Multi-task cooperative optimization scheduling method in low-altitude inspection of unmanned aerial vehicle

Abstract

The application discloses a multi-task cooperative optimization scheduling method in low-altitude inspection of unmanned aerial vehicles, and comprises the following steps: S1, defining general survey tasks and potential review tasks, constructing a task directed graph, and inserting a reserved period and a geographical standby area in an original task sequence of the unmanned aerial vehicles to form a buffer area; S2, based on the task directed graph, performing initial scheduling through a first stage and a second stage, and generating a candidate execution plan; S3, according to the candidate execution plan, establishing a buffer area in a selected set of the unmanned aerial vehicles, calculating a matching value, and distributing the review tasks to the most matched unmanned aerial vehicle; S4, determining a dispatched unmanned aerial vehicle and a newly activated review task according to the matching value, judging a state category in which the dispatched unmanned aerial vehicle is currently located, and performing conflict detection and priority adjustment on the dispatched unmanned aerial vehicle; and S5, monitoring key indicators and dynamically adjusting a coefficient. Through the dynamic distribution step of the unmanned aerial vehicles, real-time matching of heterogeneous unmanned aerial vehicle resources and dynamic task chains is realized.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) technology, and specifically to a multi-task collaborative optimization scheduling method for low-altitude UAV inspection. Background Technology

[0002] With the widespread application of drones in fields such as power line inspection, pipeline inspection, and disaster reconnaissance, wide-area, efficient, and autonomous collaborative operations based on multiple drones have become a research hotspot. Currently, a typical drone inspection task process usually includes two stages: the first stage is a wide-area rapid general survey, using payloads such as visible light to identify potential anomalies, i.e., suspected points; the second stage is a targeted and detailed review, using payloads such as infrared and hyperspectral imaging to conduct close-up detailed inspections of suspected points to confirm the problem.

[0003] For example, Chinese patent application CN 121143381 A discloses a method, apparatus, electronic device, and storage medium for low-altitude collaborative inspection of unmanned aerial vehicles (UAVs) based on data acquisition. This method rapidly assembles a formation after analyzing the task area and achieves parallel path coverage through unified scheduling, significantly increasing the inspection area per unit time. In multi-UAV collaborative mode, each UAV shares environmental perception information and dynamically adjusts its flight path to avoid duplicate flights and path conflicts, further optimizing resource utilization. Therefore, this method effectively solves the problem of low inspection efficiency in existing technologies, achieving efficient, rapid, and complete coverage of large areas.

[0004] However, most current scheduling systems typically allocate all tasks at once before the mission begins, including known survey points and yet-to-be-defined verification points. However, in actual operations, the number, location, and urgency of verification points can only be partially determined after the first round of surveys is completed. This results in pre-reserved UAV resources for verification tasks being idle for extended periods or frequently returning to base awaiting new instructions, leading to resource waste and inefficiency. Furthermore, the requirements for UAV platforms and flight modes differ significantly between survey and verification tasks. Existing methods often employ homogeneous UAV swarms or simple division of labor, failing to match the most suitable heterogeneous UAV resources in real-time based on dynamically generated task chains, thus limiting overall operational efficiency. Summary of the Invention

[0005] To address this issue, the present invention provides a multi-task collaborative optimization scheduling method for low-altitude UAV inspections, which solves the problem in the prior art that the most suitable heterogeneous UAV resources cannot be matched in real time according to the dynamically generated task chain, thus limiting the overall operational efficiency.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A multi-task collaborative optimization scheduling method for low-altitude UAV inspection includes the following steps:

[0008] S1: Define the inspection tasks as general survey tasks and potential review tasks according to preset rules, construct a task directed graph, and insert reserved time periods and geographical standby areas into the original task sequence of the UAV to form a buffer.

[0009] S2: Based on the task directed graph, initial scheduling is performed by combining the first and second phases, and candidate execution plans are generated;

[0010] S3: Based on the candidate execution plan in S2, plan and establish a buffer zone for drones in the drone candidate set, and calculate the matching value. The review task will be dynamically assigned to the most suitable drone.

