A planning method for unmanned aerial vehicle monitoring operation and maintenance of power transmission lines

Abstract

The application relates to a planning method for monitoring and operating a power transmission line by using a UAV, and relates to the technical field of UAV planning. A scene model of a scene where a power transmission line is located is created; a UAV is modeled; a planner is constructed by using a language model; the planner is controlled by a prompt word to generate an action of controlling the UAV to go to a specified power transmission line region in the scene model according to current and previous UAV states, a scene model, previous UAV group actions and filtering experience pool experience; the planner is controlled to optimize a local inspection route of the UAV in each power transmission line region for the purpose of minimizing inspection flight cost; an unreasonable task action and an inspection path formed by planner hallucination are filtered through a planning filter to avoid behaviors that do not conform to safety or reality, and filtering experience generated by the planning filter is provided to the language model to allow the language model to continuously improve decision-making based on the filtering experience.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) inspection and scheduling technology, and in particular to a planning method for monitoring and maintaining power transmission lines using UAVs. Background Technology

[0002] With the rapid development of drone technology, drone inspection, as an emerging technology, has been widely used in the field of power transmission line monitoring and maintenance.

[0003] Existing drone inspection systems typically include a drone platform, a ground control station, and image recognition modules. The drone platform, equipped with sensors such as high-definition cameras, inspects power transmission lines via preset routes or manual control. The ground control station is responsible for tasks such as drone takeoff, landing, route planning, and data transmission. The image recognition module is used for automatic identification and anomaly detection of images captured by the drone.

[0004] However, existing drone inspection systems still have some problems and shortcomings. First, regardless of whether normal conditions are observed or abnormalities occur, the drone inspection path planning is relatively fixed. When a drone discovers an anomaly near a power transmission line during inspection, it typically needs to record and transmit the anomaly type and complete the subsequent inspection path. The entire drone system lacks an intelligent linkage mechanism for anomalies, requiring manual intervention in drone inspection control based on the recorded anomalies. The flexibility of intelligent planning and anomaly adaptability are insufficient. To address these issues, some researchers have proposed power transmission line inspection systems based on multi-drone collaboration. These systems, by introducing multi-drone collaboration technology, achieve information sharing and collaborative operation among drones. However, these systems still have some shortcomings. For example, in terms of anomaly detection and response, although information sharing among drones is achieved, the response mechanism after anomaly detection remains relatively simple, lacking intelligent path planning and autonomous control capabilities.

[0005] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] To solve the above-mentioned technical problems, or at least partially solve them, the present invention provides a planning method for monitoring and maintaining power transmission lines using unmanned aerial vehicles (UAVs).

[0007] This invention provides a planning method for monitoring and maintaining power transmission lines using unmanned aerial vehicles (UAVs), comprising: Use grid maps to create digital scene models that describe the environment in which power transmission lines are located; Modeling of UAVs used for monitoring and maintaining power transmission lines is conducted from four aspects: function, operating status, energy consumption, and battery. A planner is constructed using a language model. The planner is controlled by prompt words to generate conditional probabilities for controlling the drone to move to a specified power transmission line area in the scene model based on the current and previous drone status, scene model, the actions of each drone in the previous drone group, and the experience of the filtered experience pool. The planner optimizes the local inspection route of the drone inspection in each power transmission line area with the goal of minimizing the inspection flight cost. By using a reinforcement learning-based planning filter to filter out unreasonable task actions and inspection paths generated by planner illusions, unsafe or unrealistic behaviors are avoided. The filtering experience generated by the planning filter is provided to the language model, allowing the language model to refer to past filtering experience to continuously improve decision-making.

[0008] Furthermore, the scene model's grid map contains several grids, each configured with the following attributes: altitude, no-fly zone, entity attribute, and entity state attribute. Environmental information is represented through the grid and its attributes. Specifically, the altitude attribute describes the altitude of the corresponding environmental entity, and the grid's location and altitude attribute fully describe the structure of the entity. The no-fly zone attribute describes whether the corresponding environmental area is no-fly zone. The entity attribute describes the type of the corresponding environmental entity. The entity state attribute describes the state of the corresponding entity.

