A method and system for processing route planning data for power line inspection

Abstract

The application discloses a kind of for power line inspection route planning data processing method and system, comprising: obtaining line distribution and topographic feature data;Power network area is gridded and determines the priority of inspection, generates regional priority sequence;Decompose overall inspection task, generate multiple task subsets and assign one for each unmanned aerial vehicle, each task subset includes starting position and target line section;Based on topographic feature generates initial route path;Comprehensive analysis all initial paths, identify potential airspace intersection and conflict prediction, by height stratification adjustment to relevant path, generate conflict-free route path set;Calculate the expected flight time and energy consumption of each unmanned aerial vehicle, according to the load balancing principle, reassign boundary tasks, finally generate balanced multi-unmanned aerial vehicle cooperative route scheme.The application can generate safe, efficient multi-machine cooperative inspection route automatically through systematic task decomposition, conflict prediction and load balancing optimization.

Description

Technical Field

[0001] This invention relates to the field of power line inspection technology, and in particular to a method and system for processing route planning data for power line inspection. Background Technology

[0002] In modern power system operation and maintenance, the stable operation of ultra-high voltage transmission lines is crucial. With the continuous expansion of the power grid and the increasing complexity of the geographical environment, traditional manual inspection methods face severe challenges in terms of efficiency, safety, and coverage. Therefore, utilizing drones for automated inspection has become an inevitable trend in the industry. Drones can easily traverse complex terrains such as mountains and deserts, quickly acquiring information about line conditions, greatly improving the breadth and speed of inspections. However, when we shift our focus from the operation of a single drone to large-scale, collaborative swarm operations, a deeper bottleneck emerges: how to efficiently and safely direct multiple drones to work collaboratively within the vast power grid corridor, avoiding mutual interference and ensuring that inspection tasks are neither missed nor duplicated.

[0003] Currently, the industry's common solution is to pre-set a fixed flight path for each drone. This method is manageable with a small number of drones and simple routes, but its drawbacks become glaringly apparent when faced with complex networks spanning hundreds or thousands of kilometers, dotted with numerous poles and crossing points. The biggest problem is that these isolated route plans lack a system-wide collaborative perspective. They cannot effectively anticipate and avoid airspace conflicts between multiple drones. At complex route nodes, path intersections are highly likely to occur, posing a collision risk and forcing pilots to intervene manually in emergencies, which reduces efficiency and increases safety pressure. Furthermore, due to the lack of unified planning and allocation of the overall workload, it is easy for some drones to be overworked while others have no work to do, resulting in uneven workloads. Some route segments are repeatedly inspected, while other sections that urgently need inspection may be overlooked, directly leading to resource waste and the creation of blind spots in inspections. Summary of the Invention

[0004] Therefore, the technical problem to be solved by the present invention is to overcome the problems of airspace conflict and uneven task load caused by isolated route planning and lack of coordination in the existing technology of multi-UAV power line inspection. The invention provides a method and system for processing route planning data for power line inspection, which can generate safe and efficient multi-UAV collaborative route schemes through systematic task decomposition, conflict prediction and load balancing optimization.

[0005] To address the aforementioned technical problems, this invention provides a method for processing route planning data for power line inspection, comprising the following steps:

[0006] Acquire power line data, including: line distribution data and terrain feature data;

[0007] Based on line distribution data, the entire power network area is divided into multiple grids, the inspection priority of each grid is determined, and a regional priority sequence is generated.

[0008] Based on the regional priority sequence, the overall patrol mission data is processed and decomposed to generate multiple mission subset data, and each UAV is assigned a mission subset data, where each mission subset data corresponds to a starting position data and a target route segment data.

[0009] For each UAV's mission subset data, terrain feature data is processed to generate initial flight path data from its starting position to the target route segment;

[0010] Analyze and process the initial flight path data of all UAVs to identify potential airspace intersections in the path data;

[0011] Conflict prediction analysis is performed on the identified potential airspace intersection data, flight path data is adjusted, and the flight path data of relevant UAVs is adjusted by height hierarchy to generate a set of conflict-free flight path data.

[0012] Based on the set of conflict-free flight path data, the estimated flight time and energy consumption data of each UAV are calculated. According to the load balancing principle, the boundary task data between the task subset data are redistributed to generate balanced multi-UAV collaborative flight path scheme data.

[0013] In one embodiment of the present invention, the entire power grid area is divided into multiple grids based on line distribution data, specifically including:

[0014] Identify all pole nodes in the line distribution data and connect adjacent pole nodes to form continuous line segments;

[0015] Centered on each tower node, and combined with the preset effective line-of-sight distance for drone inspection, an initial circular area of ​​influence is determined;

[0016] The overlapping circular influence areas are merged to form one or more continuous region clusters. For each region cluster, its bounding rectangle is calculated, and a regular rectangular mesh is generated using the geometric center of the bounding rectangle as the base point.

[0017] By spatially associating the rectangular grid with the original line segments and filtering out blank grids that do not contain any line segments, multiple grids covering the entire power network area are finally obtained.

[0018] In one embodiment of the present invention, determining the inspection priority of each grid and generating a region priority sequence specifically includes:

[0019] Multiple independent priority levels are defined for each grid. The priority levels include at least the first level: line criticality level, which characterizes the importance of the lines in the grid in the power grid structure; the second level: environmental risk level, which characterizes the complexity and severity of the geographical environment in which the grid is located; and the third level: operational risk level, which characterizes the stability of the lines in the grid in the near future.