[0011] Matching value The calculation formula is as follows:

[0012]

[0013] in, For the review of newly produced goods, The coordinates of the center of the geographic waiting area in the buffer zone. To verify the feasible flight distance between the mission point and the center of the buffer zone, To verify the time difference between the task trigger time and the midpoint of the reserved time period in the buffer, For load adaptability, , and All are preset weighting coefficients;

[0014] S4: Based on the matching value Identify the assigned drone and its corresponding newly activated review task, and determine the current status category of the assigned drone; then perform conflict detection and priority adjustment on the assigned drone;

[0015] S5: During task execution, continuously monitor key indicators, and then dynamically adjust the trade-off coefficients in the objective function of S2 and the matching values ​​in S3 based on the key indicators. The preset weighting coefficients.

[0016] Furthermore, the nodes in the directed graph represent all census tasks and all potential review tasks, with directed edges pointing from census task nodes to potential review task nodes that they may trigger.

[0017] Furthermore, the first stage decision-making allocates resources and path schemes for the census task, while the second stage embeds the response performance of possible future review tasks into the objective function in the form of expected values, in order to generate candidate execution plans for potential review tasks.

[0018] Furthermore, the specific content of S2 is as follows:

[0019] 1) Calculate the probability value of triggering potential review tasks associated with the census task after the census task is performed, based on historical inspection data. The calculation formula is as follows:

[0020]

[0021] in, This represents the number of times the review task was triggered in the past N census tasks.

[0022] 2) Based on the first-phase plan, calculate each... Confirmation delay when a potential review task associated with a directed edge is actually triggered Meanwhile, confirmation of delay will be made. Select the top 10 drones in ascending order to form a drone candidate set, and confirm the delay. The calculation formula is as follows:

[0023]

[0024] in, This refers to the time when the drone departs for the verification task point after completing the census task. The time it takes for the drone to fly to the location of the verification task. The estimated completion time of the census task that triggers the review task;

[0025] 3) Generate candidate execution plans, including a set of UAV candidates, using an objective function as follows:

[0026] in, A collection of census tasks. for The set, For the first A census task, To execute The comprehensive costs incurred at that time The preset tradeoff coefficients, For the first One review task.

[0027] Furthermore, S3 includes the following steps:

[0028] S3.1: Identify the location, payload, and remaining battery power of each drone in the candidate drone set of the candidate execution plan;

[0029] S3.2: Associate a dynamic task queue with each UAV buffer. The dynamic task queue is initially empty. When a census task is completed during the inspection process and a review task is actually triggered, the review task is assigned to the dynamic task queue.

[0030] S3.3: Calculate the matching value of the dynamic task queue, which is used to match the review task to the dynamic task queue of the optimal buffer, and select the drone with the closest drone location, the best matching load, and the most remaining power as the response unit of the potential review task.

[0031] Furthermore, the specific steps of S4 are as follows:

[0032] S4.1: Retrieve based on matching value Assigned drones and review tasks, and determine the current status category of the assigned drones;

[0033] S4.2: The newly activated review task is treated as a new node and inserted into the dynamic task queue of the dispatched drone according to the preset importance level and urgency, and conflict detection is performed on the queue.

[0034] S4.3: Taking the current location, real-time battery level, payload configuration, and updated dynamic task queue of the dispatched UAV as input, and minimizing the total time to complete all tasks in the dynamic task queue as the objective, replan the local task execution sequence for the dispatched UAV starting from the current moment.

[0035] Furthermore, the current state of the drone includes: being in its reserved time period and on standby in the associated geographic standby area; performing a census task or a previously assigned dynamic review task in its original mission sequence; or being in other states such as flight transfer, return, or charging.

[0036] Furthermore, the aforementioned comprehensive cost The calculation formula is as follows:

[0037] in, For drones in execution During the mission, fly from the previous location to The actual feasible path length, For the drone's cruising speed, This represents the energy consumption per unit distance of flight for the drone. This represents the energy consumption per unit hovering time of the drone. In order to be in Task point hovering operation time. Weighted by time cost, Energy consumption cost weighting.

[0038] This invention has the following advantages: By introducing the concept of a buffer consisting of a reserved time period and a geographical standby area into the initial scheduling, and embedding the expected response delay optimization for potential review tasks into the objective function, the system reserves flexible response capabilities for uncertain tasks during the planning stage, avoiding resource idleness or ineffective round trips caused by UAVs simply waiting for possible review instructions in the early stages of operation. When a review task is dynamically activated, the rapid allocation mechanism based on matching values ​​and priority adjustment can achieve near real-time task assignment and local optimization of the execution sequence.