[0009] Furthermore, the modeling of UAVs used for power transmission line monitoring and maintenance is conducted from four aspects: functionality, operational status, energy consumption, and battery. Specifically, the UAV model's functional description attributes describe the power transmission line monitoring and maintenance functions supported by the UAV; the UAV's operational status attributes describe its location and whether it is charging, operating, or idle; energy consumption modeling is performed for hovering and propulsion states; the battery model includes battery discharge power and SOC constraints; and based on energy conservation and energy conversion efficiency... x=1,2, where, For battery energy conversion efficiency, This refers to the battery discharge power. For non-driving power consumption, when x=1, For the power of the drone in hovering state, when x=2, Power for drone propulsion.

[0010] Furthermore, the power output in hovering mode is: ; in, This represents hovering drag power, caused by the frictional drag between the blades and the air during rotor rotation, and is related to the blade drag coefficient. air density Number of leaves N, leaf chord length Rotation speed cubic and blade radius The fourth power is directly proportional; This represents the hovering induced power, which is the energy required to generate a reaction force by accelerating the air downwards to counteract gravity. mg is the weight of the drone. This is the induced power correction factor.

[0011] Furthermore, the power during propulsion is: ; in, To improve the type resistance power correction factor compared to the hovering state, To improve the induced power correction factor compared to the hovering state, For hovering induced speed; The drag power during level flight is the power required to overcome the aerodynamic drag of the UAV during level flight, and is related to the drag coefficient. air density Number of leaves N, leaf chord length blade radius It is proportional to the cube of the velocity.

[0012] Furthermore, the prompt content includes: The meaning description of the scene model and the current scene model; you are required to understand the information of the scene in which the transmission line is located based on the provided scene model and the meaning description of the scene model; Drone location, drone status, and understanding the drone's situation based on its location and status; Filter experience pool; Responsible for drone planning in a scenario, the task is to determine the drone's mission actions based on experience from the filtered experience pool, the current scenario model, the drone's position and status; and to ensure that the selected mission actions comply with no-fly zones and drone energy constraints. Based on the experience of the filtration experience pool, the current situation of the power transmission line area, and the status of the drone, the inspection route is planned to minimize the inspection flight cost; ensure that the planned route complies with no-fly zones and drone energy constraints; If the task actions and paths may violate any constraints, return an empty decision instead of an unsafe one; and interact with the planning filter to filter the task actions and paths.

[0013] Furthermore, reinforcement learning-based planning filters include critic networks and actor networks; During training, the critic network and actor network are trained according to drone planning. Drone planning is formulated as a constrained Markov decision process, defined as follows: (S, A, r, P, γ, Ω), where S represents the state space, including: scene model, drone position, battery level, drone currently assigned to the power line area, and cumulative inspection distance; A represents the action space, including commands such as assigning the drone to the power line area, returning to the drone station, switching to charging and idle states; r is the reward function based on state and action; P corresponds to the state transition probability; γ is the discount factor; and Ω is a set of constraints that the drone must satisfy. Constraints are the operational requirements that force the drone to perform, and a cost function is constructed for each constraint. The planning filter achieves reward maximization and constraint compliance through the following Lagrange dual optimization: ; in, The decision made is represented by λ ≥ 0 as the Lagrange multiplier. This is a safety threshold; Actor Network Generated Motion Critics network assesses expected return and related constraint costs The policy gradient update is as follows: ; Where θ is the strategy parameter; The actions formed by the rule filter interact with the scene model and the drone model to generate a new state space, and (S, A, r, P, γ, Ω) are stored in the filter experience pool.

[0014] Furthermore, during training, if the cost of the action selected by the actor network exceeds a safety threshold, a safe alternative action will replace the original action.

[0015] Furthermore, during the filtering process, the actions and paths generated by the planner are directly evaluated by a trained commentator network, which assesses the associated constraint costs. The action and path generated by the planner are compared with a safety threshold to achieve filtering.

[0016] Furthermore, the experience corresponding to the filtered action paths is stored in the filtering experience pool, and the planner makes decisions and improvements based on the experience in the filtering experience pool.