[0020] Establish a hierarchical screening mechanism to sort multiple priority levels from high to low according to decision importance, forming a decision hierarchy;

[0021] Among all the grids, the grids are first filtered and grouped according to the highest priority level, and the grids that meet the highest priority level are divided into the first priority group;

[0022] Within the first priority group, the ranking is further refined according to the second highest priority level; at the same time, in the remaining grids that were not assigned to the first priority group, the second priority group is also selected according to the second highest priority level.

[0023] The selection and sorting process proceeds step by step down the decision-making hierarchy until all grids are included in a sequence with a clear order, thereby generating a regional priority sequence.

[0024] In one embodiment of the present invention, based on a regional priority sequence, the overall inspection task data is processed and decomposed to generate multiple task subset data, specifically including:

[0025] Initialize a pool of line segments to be assigned, which contains all line segments that need to be inspected, and sort the line segments in the pool according to the area priority sequence.

[0026] Calculate the expected load range for each task subset based on the total number of available drones and the total task load of the pool of line segments to be allocated;

[0027] Create an empty temporary task set, and add line segments to the temporary task set sequentially, starting from the top of the pool of line segments to be assigned, i.e. the line segment with the highest priority.

[0028] The current total load of the temporary task set is accumulated in real time, and the addition of line segments is stopped when the current total load enters the expected load range;

[0029] Define and output the temporary task set containing several line segments as a formal task subset data;

[0030] Remove all assigned line segments from the pool of line segments to be assigned, and assign task subsets to the remaining line segments until the pool of line segments to be assigned is empty or all drones have been assigned task subsets, generating multiple task subset data.

[0031] In one embodiment of the present invention, spatial clustering analysis is performed on all line segments included in the temporary task set to identify a main line segment cluster that forms a close cluster in space, as well as one or more isolated line segments that are clearly separated from it in space.

[0032] The identified isolated line segments are removed from the temporary task set and returned to the line segment pool to be assigned, for use in the subsequent task subset construction.

[0033] The temporary task set, which has been stripped and optimized and contains only the dominant line segment cluster, is defined and output as a spatially contiguous subset of task data.

[0034] In one embodiment of the present invention, for each UAV's mission subset data, initial flight path data from its starting position to the target route segment is generated based on terrain feature data processing, specifically including:

[0035] A three-dimensional route planning space is constructed, which integrates the starting position, target route segment and terrain feature data. The terrain feature data is converted into terrain elevation information and obstacle three-dimensional information in space.

[0036] In the three-dimensional route planning space, multiple alternative initial path connections are generated between the starting position and the target route segment. For each alternative initial path connection, a safety assessment is performed. The assessment includes calculating the minimum spatial distance between the initial path and the terrain elevation information and obstacle three-dimensional information, and marking all path segments that do not meet the safety distance margin.

[0037] For the marked unsafe path segments, local path reconstruction is performed in the three-dimensional route planning space. By introducing one or more auxiliary path points, the path is rerouted to ensure that the entire path meets the global safe distance requirements and generate multiple safe alternative paths.

[0038] An economic evaluation is conducted on all safe alternative routes, taking into account the total length and undulation of the routes. The route with the best evaluation results is selected as the final initial flight path data.

[0039] In one embodiment of the present invention, conflict prediction analysis is performed on the identified potential airspace intersection data, specifically including:

[0040] For each UAV, a flight status simulation model is established. Based on its initial flight path data and preset cruise speed, the model simulates the expected time stamp sequence of each waypoint on its arrival path.

[0041] For each identified potential airspace intersection, extract all UAVs that may pass through the potential airspace intersection and their corresponding flight state simulation models;

[0042] Based on the expected timestamp sequence, calculate the absolute value of the time difference between the expected arrival times of any two drones at the intersection point;

[0043] The absolute value of the time difference is compared with a dynamic safety time window threshold. When the absolute value of the time difference is less than the dynamic safety time window threshold, it is determined that there is a risk of conflict at the intersection, and the drones involved, the location of the intersection, and the expected conflict time are marked as a conflict event.

[0044] In one embodiment of the present invention, the process of determining the safety time window threshold includes:

[0045] A basic safety time window is set, which is calculated based on the physical space safety scale and cruising speed of the UAV in the intersection airspace.

[0046] Obtain the external environmental assessment parameters required for conflict prediction. The parameters should include at least the current or forecast wind speed data for the cross-point airspace and the satellite positioning signal quality index.

[0047] Based on wind speed data and signal quality index, positive compensation is applied to the basic safety time window; the higher the wind speed or the worse the signal quality, the greater the compensation, resulting in a more lenient safety time window threshold.

[0048] The time window value obtained after compensation and adjustment is determined as the dynamic safe time window threshold for the current conflict prediction analysis.

[0049] In one embodiment of the present invention, according to the load balancing principle, the boundary task data between task subset data is redistributed, specifically including:

[0050] After generating a set of conflict-free flight path data, the estimated total flight time for each task subset is calculated, and the task subset with the longest estimated time is identified as the high-load subset, and the task subset with the shortest estimated time is identified as the low-load subset.

[0051] Locate the spatial boundary between the high-load subset and the adjacent task subset, and identify the boundary line segments located in this boundary region that can be separated without disrupting the spatial continuity of their own task subset, and define them as candidate adjustment task blocks.

