A path planning method and system for multi-uav inspection point-line hybrid targets

Abstract

The present application relates to data model and unmanned aerial vehicle control technical field, specifically to a kind of multi unmanned aerial vehicle inspection point line mixed target's path planning method and system, the method includes: according to the point target information and line target information in the region to be inspected, construct target information model;The number of unmanned aerial vehicle, performance parameter and target information model are acquired, according to the total flight distance, total flight time and energy consumption of all unmanned aerial vehicles to complete inspection task, construct the point line mixed target path planning model of multi unmanned aerial vehicle cooperative inspection;Intelligent optimization algorithm is used to solve the path planning model, generate initial inspection path planning scheme;Initial inspection path planning scheme is dynamically adjusted and optimized.The present application realizes the efficient cooperative inspection of multi unmanned aerial vehicle in point line mixed target scene by the complete process of constructing target information model, establishing multi-objective path planning model, intelligent optimization solution and real-time dynamic adjustment.

Description

Technical Field

[0001] This invention relates to the fields of data modeling and unmanned aerial vehicle (UAV) control technology, and in particular to a path planning method and system for mixed targets involving multiple UAV inspection points and lines. Background Technology

[0002] With its advantages of high mobility, rapid access to complex areas, and ability to carry multiple sensors for data collection, drones have gradually been widely used in the field of inspection. From the initial single-drone inspection, it has evolved to today's multi-drone collaborative inspection, aiming to further improve inspection efficiency, expand the inspection range, and enhance the reliability and flexibility of inspection.

[0003] While multi-UAV collaborative inspections offer numerous advantages, current multi-UAV collaborative path planning methods still face several unresolved issues. Most existing methods focus solely on minimizing total flight distance or time, neglecting the crucial factor of energy consumption. In practical inspections, energy consumption directly impacts the UAV's endurance and task completion. Inappropriate path planning can lead to rapid energy depletion, preventing the UAV from completing the inspection or even causing it to crash due to insufficient power, resulting in significant losses. Summary of the Invention

[0004] Therefore, it is necessary to provide a path planning method and system for mixed targets of multiple UAV inspection points and lines to address the above problems.

[0005] This invention is implemented as follows: a path planning method for mixed targets involving multiple UAV inspection points and lines, the path planning method for mixed targets involving multiple UAV inspection points and lines includes: Based on the point target information and line target information within the area to be inspected, construct a target information model that includes the coordinates of the point targets, the starting and ending coordinates of the line targets, and the topological relationship between the three. Acquire the number of drones, performance parameters, and target information model, and construct a point-line hybrid target path planning model for multi-drone collaborative inspection based on the total flight distance, total flight time, and energy consumption of all drones completing the inspection task. The path planning model is solved using an intelligent optimization algorithm to generate an initial inspection path planning scheme; By combining real-time environmental data, the initial inspection path planning scheme is dynamically adjusted and optimized, and multiple drones are controlled to complete collaborative inspection tasks based on the adjustment and optimization results.

[0006] In one embodiment, the present invention provides a path planning system for mixed targets of multiple UAV inspection points and lines, the path planning system for mixed targets of multiple UAV inspection points and lines includes a control center and a number of UAVs communicating with the control center; The control center is used to execute the path planning method for mixed targets of multiple UAV inspection points and lines as described in any embodiment of the present invention and to send the path planning scheme to the UAVs; The drone is used to receive and execute the path planning scheme sent by the control center.

[0007] The path planning method for multi-UAV inspection of mixed point and line targets provided in this invention achieves efficient collaborative inspection of multiple UAVs in mixed point and line target scenarios through a complete process of constructing a target information model, establishing a multi-target path planning model, intelligent optimization solution, and real-time dynamic adjustment. Its core advantages are: this method integrates point and line targets into the target information model, preserving the continuous characteristics of line targets and avoiding information loss caused by discretizing line targets; in the path planning model, it creatively uses total flight distance as the optimization objective, total flight time as the dynamic search guide, and total energy consumption as an absolute hard constraint, achieving hierarchical processing of the three physical quantities and fundamentally solving the dilemma of traditional linear weighted methods struggling to balance multiple optimization objectives; by combining real-time environmental data for dynamic adjustment, the path planning scheme has the ability to adapt to unexpected situations, significantly improving the efficiency and reliability of multi-UAV collaborative inspection. Attached Figure Description

[0008] Figure 1 This is a flowchart of a path planning method for a mixed target of multiple UAV inspection points and lines in one embodiment; Figure 2 This is a block diagram of the internal structure of a computer device in one embodiment. Detailed Implementation

[0009] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0010] It is understood that the terms "first," "second," etc., used in this invention may be used to describe various elements herein, but unless specifically stated otherwise, these elements are not limited by these terms. These terms are used only to distinguish one element from another. For example, without departing from the scope of this invention, a first script may be referred to as a second script, and similarly, a second script may be referred to as a first script.