[0017] The technical solutions provided in the embodiments of the present invention have the following advantages compared with the prior art: This application creates a scene model of the power transmission line location; models the drone; constructs a planner using a language model, and controls the planner to generate actions to guide the drone to a specified power transmission line area in the scene model based on the current and previous drone states, the scene model, the actions of each drone in the previous drone group, and the experience of the filtering experience pool, through prompts. The planner optimizes the local inspection route of the drone inspection in each power transmission line area with the goal of minimizing the inspection flight cost; and filters unreasonable task actions and inspection paths generated by the planner's illusions through a planning filter to avoid unsafe or unrealistic behaviors, and provides the filtering experience generated by the planning filter to the language model, allowing the language model to continuously improve its decisions based on the filtering experience. This application uses a reinforcement learning-based planning filter to filter unreasonable task actions and inspection paths generated by the planner's illusions to avoid unsafe or unrealistic behaviors, and provides the filtering experience generated by the planning filter to the language model, allowing the language model to refer to past filtering experience to continuously improve its decisions. Combining the generalization advantages of the language model with the adaptability and real-time learning capabilities of reinforcement learning, it ensures intelligent generation of drone planning in the dynamic environment of power transmission line inspection. Attached Figure Description

[0018] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of a planning method for monitoring and maintaining power transmission lines using unmanned aerial vehicles (UAVs) according to an embodiment of the present invention.

[0021] Figure 2 This is a schematic diagram of a scene model provided in an embodiment of the present invention.

[0022] Figure 3 This is a schematic diagram of the network architecture provided in an embodiment of the present invention.

[0023] Figure 4 This is a schematic diagram of a planning device for monitoring and maintaining power transmission lines using unmanned aerial vehicles (UAVs) according to an embodiment of the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. 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.

[0025] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0026] Example 1 See Figure 1 As shown, to address the problem that conventional fixed-path planning cannot adapt to dynamic environments and abnormal situations, this invention provides a planning method for monitoring and maintaining power transmission lines using unmanned aerial vehicles (UAVs), comprising: A digital scene model is created to describe the environment in which the power transmission line is located. This environment includes the power transmission line itself, power transmission equipment, drone stations, and environmental entities that may obstruct drone operations. The drone stations are equipped with drone charging facilities for drones to charge after landing. An example scene model is created based on a grid map. The grid map contains several grids, and each grid is configured with the following attributes to better describe the environment: altitude, no-fly zone, entity attribute, and entity status attribute. The environment information is represented through the grids and their attributes.

[0027] Specifically, the height attribute configured for a grid describes the altitude information of the environmental entity corresponding to the grid. The grid position and its height attribute fully describe the structure of the environmental entity, which is of great significance for UAV planning and obstacle avoidance. The no-fly zone attribute configured for a grid describes whether the environmental area corresponding to the grid is no-fly zone. The entity attribute configured for a grid describes the type of the environmental entity corresponding to the grid. Entity attributes can effectively describe the area where power transmission lines, power transmission equipment, and UAV stations are located. In specific implementation, entity attributes are used to divide the scene where the power transmission line is located into multiple power transmission line areas containing the points to be measured. For any power transmission line area i, its contents are set to include... There are 10 test points; the entity state attributes configured in the grid can describe the state of the entity corresponding to the grid, especially entities that can interact with drones, such as whether the charging station entity in the drone station is occupied, whether the power transmission line has completed the inspection, and the abnormal state of the power transmission equipment of the power transmission line.

[0028] Compared to traditional grid maps, the scene model in this application introduces new attributes to the grid. In practice, the scene model is represented by a multi-channel high-dimensional matrix, where the height and width of the matrix correspond to the grid arrangement. Each channel represents a specific attribute: one channel represents altitude, one channel represents no-fly zone status, one channel represents entity type, and multiple channels represent entity status. In practice, no-fly zone status and entity status are represented by 0 and 1 respectively, while entity type and entity status are encoded numerically. The matrix-based scene model can be understood and perceived by a language model used as a planner.

[0029] In the scenario of power transmission line monitoring and maintenance, various drones with different functions may be used. In order to achieve intelligent planning and scheduling, drones are modeled.