[0052] From the candidate adjustment task blocks, select the task block whose estimated flight time is closest to half the difference between the high load and low load subsets as the proposed adjustment task block;

[0053] The task blocks to be adjusted are removed from the high-load subset and merged into the low-load subset, while the route path data and estimated total flight time of these two subsets are updated.

[0054] Repeat the above steps to iterate the load balancing process until the maximum difference in the estimated total flight time between all task subsets is lower than the preset balancing threshold, and finally generate a balanced task load distribution.

[0055] To address the aforementioned technical problems, this invention also provides a system for processing route planning data for power line inspection, used to implement the above method, comprising:

[0056] The data acquisition module is used to acquire power line data, including: line distribution data and terrain feature data;

[0057] The area division and priority sorting module is used to divide the entire power network area into multiple grids based on the line distribution data, and determine the inspection priority of each grid to generate an area priority sequence.

[0058] The task decomposition and allocation module is used to process and decompose the overall patrol task data based on the regional priority sequence, generate multiple task subset data, and allocate one of the task subset data to each UAV, wherein each task subset data corresponds to a starting position data and a target route segment data.

[0059] The initial flight path planning module is used to process the terrain feature data based on the mission subset data of each UAV and generate its initial flight path data from the starting position to the target route segment.

[0060] The airspace conflict identification module is used to analyze and process the initial flight path data of all UAVs and identify potential airspace intersection points in the path data.

[0061] The conflict resolution and flight path adjustment module is used to perform conflict prediction analysis on the identified potential airspace intersection data and to perform flight path data adjustment. It performs height-level adjustment on the flight path data of relevant UAVs to generate a set of conflict-free flight path data.

[0062] The load balancing and scheme generation module is used to calculate the estimated flight time and energy consumption data of each UAV based on the conflict-free flight path set data, and to redistribute the boundary task data between the task subset data according to the load balancing principle, so as to finally generate balanced multi-UAV collaborative flight path scheme data.

[0063] The technical solution of the present invention has the following advantages compared with the prior art:

[0064] This invention provides a method for processing flight path planning data for power line inspection. Through proactive and systematic intelligent planning, it significantly improves the safety and reliability of multi-UAV collaborative operations, effectively preventing mid-air collisions. Through scientific task decomposition and load balancing optimization, it significantly improves the working efficiency and resource utilization of the entire inspection fleet, avoiding task overlap or omissions and ensuring the uniformity and integrity of inspection quality. This method provides solid technical support for building an intelligent and automated power grid operation and maintenance system, making rapid, accurate, and routine inspections of large-scale transmission lines possible. Attached Figure Description

[0065] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein:

[0066] Figure 1 This is a flowchart of the steps of the method for processing route planning data for power line inspection according to the present invention;

[0067] Figure 2 This is a flowchart of the steps for dividing multiple grids according to the present invention;

[0068] Figure 3 This is a flowchart of the steps for generating a region priority sequence according to the present invention;

[0069] Figure 4 This is a flowchart of the steps for generating multiple task subsets according to the present invention;

[0070] Figure 5 This is a flowchart of the steps for generating the initial flight path according to the present invention;

[0071] Figure 6 This is a flowchart of the conflict prediction analysis steps of the present invention;

[0072] Figure 7 This is a flowchart of the steps for reallocating task subset data according to the present invention;

[0073] Figure 8 This is a structural framework diagram of the system for processing route planning data for power line inspection according to the present invention. Detailed Implementation

[0074] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0075] Reference Figure 1As shown, the present invention provides a method for processing route planning data for power line inspection. Through a systematic data processing flow, the original power line data is transformed into an executable multi-UAV collaborative route scheme, including the following steps:

[0076] First, power line data for the area to be inspected is obtained from the power asset management system. This data includes line distribution data (such as tower coordinates and line direction) and terrain feature data (such as elevation information and obstacle distribution). Based on the line distribution data, a preset grid partitioning algorithm is used to divide the entire power network area into multiple equal-sized grid cells. Subsequently, an inspection priority score is calculated and assigned to each grid cell according to preset rules, ultimately generating a priority sequence of areas ordered by priority. This step enables scientific quantification and focused attention on the vast inspection area.

[0077] Upon receiving the overall inspection mission command, the task decomposition process is initiated based on the generated area priority sequence. Specifically, according to the number of available UAVs, a task allocation algorithm is used to decompose the overall mission into multiple spatially contiguous task subsets with similar workloads. Each task subset contains a defined starting position and one or more target route segments. A unique task subset is assigned to each UAV, and the task subset data is sent to the corresponding UAV flight control system via a data interface.

[0078] For each UAV, based on the acquired terrain feature data, a path planning algorithm is run to calculate an optimal or suboptimal initial flight path from the starting position to the target route segment. This path must meet the UAV's flight performance constraints and avoid known fixed obstacles, generating a set of initial flight path data containing a series of path point coordinates.

[0079] The initial flight path data of all UAVs are analyzed within the same airspace model. Through geometric calculations and spatial indexing techniques, potential airspace intersections between all paths are identified. The built-in collision prediction model analyzes these intersections. If the predicted spatial distance between two or more UAVs at a future time is less than a preset safety threshold, a conflict risk is identified. For conflicting flight paths, the flight path data is automatically adjusted, prioritizing an altitude-layered strategy. This assigns different flight altitude layers to the conflicting UAVs, thereby generating a set of conflict-free flight path data.