[0039] Meanwhile, in the dynamic allocation of drones, the calculation of the matching value takes into account key factors such as distance, time window and payload suitability, so that newly activated review tasks can be intelligently matched with heterogeneous drones that are closest in location, have the most suitable payload and the most abundant remaining power, thus realizing real-time matching of heterogeneous drone resources and dynamic task chains.

[0040] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. Attached Figure Description

[0041] To more intuitively illustrate the prior art and this application, exemplary drawings are provided below. It should be understood that the specific shapes and structures shown in the drawings should not generally be regarded as limiting conditions for implementing this application; for example, based on the technical concept disclosed in this application and the exemplary drawings, those skilled in the art are able to easily make conventional adjustments or further optimizations to the addition / reduction / classification, specific shapes, positional relationships, connection methods, size ratios, etc. of certain units (components).

[0042] Figure 1 This is a flowchart illustrating the implementation of the multi-task collaborative optimization scheduling method for low-altitude UAV inspection according to the present invention. Detailed Implementation

[0043] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. It should be understood that these embodiments are merely for further explanation of the present invention and should not be construed as limiting the scope of protection of the present invention. Technical engineers in the field can make some non-essential improvements and adjustments to the present invention based on the above-described content. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] Please see Figure 1A multi-task collaborative optimization scheduling method for low-altitude UAV inspection includes the following steps:

[0045] S1: Inspection tasks are defined as census tasks and potential review tasks according to preset rules. A directed task graph is constructed, and reserved time slots (e.g., 10:15–10:30) and geographical standby areas (e.g., a low-altitude airspace with a radius of 500 meters centered on a certain tower) are inserted into the original UAV task sequence to form a buffer zone. The directed task graph includes known census task nodes, potential review task nodes, and directed edges representing the generation of review tasks from census tasks. This facilitates the representation of dynamic probabilistic dependencies between tasks, providing a unified data foundation and analysis framework for subsequent scheduling.

[0046] The census task includes the geographic coordinates of the task, the estimated operation time, the predefined importance level, and the static known information of the required payload type, such as a visible light camera.

[0047] The activation of a potential review task depends on the execution result of the census task. If an anomaly is found, the review task is initiated. The dedicated load and operation time required for the review task are associated with the census task that triggered the review task through predefined rules.

[0048] A task-directed graph is constructed using census tasks and potential review tasks. Nodes in the graph represent all census tasks and all potential review tasks. Directed edges point from census task nodes to potential review task nodes that they may trigger.

[0049] During construction, all census task nodes are first generated based on the known list of inspection areas. Then, a verification task node that may be triggered is created for each census task through a preset knowledge base (such as "X% probability that a certain type of equipment inspection requires infrared verification"). The generation probability and associated constraints are labeled in the form of weights, and finally, a complete graph data structure is output.

[0050] S2: Based on the task directed graph, initial scheduling is performed through a two-stage decision-making process. The first stage allocates resources and path schemes for the census task. The second stage embeds the expected performance of future review tasks into the objective function in the form of expected values ​​to generate candidate execution plans for potential review tasks. The specific content of S2 is as follows:

[0051] 1) Calculate the probability value of triggering potential review tasks associated with the census task after the census task is performed, based on historical inspection data. The calculation formula is as follows:

[0052]

[0053] in, This represents the number of times a review task was triggered in the past N surveys; for example, if a certain type of transmission tower was found to have suspected defects requiring infrared review in 24 out of the past 200 surveys, then... =24 / 200=0.12.

[0054] 2) Based on the first-phase plan, calculate each... Confirmation delay when a potential review task associated with a directed edge is actually triggered Confirmation delay This refers to the delay from the completion of the census task it relies on to the arrival of the drone and the commencement of the verification task. The delay will also be confirmed. Select the top 10 drones in ascending order to form a drone candidate set. Confirmation delay. The calculation formula is as follows:

[0055]

[0056] in, This refers to the time when the drone departs for the verification task point after completing the census task. The time it takes for the drone to fly to the location of the verification task. The estimated completion time of the census task that triggers the review task.