[0011] like Figure 1As shown, in one embodiment, the present invention proposes a path planning method for mixed targets of multiple UAV inspection points and lines, which may specifically include the following steps: S1. Based on the point target information and line target information within the area to be inspected, construct a target information model that includes the coordinates of the point targets, the starting and ending coordinates of the line targets, and the topological relationship between the three. S2. Obtain the number of drones, performance parameters, and target information model. Based on the total flight distance, total flight time, and energy consumption of all drones completing the inspection task, construct a point-line hybrid target path planning model for multi-drone collaborative inspection. S3. Solve the path planning model using an intelligent optimization algorithm to generate an initial inspection path planning scheme; S4. Combine real-time environmental data to dynamically adjust and optimize the initial inspection path planning scheme, and control multiple drones to complete collaborative inspection tasks based on the adjustment and optimization results.

[0012] In this embodiment, the method aims to provide an efficient point-line mixed target path planning scheme for multi-UAV collaborative inspection, fundamentally solving the problems of traditional methods being unable to handle mixed target types, balancing multiple optimized targets, and lacking dynamic adaptability.

[0013] In step S1, the system first constructs a unified target information model based on point target information (such as the location coordinates of power towers) and line target information (such as the coordinates of the start and end points of transmission lines) within the area to be inspected. This model not only includes the coordinate information of various targets but also records the topological relationships between point targets and line targets, such as whether a point target is located on a line target or whether a line target passes near a point target. The purpose of this step is to transform the raw geographic information into structured data that the algorithm can process, laying the data foundation for subsequent path planning.

[0014] In step S2, the system constructs a path planning model for multi-drone collaborative inspection based on the number of drones (e.g., 3 drones), the performance parameters of each drone (such as maximum flight distance, maximum flight speed, and endurance), and the target information model built in step S1. This model uses the total flight distance of all drones completing the inspection task as the optimization objective, the total flight time as the dynamic search guide, and the total energy consumption as a hard constraint. The purpose of this is to find the feasible path with the shortest total distance by guiding the search direction with time information, while ensuring that energy consumption does not exceed limits.

[0015] In step S3, the system uses an intelligent optimization algorithm (such as an improved hybrid frog-leaping algorithm) to solve the path planning model constructed in step S2, and generates an initial inspection path planning scheme through iterative optimization. This scheme assigns a specific inspection target and flight path to each UAV.

[0016] In step S4, the system dynamically adjusts and optimizes the initial inspection path planning scheme by combining real-time environmental data monitored by sensors (such as wind speed, wind direction, and sudden obstacles), and controls multiple drones to actually perform collaborative inspection tasks based on the adjusted and optimized results. The purpose of this step is to enable the path planning scheme to respond to unexpected situations in the real environment, ensuring the smooth completion of the inspection task.

[0017] The path planning method for multi-UAV inspection of mixed point and line targets provided in this invention achieves efficient collaborative inspection of multiple UAVs in mixed point and line target scenarios through a complete process of constructing a target information model, establishing a multi-target path planning model, intelligent optimization solution, and real-time dynamic adjustment. Its core advantages are: this method integrates point and line targets into the target information model, preserving the continuous characteristics of line targets and avoiding information loss caused by discretizing line targets; in the path planning model, it creatively uses total flight distance as the optimization objective, total flight time as the dynamic search guide, and total energy consumption as an absolute hard constraint, achieving hierarchical processing of the three physical quantities and fundamentally solving the dilemma of traditional linear weighted methods struggling to balance multiple optimization objectives; by combining real-time environmental data for dynamic adjustment, the path planning scheme has the ability to adapt to unexpected situations, significantly improving the efficiency and reliability of multi-UAV collaborative inspection.