[0080] Based on a set of conflict-free flight paths, the estimated flight time and energy consumption data required for each UAV to perform its mission are simulated and calculated. Analyzing this data, if the difference in mission load between different UAVs exceeds a preset balancing threshold, a load balancing process is initiated. This process redistributes tasks by fine-tuning the ownership of boundary tasks (i.e., poles or line segments at the edge of the mission area that can be reassigned) between adjacent task subsets until the estimated load of all UAVs tends to be balanced. Finally, all optimization results are integrated to generate a final collaborative flight path scheme data containing the precise flight paths, altitude layers, and time nodes of each UAV, which can be output to the dispatch center or each UAV execution terminal.

[0081] In summary, the flight path planning data processing method in this embodiment no longer treats the flight path of each UAV as an independent entity, but rather considers it holistically within a unified and intelligent data processing flow. Through a series of interconnected data transformation and optimization steps, a large and complex overall inspection task is transformed into a clear, safe, and efficient collaborative operation plan. The entire process begins with in-depth mining of the power line data itself. Through gridding and priority sorting, it can intelligently identify which sections are of paramount importance for inspection, thus laying a scientific foundation for subsequent task allocation. Next, using task decomposition and allocation logic, the macro-task is broken down into task packages matching the number of UAVs, ensuring that each UAV has a patrol area of ​​comparable magnitude and clearly defined responsibility before the task begins, thus avoiding uneven workloads from the outset. Subsequently, after planning the initial path for each task package, airspace conflict analysis and prediction are introduced. Through algorithms, the flight paths of all UAVs are simulated in advance, accurately locating potential intersection hazards and proactively adopting highly layered strategies for avoidance. Finally, the initially formed conflict-free routes will be optimized, and the mission boundaries will be fine-tuned based on flight time and energy consumption data to achieve a high degree of balance in the mission load of the entire fleet. What is generated is not just a set of route coordinates, but an optimal collaborative solution that takes into account mission priority, flight safety, and resource efficiency.

[0082] Reference Figure 2As shown, this embodiment further discloses the steps of dividing the entire power network area into multiple grids based on line distribution data. The specific implementation is as follows: First, the acquired line distribution data is parsed to identify the geographic coordinates of all pole nodes; then, according to the line topology, adjacent pole nodes are connected in pairs to reconstruct continuous line segments in digital space; after that, the core process of grid division is initiated: taking each identified pole node as the center, and based on the effective line-of-sight distance of the selected UAV inspection platform (this parameter can be preset and adjusted, and different UAV inspection platforms have different effective line-of-sight distances), an initial circular influence area is determined in the geographic information system. This area represents a basic inspection unit with this pole as the core. Since the distance between adjacent towers is usually less than the effective line-of-sight range of drones, these circular areas often overlap. Through spatial clustering algorithms, these overlapping circular influence areas are automatically merged to form one or more spatially continuous area clusters, each corresponding to a line or a dense line segment. The minimum bounding rectangle of each merged area cluster is calculated, and the geometric center of this rectangle is used as the reference point to divide it into regular, uniformly sized rectangular grids. Finally, the divided rectangular grids are accurately spatially associated and matched with the initially reconstructed line segments, and blank grids that do not contain any line segments are automatically filtered out, resulting in a set of multiple grids that completely cover the entire effective area of ​​the power network and closely match the line direction.

[0083] The grid division method adopted in this embodiment can fundamentally avoid the problems that may be caused by mechanical grid division, such as the route being cut off or the grid being disconnected from the inspection target. By "starting from the route node, merging the affected area, and then regularizing the division", it is ensured that each final generated grid contains the actual inspection target and its size matches the actual inspection capability of the UAV. This not only lays a reasonable spatial foundation for the accurate evaluation of subsequent priorities, but also makes grid-based task allocation more operable and efficient, thereby improving the rationality and practicality of the route planning data as a whole.

[0084] Reference Figure 3As shown, this embodiment further discloses the process of determining the inspection priority of each grid to generate a regional priority sequence. The specific implementation method is as follows: Define and calculate multiple independent priority level evaluation indicators for each grid; these indicators include at least: the first level, namely the line criticality level, which is determined by analyzing the voltage level of the line within the grid, whether it belongs to the backbone network or the power supply line for important loads, and is used to quantitatively characterize the importance of this section of the line in the entire power grid topology; the second level, namely the environmental risk level, which is assessed by integrating geographic information system data and comprehensively considering factors such as the terrain undulation, vegetation density, presence of easily floating debris sources, and historical meteorological disaster records within the grid, and is used to characterize the potential threat posed by the geographical environment to line safety and the difficulty of inspection operations; the third level, namely the operational risk level, which is analyzed by accessing real-time or historical operational data from the power grid dispatching system, analyzing recent load fluctuations, fault alarm records, and insulator deterioration trends of the lines within the grid, and is used to characterize the stability of the line's own operation and the probability of potential faults.

[0085] To integrate these three dimensions of evaluation into an operational and unified decision-making sequence, a hierarchical screening mechanism was established. This mechanism first, based on the core principle of "safety first, prevention foremost" in power grid inspection, pre-orders the aforementioned priority levels according to their decision-making importance from highest to lowest, for example, forming a decision hierarchy of line criticality level > environmental risk level > operational risk level. In the actual sorting process, all grids to be inspected are first screened and grouped according to the highest-ranked line criticality level. All grids belonging to the UHV main network or involving the power supply of important cities, i.e., the grids with the highest scores at that level, are uniformly classified into the first priority group. Subsequently, the sorting logic moves to the next level: within the first priority group, all grids within the group are further sorted according to the second-highest environmental risk level, thus determining the order within the group. Simultaneously, other grids are not ignored; from the remaining grids not classified into the first priority group, the grids with the highest environmental risk level scores are also selected to form the second priority group. This process continues in the same manner, filtering, grouping, and sorting within each group according to the preset decision-making hierarchy, ensuring that each grid is assigned a clear and unique sequential position in the entire inspection sequence based on its comprehensive performance at each level.