[0057] 3) After completing the above parameter preparation, a candidate execution plan including a set of UAV candidates is generated through the objective function. This objective function drives the optimization algorithm to find a scheduling scheme, which can minimize the overall expected response delay of future review tasks while efficiently completing the deterministic census task.

[0058] The objective function is as follows:

[0059] in, A collection of census tasks. for The set, For the first A census task, To execute The comprehensive costs incurred at that time The preset tradeoff coefficients, For the first One review task. Overall cost. The calculation formula is as follows:

[0060] in, For drones in execution During the mission, fly from the previous location (base or above the mission point) to... The actual feasible path length. For the drone's cruising speed, This represents the energy consumption per unit distance of flight for the drone (unit: Wh / m). This represents the energy consumption per unit hovering time of the drone (unit: Wh / s). In order to be in Task point hovering operation time. Weighted by time cost, Energy consumption cost weighting.

[0061] Energy consumption per unit distance of flight The calculation formula is as follows:

[0062] in, For the distance the drone flies in a straight line at a constant cruising speed Total energy consumption.

[0063] Energy consumption per unit hovering time The calculation formula is as follows:

[0064]

[0065] in, This represents the real-time power output of the battery during flight. The unit is W, 1Wh = 3600W·s, so power is converted to Wh / s by dividing by 3600.

[0066] S3: Based on the candidate execution plan in S2, plan and establish a buffer zone for drones in the drone candidate set, and calculate the matching value. The review task will be dynamically assigned to the most suitable drone.

[0067] S3 includes the following steps:

[0068] S3.1: Identify the location, payload, and remaining battery power of each drone in the candidate drone set of the candidate execution plan.

[0069] S3.2: Associate a dynamic task queue with each UAV buffer. The dynamic task queue is initially empty. When a survey task is completed during the inspection process and a review task is actually triggered, the review task is assigned to the dynamic task queue.

[0070] S3.3: Calculate the matching value of the dynamic task queue, used to match review tasks to the dynamic task queue in the optimal buffer, selecting the drone with the closest drone location, the best payload match, and the most remaining battery power as the response unit for potential review tasks. Matching value The calculation formula is as follows:

[0071]

[0072] in, For the review of newly produced goods, The coordinates of the center of the geographic waiting area in the buffer zone. To verify the feasible flight distance between the mission point and the center of the buffer zone, This is to verify the time difference between the task trigger time and the midpoint of the reserved time period in the buffer. For load fit, if the required load for the verification task is the same as the load of the UAV, the load fit is... The value is 1 if it is not 0 otherwise. , and All of these are preset weighting coefficients.

[0073] S4: Based on the matching value Identify the assigned drone and its corresponding newly activated review task, and determine the current status category of the assigned drone; then perform conflict detection and priority adjustment on the assigned drone.

[0074] The specific steps for S4 are as follows:

[0075] S4.1: Retrieve based on matching value Assigned drones and review tasks, and determine the current status category of the assigned drones;

[0076] The current status of the drone includes: 1) being in its reserved time slot and on standby in the associated geographic standby area; 2) performing a census task or a previously assigned dynamic review task in its original mission sequence; 3) being in other states such as flight transfer, return, or charging.

[0077] S4.2: Treat newly activated review tasks as new nodes, insert them into the dynamic task queue of the dispatched drone according to their corresponding preset importance level and urgency, and perform conflict detection on the queue.

[0078] Conflict detection is used to check for constraints such as overlapping time windows or mismatch between the required payload of a task and the current payload of the UAV. For tasks with conflicts, the queue order is adjusted according to preset priority rules, or tasks that cannot meet the constraints are marked as temporarily infeasible.

[0079] S4.3: Taking the current location, real-time battery level, payload configuration, and updated dynamic task queue of the dispatched UAV as input, and minimizing the total time to complete all tasks in the queue, replan the local task execution sequence for the dispatched UAV starting from the current moment.

[0080] S5: During task execution, continuously monitor key indicators, and then dynamically adjust the trade-off coefficients in the objective function of S2 and the matching values ​​in S3 based on the key indicators. The preset weighting coefficients.

[0081] For example, if the average delay of review tasks is found to be too high, then the time should be increased appropriately. or The weighting makes scheduling more inclined to respond quickly to review tasks.