[0018] In one embodiment of the present invention, constructing a target information model including point target coordinates, line target start coordinates and end coordinates, and the topological relationship between the three includes: Using Geographic Information System (GIS) technology or manual measurement methods, obtain the coordinates of all point targets within the area to be inspected, and represent them in two-dimensional or three-dimensional coordinate form; Identify all line targets that can be formed from point targets, and record the coordinates of the starting and ending points. Also record one or more attribute information of the line target, such as its type, length, and direction. Add topological relationship descriptions between point targets and line targets to construct a target information model. The topological relationships include whether a point target is located on a line target and whether a line target passes near a point target.

[0019] In this embodiment, the specific implementation method of "constructing the target information model" in step S1 is defined. First, the system uses Geographic Information System (GIS) technology to extract the coordinate data of all point targets within the area to be inspected from an existing geographic information database, or obtains these coordinates through manual measurement methods (such as on-site surveys, satellite remote sensing positioning, etc.). The coordinates of the point targets can be represented in two-dimensional coordinate form (such as longitude and latitude) or three-dimensional coordinate form (such as longitude, latitude, and altitude), depending on the actual needs of the inspection task. For example, in a power inspection scenario, the point target can be the location of a power tower, and its coordinates can be directly obtained through a GIS system. Then, the system determines all line targets that can be formed based on the spatial distribution of the point targets and preset connection rules. For each line target, its starting point coordinates and ending point coordinates are recorded, and the system can also record attribute information such as the type of the line target (such as transmission line, pipeline, road, etc.), length (calculated from the starting point and ending point coordinates), and direction (such as north-south direction, east-west direction, etc.). For example, if a transmission line exists between two power towers, the coordinates of these two towers can form a line target. This line target is classified as a "transmission line," and its length is the straight-line distance between the two towers or the actual length of the line. Finally, the system adds a topological relationship description between point targets and line targets, forming a complete target information model. The topological relationship specifically includes: whether a point target is located on a line target (e.g., whether a tower is on a line), and whether a line target passes near a point target (e.g., the line is within a certain range of a device node). The purpose of this is to transform the raw spatial data into a structured information model containing rich semantic relationships, enabling subsequent path planning algorithms to accurately understand the spatial distribution and interrelationships of inspection targets, thereby formulating a more reasonable inspection path.

[0020] In one embodiment of the present invention, the step of acquiring the number of drones, performance parameters, and target information model, and constructing a point-line hybrid target path planning model for multi-drone collaborative inspection based on the total flight distance, total flight time, and energy consumption of all drones completing the inspection task, includes: Set the performance parameters of the drone, which include one or more of the following: maximum flight distance, maximum flight speed, minimum turning radius, endurance, and communication range; Based on the target information model, the inspection task is divided into point target tasks and line target tasks, and each UAV is assigned a number of point target tasks and / or a number of line target tasks. Using the total flight distance of all UAVs completing inspection tasks as the optimization objective, the total flight time as the dynamic search guide, and the total energy consumption as the absolute hard constraint, a complete point-line hybrid target path planning model for multi-UAV collaborative inspection is constructed.

[0021] In this embodiment, the specific implementation of "constructing a path planning model" in step S2 is defined. First, the system sets the performance parameters of each UAV participating in the inspection. These parameters specifically include: maximum flight distance (the maximum distance a UAV can fly after a single charge), maximum flight speed (the highest speed a UAV can reach under safe conditions), minimum turning radius (the minimum turning radius required for a UAV to change its flight direction, limited by the UAV's maneuverability), endurance time (the longest flight time a UAV can fly after a single charge), and communication range (the maximum distance a UAV can maintain communication with the control center or other UAVs). These parameters constitute the basic input data for the path planning model. Then, based on the target information model constructed in step S1, the system divides the inspection task into point target tasks (requiring the UAV to fly to precise coordinates for fixed-point detection) and line target tasks (requiring the UAV to continuously scan along the line segment direction). On this basis, the system assigns several point target tasks and / or several line target tasks to each UAV. Task allocation needs to consider the differences in the performance parameters of each UAV; for example, UAVs with strong endurance can be assigned more tasks or longer inspection routes. Finally, the system uses the total flight distance of all UAVs completing the inspection task as the optimization objective (i.e., minimizing the sum of the flight distances of all UAVs), the total flight time of all UAVs completing the inspection task as the dynamic search guide (i.e., using flight time information to guide the algorithm's direction and pace during the search process), and the total energy consumption of all UAVs completing the inspection task as an absolute hard constraint (i.e., the total energy consumption must not exceed the total energy limit of all UAVs, and the energy consumption of each UAV must not exceed its own maximum battery capacity). Simultaneously, other necessary constraints are constructed (such as safe distance constraints, obstacle avoidance constraints, etc.), ultimately forming a complete point-line hybrid target path planning model for multi-UAV collaborative inspection. The purpose of this is to transform the actual inspection problem into a mathematical optimization problem, enabling subsequent intelligent optimization algorithms to solve it.