[0086] This prioritization method, through a multi-dimensional and structured evaluation system, overcomes the limitations of decision-making based on a single indicator (such as voltage level alone), making the priority assessment results more scientific and comprehensive, and truly reflecting the urgency of the grid's comprehensive inspections. More importantly, its hierarchical screening mechanism simulates the thinking of human experts in complex decision-making, prioritizing the resolution of "critical" issues before addressing "relatively important" ones. This ensures that limited inspection resources are always prioritized for the lines with the highest risk and greatest impact, greatly enhancing the strategic nature of inspection work and overall security capabilities. The resulting regional priority sequence provides a clear and reliable data-driven basis for subsequent task decomposition and resource allocation, optimizing the efficiency of the entire collaborative inspection process from the decision-making source.

[0087] Reference Figure 4 As shown, this embodiment further discloses the specific implementation process of processing and decomposing the overall inspection task data based on the regional priority sequence to generate multiple task subsets. First, an allocation task pool containing all the line segments that need to be inspected is initialized, and the task pool is strictly sorted according to the regional priority sequence to ensure that the line segment with the highest priority is always at the top of the queue. Then, based on the total number of currently available UAVs and the estimated total task load of the allocation line segment pool (which can be estimated based on the total length of the line segment and the average inspection time), the ideal expected load range of each task subset is calculated, aiming to achieve a rough balance of workload for each UAV.

[0088] Specifically, the core decomposition process is as follows: Create a temporary empty task set, and starting from the top of the waiting pool, add the top-ranked line segments to this set in sequence; accumulate the current total load of this temporary set in real time, and stop adding once the accumulated load enters the preset expected range, and define and output this temporary set as a formal task subset data, which is then assigned to a standby drone; subsequently, remove these assigned line segments from the waiting pool, and repeat the above process to build a new task subset for the next drone, until the task pool is cleared or all drones have been assigned tasks.

[0089] To further improve the operational efficiency of each task subset, after the aforementioned temporary task set is formed, spatial clustering analysis is performed on all the line segments it contains. This identifies the main line segment clusters that are closely clustered in spatial geographic location, as well as isolated line segments that are far from the main clusters and clearly separated in space. The existence of these isolated line segments may cause UAVs to need to travel long distances ineffectively between different areas when performing tasks, seriously wasting operational time and energy. Therefore, these identified isolated line segments will be separated from the current temporary set and re-assigned to the line segment pool to be integrated into other more suitable task subsets later. After this separation and optimization, the final output task subset data will only contain spatially continuous and clustered main line segment clusters.

[0090] The task decomposition method in this embodiment, through priority ranking and load control, macroscopically ensures that inspection resources are concentrated on critical paths and achieves task load balancing among multiple UAVs. Through spatial clustering and isolated segment stripping optimization, it microscopically ensures that the tasks assigned to each UAV are geographically compact and continuous. This significantly reduces the ineffective flight distance and time of UAVs during task execution, thereby significantly improving single-aircraft operational efficiency and overall collaborative effectiveness, and effectively reducing the overall energy consumption of UAVs.

[0091] Reference Figure 5 As shown, this embodiment further discloses the specific implementation process of generating initial flight paths for each UAV's mission subset data. First, a refined three-dimensional flight path planning space is constructed. This space is a digital three-dimensional environment that integrates the UAV's starting position coordinates, the spatial orientation of the target route segments included in the mission subset, and high-precision terrain feature data. The terrain feature data is converted into a terrain elevation grid in this space, while fixed obstacles such as buildings and vegetation are represented as no-fly zones with three-dimensional outlines.

[0092] Within this three-dimensional space, instead of directly planning a single route, multiple alternative initial path lines are first generated from the starting position to the target route segment. These lines form the baseline for path search. Subsequently, a rigorous safety assessment is performed on each alternative path, the core of which is calculating the minimum spatial distance between each point on the path and the surrounding terrain elevation and obstacle 3D information. Based on safety requirements, a preset safety distance margin threshold is established; for example, the safety distance margin threshold can be set to 5-10 meters. Any path segment whose minimum spatial distance is lower than this threshold is marked as an unsafe segment by the system. For paths with unsafe segments, a local path reconstruction algorithm is initiated in three-dimensional space. This algorithm introduces one or more new auxiliary path points near the unsafe segments of the original path, guiding the path to perform detours, thereby effectively avoiding terrain and obstacles and ensuring that the entire path meets the global safety distance requirements. After this round of correction, several alternative paths that all meet the basic safety requirements are obtained.

[0093] Next, the decision-making phase begins, where these safe alternative routes undergo an economic evaluation. This evaluation is a multi-objective trade-off process, which not only calculates the total flight length of the route but also analyzes the undulation of the route (which directly affects flight energy consumption and stability) and may take into account factors such as turning angles. The evaluation is then quantitatively compared using a weighted scoring model. Finally, the route with the best economic evaluation result is selected from all alternative routes, and its path point coordinate sequence is output as the final initial flight path data.

