution process.

3. The modular UAV-UGV self-organizing network coordination method of claim 1, wherein, The step of determining the target successor action from the candidate set of successor actions under the constraints of backup nesting site resources and emergency control link resources includes: Determine at least one candidate successor action from the candidate successor action set that satisfies the preset constraints; Calculate the task execution context inheritance metric and action switching cost for each of the candidate successor actions; Based on the context inheritance metric and action switching cost, the target replacement action is determined from each of the candidate replacement actions.

4. The modular UAV-UGV ad hoc networking coordination method of claim 3, wherein, The calculation steps for the task execution context inheritance metric include: The task execution context inheritance metric is calculated based on the inheritance degree of the coverage graph, the inheritance degree of the target tracking, the inheritance degree of the data not returned, the version matching degree, and the interface compatibility degree. The calculation formula is as follows: in, Indicates that candidate ground-based unmanned vehicles are inheriting snapshots The state of the overlay map that can be restored later; This indicates the target tracking state that the candidate ground unmanned vehicle can recover after inheriting the snapshot; This represents the set of data indexes that are expected to be lost or need to be retransmitted during the inheritance process; This represents the interface compatibility determination function between the aerial drone i and the candidate ground unmanned vehicle j; This indicates the model or task rule version identifier corresponding to the current snapshot of aerial drone i; Indicates candidate ground unmanned vehicles Supported model or task rule version identifier; to For normalized weights.

5. The modular UAV-UGV self-organizing network coordination method of claim 4, wherein, The cost of the action switching is defined by the following formula: in, This indicates that the coverage area needs to be redone. Indicates the planned coverage area. Indicates the continuity retention rate. Indicates recovery delay. Indicates the restoration of the time delay budget, This indicates the number of switching oscillations within the statistics window. This indicates the upper limit of the allowed number of oscillations. to For normalized weights.

6. The modular UAV-UGV ad hoc networking and collaboration method of claim 1, wherein, The steps for performing platform switching based on the target replacement action include: Based on the target takeover action, a preparation request is initiated, requesting the pre-occupied nesting site and link lease of the target ground unmanned vehicle, and receiving a snapshot of the task execution breakpoint; If the target ground-based unmanned vehicle sends back a confirmation message, and the task execution breakpoint snapshot verification is successful, then the task continuity token is updated and issued to the ground-based unmanned vehicle and the aerial drone that inherits the task, so that the aerial drone that inherits the task can continue the task with the support of the ground-based unmanned vehicle, and the following steps are executed: Continue executing unfinished tasks, perform supplementary scanning for missing task coverage areas, perform partial reconstruction for lost tracking status, and perform retransmission for unconfirmed data.

7. A modular UAV-unmanned vehicle self-organizing network collaborative system, characterized in that, The system includes: The aerial drone module includes several aerial drones; The ground unmanned vehicle module includes several ground unmanned vehicles, each of which includes a drone nest for carrying the aerial drone; The scheduling center module is used to execute the steps of the modular UAV-unmanned vehicle self-organizing network cooperative method according to any one of claims 1 to 6; The aerial drone module, the ground unmanned vehicle module, and the dispatch center module are interconnected.

Images