[0022] In one embodiment of the present invention, the total flight distance of all UAVs completing the inspection task is used as the optimization target, i.e. , where n is the number of drones. Let be the flight distance of the i-th drone; Using the total flight time of all drones completing the inspection mission as a dynamic search guide, the frog is determined by the following formula: Among them, X w X: The position vector of the worst frog in the current subgroup; best_D X: The position of the frog with the shortest total distance in the current subgroup (spatial optimal solution); best_Tλ: The position of the frog with the shortest total flight time in the current subgroup (optimal speed solution); D λ is the gravitational coefficient learned from the "shortest distance solution". T The gravitational coefficient learned from the solution in the shortest time; The total energy consumption of all drones completing the inspection task is used as a hard constraint, i.e. E ( X )≤ E sum_max And any single machine Where E(X) represents the total energy consumption of all drones completing the inspection task. E sum_max E represents the maximum energy consumption for all drones to complete the inspection mission. k Energy consumption for a single drone to complete an inspection task. This represents the maximum energy consumption required for a single drone to complete an inspection task.

[0023] In this embodiment, the mathematical definitions and roles of the three core physical quantities in the path planning model are specifically defined. Total flight distance D (X) is set as the optimization objective, calculated as the sum of the flight distances of all drones. For example, if there are 3 drones with flight distances of 120 km, 95 km, and 110 km respectively, the total flight distance is 325 km. The fitness function of the intelligent optimization algorithm directly uses the total flight distance, meaning that solutions with shorter distances are considered better. The total flight time is set as a dynamic search guide and does not directly participate in fitness calculations; instead, it is used to adjust the update direction and stride of individual frogs in the algorithm. In the improved hybrid frog-jumping algorithm, the worst-performing frog X in each subgroup... w Instead of learning from a single "comprehensive optimal" solution, it learns simultaneously from two solutions with different physical optimalities: one is frog X, which has the shortest total flight distance in the subgroup. best_D The other is frog X, which has the shortest total flight time in the subgroup. best_T Gravitational coefficient λ D and λ T The distance D of the frog itself at the moment w and time T w The ratio is dynamically determined, and its calculation formula is: in, α i For adaptive shrinkage coefficient, ϵ For a very small positive number (such as 10) -10 (), used to prevent the denominator from being zero. When a frog's flight time... T w When it is too big, λD As the distance increases, the algorithm prioritizes guiding the learning towards the "shortest distance solution" to shorten the journey; when the flight distance of a frog... D w When it is too big, λ T To optimize speed, the algorithm prioritizes learning from the "shortest time solution." Total energy consumption is set as an absolute hard constraint; any solution whose total energy consumption after decoding exceeds the sum of the total battery capacities of all drones will be penalized. E sum_max Or the energy consumption of any drone E k Exceeding its own battery capacity If the distance is not specified, the solution is directly deemed infeasible and discarded. The purpose of this is to process the three physical quantities in layers: distance is the "optimization objective" (the object that the algorithm wants to minimize), time is the "search guide" (the sensor that the algorithm uses to determine direction and stride), and energy consumption is the "safety red line" (the physical limit that cannot be crossed). Each of the three performs its own function and works together to guide the algorithm to find a safe and efficient inspection path.

[0024] In one embodiment of the present invention, the constraint conditions include at least one of the following: Safety distance constraint: Ensures that the distance between any two drones is always greater than the set safety distance during flight. ,Right now: In the formula, and Let be the position coordinates of the i-th and j-th UAVs, respectively; Collision avoidance constraints: The drone must not collide with obstacles during flight. By establishing an obstacle model, it is ensured that the drone's flight path maintains a certain distance from the obstacles. Flight altitude constraint: The drone's flight altitude h must be within the set minimum flight altitude. and maximum flight altitude Between, that is ; Speed ​​constraint: The flight speed of the drone It needs to be at the set minimum flight speed and maximum flight speed Between, that is ; Task completion time constraint: The time for each drone to complete its assigned task cannot exceed its flight time. ,Right now ,in Let be the mission completion time of the i-th drone.