[0094] The initial flight path generation method in this embodiment significantly improves the basic quality and reliability of the flight path. By constructing a three-dimensional planning space and performing rigorous safety distance assessment and local reconstruction, it fundamentally eliminates the risk of collisions between the flight path and terrain or obstacles, providing primary protection for UAV flight safety. Furthermore, by introducing a multi-path alternative and economic evaluation mechanism, it overcomes the efficiency shortcomings that may exist in single-path planning. Under the premise of ensuring safety, it can proactively find an optimized path with a shorter total flight distance and more stable and economical flight. This not only directly reduces the flight time and energy consumption of a single mission, but also provides a better initial solution with a lower probability of conflict for subsequent swarm collaborative conflict resolution, thus laying a solid foundation for the entire collaborative inspection mission in terms of both safety and efficiency.

[0095] Reference Figure 6As shown, this embodiment further discloses the specific implementation process of conflict prediction analysis. First, a flight state simulation model is established for each UAV participating in the collaborative inspection. This model is based on the initial flight path data of the UAV and combined with its preset cruising speed. By calculating and deriving the precise estimated timestamp sequence of each path point on the flight path, the flight trajectory of each UAV in the time dimension is simulated. Subsequently, for each potential airspace intersection identified in the previous steps, all UAVs whose flight paths will pass through this intersection and their corresponding flight state simulation models are retrieved and extracted. Based on the estimated timestamp sequence provided by these models, the absolute value of the time difference between any two UAVs expected to arrive at the intersection is calculated. This time difference is the core basis for conflict judgment. This time difference is compared with a preset safe time window threshold. When the absolute value of the time difference is less than the threshold, it is determined that there is a risk of conflict between the two UAVs in the airspace at the intersection. The UAVs involved, the spatial location of the intersection, and the estimated conflict time are collectively marked as a specific conflict event, providing clear input for subsequent conflict resolution.

[0096] Specifically, to improve the accuracy and environmental adaptability of conflict prediction, the safety time window threshold is not a fixed value but a dynamically adjusted parameter. Its determination begins with a basic safety time window, calculated by dividing the minimum physical safety interval that the UAV needs to maintain in the intersection airspace by its cruising speed. Based on this, external environmental assessment parameters required for conflict prediction are obtained. These parameters include at least the current or predicted wind speed data in the intersection airspace and the satellite positioning signal quality index of the area. The basic safety time window is positively compensated and adjusted based on these real-time environmental parameters. The logic is as follows: higher wind speeds increase the difficulty for the UAV to maintain its predetermined trajectory and may increase positioning errors; poorer satellite signal quality reduces the UAV's positioning accuracy and control system stability. Both of these situations require a larger safety buffer. Therefore, the compensation rule is that the higher the wind speed or the worse the signal quality, the greater the compensation amount, thus calculating a more lenient and safer dynamic safety time window threshold, which is then used for the current round of conflict prediction analysis.

[0097] The conflict prediction and analysis method in this embodiment improves the system's accuracy and robustness in perceiving potential risks. By establishing a time-based model for each UAV, it transforms the spatial path conflict problem into a more fundamental temporal conflict problem, making the prediction results more accurate and intuitive. More importantly, by introducing a dynamic safety time window based on wind speed and positioning signal quality, this method transforms conflict prediction from a static judgment based on an ideal laboratory environment into a dynamic risk perception that fully considers the complex and ever-changing environmental factors in actual field operations. This significantly reduces the risk of missed detections caused by flight path deviations or positioning signal drift due to gusts, while also avoiding false alarms caused by overly strict thresholds under ideal conditions. Thus, while ensuring absolute flight safety, it also guarantees the efficiency and smoothness of the overall inspection mission planning.

[0098] Reference Figure 7 As shown, this embodiment further discloses the specific implementation process of redistributing task subset data according to the load balancing principle. First, after generating a set of conflict-free flight paths, the estimated total flight time corresponding to each task subset is calculated based on the precise path, thereby quantifying the task load of each UAV. Then, the task subset with the longest estimated time is identified and marked as the high-load subset, while the task subset with the shortest estimated time is identified and marked as the low-load subset.

[0099] Subsequently, the spatial boundaries between the high-load subset and one or more adjacent task subsets are located. Using spatial analysis algorithms, line segments that can be relatively independently separated from the high-load subset near the boundary without severely disrupting the spatial continuity of the main body of its original subset are identified. These line segments are defined as candidate adjustment task blocks. The estimated flight durations of these candidate task blocks are evaluated, and the task block whose duration is closest to half the current total duration difference between the high-load and low-load subsets is selected as the proposed adjustment task block for this iteration. This selection strategy aims to minimize the load gap with the fewest adjustments.

[0100] After identifying the task block to be adjusted, a task transfer operation is performed: the task block is formally removed from its original high-load subset and spatially merged with its task content into the low-load subset; then, the internal flight paths of the two affected task subsets are automatically replanned, and their respective estimated total flight times are updated. The entire process constitutes a complete load balancing iteration. This iteration is repeated, with each iteration fine-tuning the task block between the two most unbalanced subsets, until the maximum difference in estimated total flight times among all task subsets stabilizes below a preset balance threshold. At this point, the load is considered to have reached a satisfactory balance, and the final balanced task load distribution is output.