[0030] The constructed drone model includes the following: The functional description attribute of the drone model describes the power transmission line monitoring and maintenance functions supported by the drone. When scheduling drones based on anomalies, drones with corresponding functional description attributes are scheduled according to the abnormal state in the scenario model.

[0031] Drone operational status attributes describe the drone's location and whether it is charging, operating, or idle.

[0032] For a system containing M drones, construct an M-dimensional, three-channel tensor to model the functional description and operational status of each drone. One channel consists of the functional description encoding of the M drones, one channel consists of the position encoding of the M drones, and one channel consists of the charging, working, or idle status encoding of the M drones.

[0033] To accurately describe the energy consumption of drones under different operating conditions, this application divides the operating conditions into two states: hovering and propulsion, and models the energy consumption for each state.

[0034] The power model in hovering state is: ; in, This represents hovering drag power, caused by the frictional drag between the blades and the air during rotor rotation, and is related to the blade drag coefficient. air density Number of leaves N, leaf chord length Rotation speed cubic and blade radius The fourth power is proportional to the energy consumption, reflecting the influence of air, rotor geometry, and rotational speed on energy consumption during hovering.

[0035] This represents the hovering induced power, which is the energy required to generate a reaction force by accelerating the air downwards to counteract gravity. mg is the weight of the drone. is the induced power correction coefficient, used to correct for losses in non-ideal flow fields, k∈[0.1,0.3].

[0036] Power during propulsion is: ; in, To improve the shape drag power correction factor compared to the hovering state, and to correct the rotor's shape drag power during level flight, the effect of forward speed on blade load distribution is considered. At low speeds, , The speed in the middle is ignored, becoming The drag power is consistent with that in the hovering state. At high speeds, the drag power increases significantly with increasing speed, and the drag increases nonlinearly due to local speed differences in the blades.

[0037] To improve the induced power correction factor compared to the hovering state, the change of induced power with speed v during UAV level flight is corrected based on the forward ratio theory. This is the hovering-inducing speed. At low speeds... , Approximately degenerates into hovering induced power At high speed The induced power decreases as the speed v increases because the forward speed reduces the airflow that needs to be accelerated.

[0038] This represents the power required to overcome aerodynamic drag during level flight, specifically the power needed for the UAV to overcome its fuselage aerodynamic drag. It is related to the fuselage drag coefficient. air density Number of leaves N, leaf chord length blade radius It is proportional to the cube of the velocity.

[0039] Battery Model: The drone is powered by a battery, and the battery discharge power follows: ;in, The open-circuit voltage associated with SOC. The internal resistance is related to the State of Charge (SOC); the SOC follows the following rules during battery charging and discharging: .

[0040] According to the law of conservation of energy, and based on energy conversion efficiency... x=1,2. For battery energy conversion efficiency, This refers to the battery discharge power. For non-driving power consumption, when x=1, For the power of the drone in hovering state, when x=2, Power for drone propulsion.

[0041] The drone model combines energy consumption and battery models, providing effective constraints for drone planning in power transmission line monitoring and maintenance scenarios.

[0042] A planner is constructed using language models, enabling flexible planning for UAVs in power transmission line monitoring and maintenance scenarios. Example language models include Qwen32B, gpt-4o, and deepseek.

[0043] The planner, controlled by prompts, generates conditional probabilities for controlling drones to travel to a specified power transmission line area in the scene model based on the current and previous drone states, the scene model, the actions of individual drones in the previous drone swarm, and the experience from the filtered experience pool. .in, This represents the mission actions of each drone in the preceding drone swarm before time t. The predicted task action at time t is represented by X, where X represents the current and preceding drone state, scene model, and filtered experience pool experience.

[0044] For any transmission line region i, define it as containing There are several test points, which form a permutation sequence. If the drone inspects the test points according to their order, a planner is needed to optimize the local inspection route for each power transmission line area to minimize the inspection flight cost. The inspection flight cost is calculated using a drone energy consumption model, a battery model, and the planned route.