[0082] This invention introduces the concept of a buffer zone consisting of reserved time periods and geographical standby areas into the initial scheduling, and embeds optimization of expected response delays for potential review tasks into the objective function. This allows the system to reserve flexible response capabilities for uncertain tasks during the planning phase, avoiding resource idleness or ineffective round trips by UAVs in the early stages of operations due to simply waiting for possible review instructions. When a review task is dynamically activated, a rapid allocation mechanism based on matching values ​​and priority adjustment can achieve near real-time task assignment and local optimization of the execution sequence.

[0083] Meanwhile, in the dynamic allocation of drones, the calculation of the matching value takes into account key factors such as distance, time window and load adaptability, so that newly activated review tasks can be intelligently matched with the heterogeneous drones that are closest in location, have the most suitable load and the most abundant remaining power. This achieves real-time matching of heterogeneous drone resources and dynamic task chains, overcoming the efficiency bottleneck of the fixed division of labor mode.

[0084] An example is given below:

[0085] Take the drone inspection of power transmission lines in a certain area as an example. Before the inspection, it was known that 100 towers needed to undergo visible light surveys. According to the historical knowledge base, 30 towers located in old areas had a 20% probability of requiring infrared verification after the survey.

[0086] First, execute S1 to construct a directed task graph. Generate 100 survey task nodes (attributes include coordinates, operation duration, importance level, and visible light payload), and create corresponding potential review task nodes (infrared payload) for each of the aforementioned 30 towers, forming directed edges from the survey nodes to the review nodes, with edge weights containing probability p=0.2. Insert multiple reserved time slots (e.g., 10:00-10:15) and corresponding geographical standby areas (centered around several key towers) into the initial task sequence of each participating UAV.

[0087] Next, S2 is executed for initial scheduling. The first phase assigns execution paths and times for 100 census tasks to the four drones. The second phase calculates the expected confirmation delay for each potential review task node and optimizes the overall scheme using an objective function. Finally, candidate execution plans are generated, specifying the top few drones in the candidate set that respond fastest if a specific tower (e.g., tower #15) triggers a review (e.g., infrared drone #2 ranks first due to its reserved time slot and geographical standby area being close to #15).

[0088] After the operation began, the drone carried out the survey as planned. Assuming that at 09:50, drone #1 completed the visible light survey of tower #15 and, through real-time image analysis, discovered a suspected hotspot, then activated the corresponding infrared verification task.

[0089] The system immediately executes S3, calculating the matching value between the newly activated review task and each UAV buffer based on the latest status. The calculation reveals that infrared UAV #2 is currently within its reserved time slot, waiting in its geographic standby area, and has a payload fit of 1, resulting in the optimal matching value. Therefore, the system assigns this review task to the dynamic task queue associated with UAV #2.

[0090] Subsequently, S4 was triggered, and the system determined that Drone #2 was in the "standby during the reserved time period" state. The verification task was inserted into its queue as a high-priority node, and no conflict was detected. It was instructed to immediately fly from its current standby point to tower #15 to perform infrared verification, and upon completion, return to its original geographical standby area or act according to new instructions. The entire process was completed within seconds; the response time for Drone #2 to switch from standby to performing the verification task was extremely short.

[0091] Throughout the day, the system continuously executes S5 to monitor various key indicators. For example, if the average confirmation delay of infrared verification tasks is found to have increased in the afternoon, the system will automatically fine-tune the trade-off coefficient, making subsequent incremental scheduling decisions more inclined to prioritize ensuring verification response speed, thereby dynamically maintaining overall operational efficiency.

[0092] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-task collaborative optimization scheduling method for low-altitude UAV inspection, characterized in that, Includes the following steps: S1: Define the inspection tasks as general survey tasks and potential review tasks according to preset rules, and construct a task directed graph. Insert reserved time periods and geographical standby areas into the original task sequence of the UAV to form a buffer. S2: Based on the task directed graph, initial scheduling is performed by combining the first and second phases, and candidate execution plans are generated; S3: Based on the candidate execution plan in S2, plan and establish a buffer zone for drones in the drone candidate set, and calculate the matching value. The review task will be dynamically assigned to the most suitable drone. Matching value The calculation formula is as follows: in, For the review of newly produced goods, The coordinates of the center of the geographic waiting area in the buffer zone. To verify the feasible flight distance between the mission point and the center of the buffer zone, To verify the time difference between the task trigger time and the midpoint of the reserved time period in the buffer, For load adaptability, , and All are preset weighting coefficients; S4: Based on the matching value Identify the assigned drone and the newly activated review task, and determine the current status category of the assigned drone; then perform conflict detection and priority adjustment on the assigned drone; S5: During task execution, continuously monitor key indicators, and then dynamically adjust the trade-off coefficients in the objective function of S2 and the matching values ​​in S3 based on the key indicators. The preset weighting coefficients.