[0101] The load balancing method disclosed in this embodiment improves the overall efficiency and resource utilization of multi-UAV collaborative operations. By accurately locating load extremes and intelligently stripping and transferring boundary task blocks, it achieves dynamic redistribution of task load, effectively avoiding situations where individual UAVs become bottlenecks in the entire inspection process due to excessive task load. This method is not a simple egalitarian approach, but rather a fine-tuning process that strives to maintain the spatial continuity of each task subset. This achieves load balancing while minimizing the additional path costs caused by task adjustments. Ultimately, it ensures that the fleet can complete its respective inspection tasks in approximately the same amount of time, shortening the overall task cycle and reducing the idle waiting time of some UAVs and operators, making the organization and execution of the entire inspection task more efficient and economical.

[0102] Reference Figure 8 As shown, in order to implement the above method, this embodiment further discloses a system for processing route planning data for power line inspection, including:

[0103] The data acquisition module is used to acquire power line data, including: line distribution data and terrain feature data;

[0104] The area division and priority sorting module is used to divide the entire power network area into multiple grids based on the line distribution data, and determine the inspection priority of each grid to generate an area priority sequence.

[0105] The task decomposition and allocation module is used to process and decompose the overall patrol task data based on the regional priority sequence, generate multiple task subset data, and allocate one of the task subset data to each UAV, wherein each task subset data corresponds to a starting position data and a target route segment data.

[0106] The initial flight path planning module is used to process the terrain feature data based on the mission subset data of each UAV and generate its initial flight path data from the starting position to the target route segment.

[0107] The airspace conflict identification module is used to analyze and process the initial flight path data of all UAVs and identify potential airspace intersection points in the path data.

[0108] The conflict resolution and flight path adjustment module is used to perform conflict prediction analysis on the identified potential airspace intersection data and to perform flight path data adjustment. It performs height-level adjustment on the flight path data of relevant UAVs to generate a set of conflict-free flight path data.

[0109] The load balancing and scheme generation module is used to calculate the estimated flight time and energy consumption data of each UAV based on the conflict-free flight path set data, and to redistribute the boundary task data between the task subset data according to the load balancing principle, so as to finally generate balanced multi-UAV collaborative flight path scheme data.

[0110] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for processing route planning data for power line inspection, characterized by, The method comprises the following steps: Obtaining power line data, including: line distribution data and terrain feature data; Dividing the entire power network area into multiple grids based on the line distribution data, specifically including: identifying all tower nodes in the line distribution data, and connecting adjacent tower nodes to form continuous line sections; taking each tower node as the center, combining the preset effective inspection distance of the unmanned aerial vehicle, and determining an initial circular influence area; merging the mutually overlapping circular influence areas to form one or more continuous area clusters, for each area cluster, calculating its circumscribed rectangle, and taking the geometric center of the circumscribed rectangle as the base point to perform regular rectangular grid division; spatially associating the rectangular grid with the original line section, and screening out blank grids that do not contain any line section, to finally obtain multiple grids covering the entire power network area; determining the inspection priority of each grid to generate a regional priority sequence; Based on the regional priority sequence, processing and decomposing the overall inspection task data to generate multiple task subset data, and assigning each unmanned aerial vehicle a task subset data, wherein each task subset data corresponds to a starting position data and a target line section data; For the task subset data of each unmanned aerial vehicle, processing and generating the initial flight path data from the starting position to the target line section based on the terrain feature data; Analyzing and processing the initial flight path data of all unmanned aerial vehicles to identify potential airspace intersection point data in the path data; Performing conflict prediction analysis on the identified potential airspace intersection point data, adjusting the flight path data, and performing high-level hierarchical adjustment on the flight path data of related unmanned aerial vehicles to generate a conflict-free flight path set data; Based on the conflict-free flight path set data, calculating the expected flight time data and energy consumption data of each unmanned aerial vehicle, and reassigning the boundary task data between the task subset data according to the load balancing principle to finally generate an balanced multi-unmanned aerial vehicle cooperative flight scheme data.

2. The method of claim 1, wherein: Determining the inspection priority of each grid to generate a regional priority sequence, specifically including: Defining multiple independent priority levels for each grid, including at least a first level: line criticality level, used to represent the importance of the grid in the power grid structure; a second level: environmental risk level, used to represent the complexity and severity of the geographical environment of the grid; a third level: operation risk level, used to represent the stability of the recent operation of the line in the grid; Establishing a hierarchical screening mechanism to sort multiple priority levels from high to low according to decision importance to form a decision hierarchy; In all grids, first, screen and group according to the highest priority level, and divide the grid that meets the highest priority level into the first priority group; In the first priority group, further refine the sorting according to the second highest priority level; at the same time, in the remaining grids that are not included in the first priority group, also screen out the second priority group according to the second highest priority level; Screen and sort step by step according to the decision hierarchy until all grids are included in a sequence with clear order, thereby generating a regional priority sequence.

3. The method of claim 1, wherein: Based on the sequence of regional priority, the overall inspection task data is processed and decomposed to generate multiple task subset data, specifically including: Initialize a pool of to-be-assigned line segments, which contains all line segments that need to be inspected, and sort the line segments in the pool according to the sequence of regional priority; According to the total number of available unmanned aerial vehicles and the total task load of the pool of to-be-assigned line segments, calculate the expected load range of each task subset; Create an empty temporary task set and add line segments from the top of the pool of to-be-assigned line segments, i.e. the highest priority line segment, to the temporary task set one by one; Real-time cumulative current total load of the temporary task set, when the current total load enters the expected load range, stop adding line segments; Define and output the temporary task set containing several line segments at this time as a formal task subset data; Remove all assigned line segments from the pool of to-be-assigned line segments, and assign task subsets to the remaining line segments until the pool of to-be-assigned line segments is empty or all unmanned aerial vehicles have been assigned task subsets, generating multiple task subset data.