[0045] In specific implementation, the prompts used to control the planner during planning include: "The meaning description of the scenario model and the current scenario model; the requirement is to understand the information of the scenario in which the transmission line is located based on the provided scenario model and the meaning description of the scenario model."

[0046] The drone's functions and status require understanding the drone's situation based on its functions and status. Filter experience pool; Responsible for drone planning in a given scenario. The task is to determine the drone's mission actions based on experience from the filtered experience pool, the current scenario model, the drone's location, and its state. Ensure the selected mission actions comply with no-fly zones and drone energy constraints in the scenario model. Based on experience from the filtered experience pool, the current power line area conditions, and the drone's state, plan an inspection path that minimizes the inspection flight cost, ensuring the planned path complies with no-fly zones and drone energy constraints in the scenario model. If the mission actions and path may violate any constraints, return to a null decision. Interact with the planning filter to filter the mission actions and paths. This application employs a reinforcement learning-based planning filter to filter out irrational task actions and inspection paths generated by planner illusions, thereby avoiding behaviors that are unsafe or unrealistic. The filtering experience generated by the planning filter is then provided to a language model, allowing the language model to continuously improve its decisions by referencing past filtering experiences. By combining the generalization advantages of the language model with the adaptability and real-time learning capabilities of reinforcement learning, intelligent UAV planning is ensured in the dynamic environment of power transmission line inspection.

[0047] In practice, reinforcement learning-based planning filters include critic networks and actor networks.

[0048] During training, the critic network and the actor network are trained according to the drone planning. Drone planning is formulated as a constrained Markov decision process, defined as follows: (S, A, r, P, γ, Ω), where S represents the state space, A represents the action space, r is the reward function based on state and action, P corresponds to the state transition probability, γ ∈ [0, 1] is the discount factor, and Ω is a set of constraints that the drone must satisfy. Specifically, the state space includes: the scene model, drone state, battery level, the drone currently assigned to the power line area, and the cumulative inspection distance. The action space includes commands to assign the drone to the power line area, return to the drone station, and switch between charging and idle states. The reward is calculated based on flight cost. Constraints are requirements that force the drone to perform operations, such as battery protection and route efficiency; a cost function c(s, a) is constructed for each constraint.

[0049] The planning filter achieves reward maximization and constraint compliance through the following Lagrange dual optimization: ; in, The decision made is represented by λ ≥ 0 as the Lagrange multiplier. This is the safety threshold.

[0050] Actor Network Generated Motion Critics network assesses expected return and related constraint costs The policy gradient update is performed as follows: ; Where θ is the policy parameter. If the cost of the selected action exceeds the safety threshold, a safe alternative action will replace the original action.

[0051] The actions formed by the rule filter interact with the scene model and the drone model to generate a new state space, and (S, A, r, P, γ, Ω) are stored in the filter experience pool.

[0052] During the filtering process, the actions and paths generated by the planner are directly evaluated by the commentator network for their associated constraint costs. The filtered actions are compared with a safety threshold. The (S, A, r, P, γ, Ω) corresponding to the filtered action paths are stored in the filtering experience pool.

[0053] Example 2 See Figure 4 As shown, this invention provides a planning device for monitoring and maintaining power transmission lines using unmanned aerial vehicles (UAVs), comprising: at least one processing unit, the processing unit being connected to a storage unit via a bus unit, the storage unit storing a computer program, and the computer program being executed by the processing unit to implement the multi-level power optimization allocation and equipment selection method for the energy and power system, including: Use grid maps to create digital scene models that describe the environment in which power transmission lines are located; Modeling of UAVs used for monitoring and maintaining power transmission lines is conducted from four aspects: function, operating status, energy consumption, and battery. A planner is constructed using a language model. The planner is controlled by prompt words to generate conditional probabilities for controlling the drone to move to a specified power transmission line area in the scene model based on the current and previous drone status, scene model, the actions of each drone in the previous drone group, and the experience of the filtered experience pool. The planner optimizes the local inspection route of the drone inspection in each power transmission line area with the goal of minimizing the inspection flight cost. By using a reinforcement learning-based planning filter to filter out unreasonable task actions and inspection paths generated by planner illusions, unsafe or unrealistic behaviors are avoided. The filtering experience generated by the planning filter is provided to the language model, allowing the language model to refer to past filtering experience to continuously improve decision-making.