2. The multi-task collaborative optimization scheduling method for low-altitude UAV inspection according to claim 1, characterized in that, The nodes in the directed graph represent all census tasks and all potential review tasks, with directed edges pointing from census task nodes to potential review task nodes that they may trigger.

3. The multi-task collaborative optimization scheduling method for low-altitude UAV inspection according to claim 1, characterized in that, The first stage decision-making allocates resources and path schemes for the census task. The second stage embeds the response performance of possible future review tasks into the objective function in the form of expected values, which is used to generate candidate execution plans for potential review tasks.

4. The multi-task collaborative optimization scheduling method for low-altitude UAV inspection according to claim 1, characterized in that, The specific content of S2 is as follows: 1) Calculate the probability value of triggering potential review tasks associated with the census task after the census task is performed, based on historical inspection data. The calculation formula is as follows: in, This represents the number of times the review task was triggered in the past N census tasks. 2) Based on the first-phase plan, calculate each... Confirmation delay when a potential review task associated with a directed edge is actually triggered Meanwhile, confirmation of delay will be made. Select the top 10 drones in ascending order to form a drone candidate set, and confirm the delay. The calculation formula is as follows: in, This refers to the time when the drone departs for the verification task point after completing the census task. The time it takes for the drone to fly to the location of the verification task. The estimated completion time of the census task that triggers the review task; 3) Generate candidate execution plans, including a set of UAV candidates, using an objective function as follows: in, A collection of census tasks. for The set, For the first A census task, To execute The comprehensive costs incurred at that time The preset tradeoff coefficients, For the first One review task.

5. The multi-task collaborative optimization scheduling method for low-altitude UAV inspection according to claim 1, characterized in that, S3 includes the following steps: S3.1: Identify the location, payload, and remaining battery power of each drone in the candidate drone set of the candidate execution plan; S3.2: Associate a dynamic task queue with each UAV buffer. The dynamic task queue is initially empty. When a census task is completed during the inspection process and a review task is actually triggered, the review task is assigned to the dynamic task queue. S3.3: Calculate the matching value of the dynamic task queue, which is used to match the review task to the dynamic task queue of the optimal buffer, and select the drone with the closest drone location, the best matching load, and the most remaining power as the response unit of the potential review task.

6. The multi-task collaborative optimization scheduling method for low-altitude UAV inspection according to claim 1, characterized in that, The specific steps of S4 are as follows: S4.1: Retrieve based on matching value Assigned drones and review tasks, and determine the current status category of the assigned drones; S4.2: The newly activated review task is treated as a new node and inserted into the dynamic task queue of the dispatched drone according to the preset importance level and urgency, and conflict detection is performed on the queue. S4.3: Taking the current location, real-time battery level, payload configuration, and updated dynamic task queue of the dispatched UAV as input, and minimizing the total time to complete all tasks in the dynamic task queue as the objective, replan the local task execution sequence for the dispatched UAV starting from the current moment.

7. The multi-task collaborative optimization scheduling method for low-altitude UAV inspection according to claim 6, characterized in that, The current status of the drone includes: being within its reserved time period and on standby in the associated geographic standby area; performing a census task or a previously assigned dynamic review task in its original mission sequence; or being in other states such as flight transfer, return, or charging.

8. The multi-task collaborative optimization scheduling method for low-altitude UAV inspection according to claim 4, characterized in that, The comprehensive cost The calculation formula is as follows: in, For drones in execution During the mission, fly from the previous location to The actual feasible path length, For the drone's cruising speed, This represents the energy consumption per unit distance of flight for the drone. This represents the energy consumption per unit hovering time of the drone. In order to be in Task point hovering operation time. Weighted by time cost, Energy consumption cost weighting.

Images