4. The processing method for flight path planning data of power line inspection according to claim 3, characterized in that: Perform spatial clustering analysis on all line segments contained in the temporary task set to identify a main line segment cluster that forms a close cluster in space and one or more isolated line segments that are obviously separated in space; Strip the identified isolated line segments from the temporary task set and reclassify them into the pool of to-be-assigned line segments for subsequent task subset construction; Define and output the temporary task set containing only the main line segment cluster after stripping and optimization as a spatially continuous task subset data.

5. The method of claim 1, wherein: For each unmanned aerial vehicle's task subset data, based on the terrain feature data, process and generate its initial flight path data from the starting position to the target line segment, specifically including: Construct a three-dimensional flight path planning space that integrates the starting position, target line segment, and terrain feature data, where the terrain feature data is converted into terrain elevation information and obstacle three-dimensional information in space; In the three-dimensional flight path planning space, generate multiple candidate initial path connections between the starting position and the target line segment, and for each candidate initial path connection, perform safety evaluation, which includes calculating the minimum spatial distance between the initial path and the terrain elevation information and obstacle three-dimensional information, and marking all path sections that do not meet the safety distance margin; For the marked unsafe path sections, perform local path reconstruction in the three-dimensional flight path planning space by introducing one or more auxiliary path points to make the path detour to ensure that the entire path meets the global safety distance requirement, generating multiple safe candidate paths; Perform economic evaluation on all safe candidate paths, considering the total length and undulation of the path, and select the path with the best evaluation result as the final initial flight path data.

6. The method of claim 1, wherein: Perform conflict prediction analysis on the identified potential airspace intersection point data, specifically including: establishing a flight state deduction model for each UAV, the model being based on initial route path data and a preset cruising speed of the UAV, and deducing a sequence of expected time stamps of the UAV reaching each waypoint on the route path; for each identified potential airspace intersection point, extracting all UAVs that are likely to pass through the potential airspace intersection point and their corresponding flight state deduction models; based on the sequence of expected time stamps, calculating an absolute value of a time difference of any two UAVs expected to reach the intersection point; comparing the absolute value of the time difference with a dynamic safety time window threshold value, and when the absolute value of the time difference is less than the dynamic safety time window threshold value, determining that the intersection point has a conflict risk, and marking the involved UAVs, the intersection point location and the expected conflict time as a conflict event.

7. The method of claim 6, wherein: The determination process of the safety time window threshold value includes: setting a basic safety time window, which is calculated based on a physical space safety dimension of the UAV in the intersection point airspace and the cruising speed; obtaining external environment assessment parameters required for conflict prediction, including at least current or forecast wind speed data of the intersection point airspace and satellite positioning signal quality index; based on the wind speed data and the signal quality index, positively compensating the basic safety time window, wherein the greater the wind speed or the worse the signal quality, the greater the compensation amount, and a more relaxed safety time window threshold value is obtained; determining the time window value obtained after compensation and adjustment as the dynamic safety time window threshold value used for current conflict prediction analysis.

8. The method of claim 1, wherein: According to the load balancing principle, the boundary task data between the task subset data is re-distributed, specifically including: after generating the conflict-free route path set data, calculating the estimated total flight time of each task subset, and identifying one task subset with the longest estimated time as a high-load subset and one task subset with the shortest estimated time as a low-load subset; locating the spatial boundary between the high-load subset and the adjacent task subset, and identifying a boundary line segment located in the boundary region and capable of being separated without destroying the spatial continuity of the task subset to which it belongs, and defining it as a candidate adjustment task block; from the candidate adjustment task block, selecting a task block with an estimated flight time closest to half the time difference between the high-load and low-load subsets as a tentative adjustment task block; removing the tentative adjustment task block from the high-load subset and merging it into the low-load subset, and updating the route path data and the estimated total flight time of the two subsets; repeating the above steps to iteratively perform load balancing until the maximum difference in the estimated total flight time between all task subsets is less than a preset balancing threshold, and finally generating balanced task load distribution.

9. A processing system for route planning data for power line inspection, for implementing the method according to any one of claims 1 to 8, characterized in that: including: a data acquisition module for acquiring power line data, including line distribution data and terrain feature data; a region division and priority sorting module for dividing the entire power network region into multiple grids based on the line distribution data and determining the patrol priority of each grid to generate a region priority sequence; a task decomposition and distribution module configured to process and decompose the overall patrol task data based on the region priority sequence to generate a plurality of task subset data, and assign each of the UAVs with one of the task subset data, wherein each of the task subset data corresponds to a start location data and a target route segment data; an initial route planning module configured to process and generate, for each of the UAVs, initial route path data from the start location to the target route segment based on the terrain feature data; a space conflict identification module configured to analyze and process the initial route path data of all the UAVs to identify potential airspace intersection point data existing in the path data; a conflict resolution and route adjustment module configured to perform conflict prediction analysis on the identified potential airspace intersection point data, and execute route data adjustment to perform highly layered adjustment on the route path data of the relevant UAVs to generate a set of conflict-free route path data; a load balancing and scheme generation module configured to calculate, based on the set of conflict-free route path data, predicted flight time data and energy consumption data of each of the UAVs, and perform re-distribution processing on boundary task data between the task subset data according to a load balancing principle to finally generate balanced multi-UAV cooperative route scheme data.

Images