[0054] Of course, the storage unit in the planning device for monitoring and maintaining power transmission lines provided in the embodiments of the present invention stores computer programs that are not limited to the operation of the method described above, and can also execute related operations in the planning method for monitoring and maintaining power transmission lines provided in any embodiment of the present invention.

[0055] Example 3 This invention provides a computer-readable storage medium storing computer instructions. When executed by a processor, these computer instructions implement the planning method for monitoring and maintaining power transmission lines using unmanned aerial vehicles (UAVs), including: Use grid maps to create digital scene models that describe the environment in which power transmission lines are located; Modeling of UAVs used for monitoring and maintaining power transmission lines is conducted from four aspects: function, operating status, energy consumption, and battery. A planner is constructed using a language model. The planner is controlled by prompt words to generate conditional probabilities for controlling the drone to move to a specified power transmission line area in the scene model based on the current and previous drone status, scene model, the actions of each drone in the previous drone group, and the experience of the filtered experience pool. The planner optimizes the local inspection route of the drone inspection in each power transmission line area with the goal of minimizing the inspection flight cost. By using a reinforcement learning-based planning filter to filter out unreasonable task actions and inspection paths generated by planner illusions, unsafe or unrealistic behaviors are avoided. The filtering experience generated by the planning filter is provided to the language model, allowing the language model to refer to past filtering experience to continuously improve decision-making.

[0056] Of course, the computer program stored in the computer-readable storage medium provided in the embodiments of the present invention is not limited to the method operation described above, but can also execute related operations in the planning method for monitoring and maintaining power transmission lines by unmanned aerial vehicles provided in any embodiment of the present invention.

[0057] In the embodiments provided by this invention, it should be understood that the disclosed structures and methods can be implemented in other ways. For example, the structural embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, structures, or units, and may be electrical, mechanical, or other forms.

[0058] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0059] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0060] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A planning method for monitoring and maintaining power transmission lines using unmanned aerial vehicles (UAVs), characterized in that, include: Use grid maps to create digital scene models that describe the environment in which power transmission lines are located; Modeling of UAVs used for monitoring and maintaining power transmission lines is conducted from four aspects: function, operating status, energy consumption, and battery. A planner is constructed using a language model. The planner is controlled by prompt words to generate conditional probabilities for controlling the drone to move to a specified power transmission line area in the scene model based on the current and previous drone status, scene model, the actions of each drone in the previous drone group, and the experience of the filtered experience pool. The planner optimizes the local inspection route of the drone inspection in each power transmission line area with the goal of minimizing the inspection flight cost. By using a reinforcement learning-based planning filter to filter out unreasonable task actions and inspection paths generated by planner illusions, unsafe or unrealistic behaviors are avoided. The filtering experience content generated by the planning filter is provided to the language model, allowing the language model to refer to past filtering experience to continuously improve decision-making. Reinforcement learning-based planning filters include critic networks and actor networks; During training, the critic network and actor network are trained according to drone planning. Drone planning is formulated as a constrained Markov decision process, defined as follows: (S, A, r, P, γ, Ω), where S represents the state space, including: scene model, drone state, battery level, drone currently assigned to the power line area, and cumulative inspection distance; A represents the action space, including commands such as assigning the drone to the power line area, returning to the drone station, switching to charging and idle states; r is the reward function based on state and action; P corresponds to the state transition probability; γ is the discount factor; and Ω is a set of constraints that the drone must satisfy. Constraints are the operational requirements that force the drone to perform, and a cost function is constructed for each constraint. The planning filter achieves reward maximization and constraint compliance through the following Lagrange dual optimization: ； in, The decision made is represented by λ≥0 as the Lagrange multiplier. This is a safety threshold; Actor Network Generated Motion Critics network assesses expected return and related constraint costs The policy gradient update is as follows: ； Where θ is the strategy parameter; The actions formed by the planning filter interact with the scene model and the drone model to generate a new state space, and (S,A,r,P,γ,Ω) are stored in the filtering experience pool.

2. The planning method for monitoring and maintaining power transmission lines using unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The scene model's grid map contains several grids, each configured with the following attributes: altitude, no-fly zone, entity attribute, and entity state attribute. Environmental information is represented through the grid and its attributes. Specifically, the altitude attribute describes the elevation of the corresponding environmental entity, and the grid's location and altitude attribute fully describe the structure of the entity. The no-fly zone attribute describes whether the corresponding environmental area is a no-fly zone. The entity attribute describes the type of the corresponding environmental entity. The entity state attribute describes the state of the corresponding entity.

3. The planning method for monitoring and maintaining power transmission lines using unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, This paper models a drone used for power transmission line monitoring and maintenance from four aspects: function, operational status, energy consumption, and battery. The drone model's functional description attributes describe the power transmission line monitoring and maintenance functions supported by the drone. The drone's operational status attributes describe its location and whether it is charging, working, or idle. Energy consumption is modeled for hovering and propulsion states. The battery model includes battery discharge power and SOC constraints. Based on energy conservation and energy conversion efficiency, [further details are needed for accurate translation]. x=1,2, where, For battery energy conversion efficiency, This refers to the battery discharge power. For non-driving power consumption, when x=1, For the power of the drone in hovering state, when x=2, Power for drone propulsion.

4. The planning method for monitoring and maintaining power transmission lines using unmanned aerial vehicles (UAVs) according to claim 3, characterized in that, Power during hovering: ; in, This represents hovering drag power, caused by the frictional drag between the blades and the air during rotor rotation, and is related to the blade drag coefficient. air density Number of leaves N, leaf chord length Rotation speed cubic and blade radius The fourth power is directly proportional; This represents the hovering induced power, which is the energy required to generate a downward reaction force by accelerating the air to counteract gravity. mg is the weight of the drone. This is the induced power correction factor.

5. The planning method for monitoring and maintaining power transmission lines using unmanned aerial vehicles (UAVs) according to claim 3, characterized in that, Power during propulsion is: ； in, To improve the type resistance power correction factor compared to the hovering state, To improve the induced power correction factor compared to the hovering state, For hovering induced speed; The drag power during level flight is the power required to overcome the aerodynamic drag of the UAV during level flight, and is related to the drag coefficient. air density Number of leaves N, leaf chord length blade radius It is proportional to the cube of the velocity.

6. The planning method for monitoring and maintaining power transmission lines using unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The prompt text includes: The meaning description of the scene model and the current scene model; you are required to understand the information of the scene in which the transmission line is located based on the provided scene model and the meaning description of the scene model; Drone functions, drone status, and understanding the drone's situation based on drone functions and drone status; Filter experience pool; Responsible for drone planning in a scenario, the task is to determine the drone's mission actions based on experience from the filtered experience pool, the current scenario model, the drone's position and status; and to ensure that the selected mission actions comply with the no-fly zone requirements and drone energy constraints in the scenario model. Based on the experience of the filtering experience pool, the current situation of the power transmission line area, and the status of the drone, the inspection route is planned to minimize the inspection flight cost; ensure that the planned route complies with the no-fly requirements and drone energy constraints in the scenario model; If the task actions and paths may violate any constraints, return an empty decision instead of an unsafe one; and interact with the planning filter to filter the task actions and paths.

7. The planning method for monitoring and maintaining power transmission lines using unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, During training, if the cost of the action selected by the actor network exceeds a safety threshold, a safe alternative action will replace the original action.

8. The planning method for monitoring and maintaining power transmission lines using unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, During the filtering process, the actions and paths generated by the planner are directly evaluated by a trained commentator network, which assesses the associated constraint costs. The action and path generated by the planner are compared with a safety threshold to achieve filtering.

9. The planning method for monitoring and maintaining power transmission lines using unmanned aerial vehicles (UAVs) according to claim 8, characterized in that, The (S,A,r,P,γ,Ω) corresponding to the action paths that are filtered out are stored in the filtering experience pool, and the planner makes decisions and improvements based on the experience in the filtering experience pool.

Images