[0025] In this embodiment, various constraints that may be included in the path planning model are listed and explained in detail. The safety distance constraint requires that at any given moment during flight, the spatial distance between any two drones must be greater than a preset safety distance threshold. For example, if set If the distance is 50 meters, then the distance between any two drones cannot be less than 50 meters. This constraint is calculated using the Euclidean distance formula. ,in and These are the position coordinates of the i-th and j-th drones, respectively. The purpose of this constraint is to prevent drones from colliding in mid-air and ensure flight safety. The collision avoidance constraint requires that the drone's flight path cannot collide with any obstacles (such as buildings, mountains, trees, etc.). In practice, obstacle models need to be pre-built (e.g., abstracting obstacles as spheres, cylinders, or polyhedra), and then the path planning checks whether the drone's flight path intersects with these obstacle models; if they do, it is considered infeasible. The flight altitude constraint limits the range of the drone's flight altitude, i.e., the flight height... h Must be within the minimum flight altitude hmin (e.g., 50 meters to avoid ground obstacles) and maximum flight altitude hmax (For example, 300 meters, limited by drone performance and airspace control). Speed ​​constraints limit the range of drone flight speeds, i.e., speed... v Must be within the minimum flight speed vmin (e.g., 5 m / s, used to maintain flight stability) and maximum flight speed vmax (For example, within 30 meters per second, limited by the drone's performance). The task completion time constraint requires each drone to complete its assigned task within a certain total time. Ti It cannot exceed its battery life. Tmax This ensures that the drone has sufficient power to complete the inspection mission and return safely. One or more of these constraints can be flexibly selected based on the actual needs of the inspection mission, collectively ensuring the safety and feasibility of the path planning scheme.

[0026] In one embodiment of the present invention, the step of solving the path planning model using an intelligent optimization algorithm to generate an initial inspection path planning scheme includes: Based on the number of drones and the complexity of the inspection task, a certain number of initial populations are randomly generated. Each individual in the population represents a multi-drone inspection path planning scheme. Based on the constructed optimization objective and constraints, the fitness value of each individual is calculated. The higher the fitness value, the better the path planning scheme represented by that individual. The initial population is divided into multiple meme groups, each meme group representing an independent search subspace; Within each meme group, an improved individual learning strategy is used for local search to update the individual's position; The individuals in all meme groups are reshuffled to facilitate information exchange; A global search strategy is used to optimize the population, further improving its diversity. The algorithm parameters are dynamically adjusted based on the evolutionary state of the population. Determine if the termination condition is met. If it is, terminate the algorithm and output the optimal solution as the initial inspection path planning scheme. Otherwise, return to continue iterating.

[0027] In this embodiment, the specific process of "solving the path planning model using an intelligent optimization algorithm" in step S3 is described in detail. This process is based on an improved hybrid frog-leaping algorithm framework, and its core innovation lies in the deep integration of the physical characteristics (distance, time, energy consumption) of the path planning problem with the search operator of the algorithm, realizing adaptive evolution driven by physical information.

[0028] First, the system randomly generates an initial population (e.g., 200 individuals) based on the number of drones (which determines the code length) and the complexity of the inspection task (which determines the size of the search space). Each individual in the population (i.e., each "frog") represents a multi-drone inspection path planning scheme, and its encoding method adopts a hybrid discrete-continuous encoding scheme, as follows: Let the total number of task points that need to be visited (i.e., the endpoints of the point and line targets to be inspected) be . N Then the position vector X of each individual is a 2 N 3D continuous vectors: Among them, the sorted gene segments R j ( j =1, 2, ..., N The value range of this segment is [0,1]. This segment of the gene does not directly represent the access order, but rather represents "priority". During decoding, this segment is... N The values ​​are sorted in ascending order, and the corresponding task point indices after sorting represent the actual access order of the drone. For example, when... N When =5, if R =[0.3, 0.8, 0.1, 0.5, 0.6], after sorting, the index order is 3→1→4→5→2. Therefore, the actual waypoint order of the UAV is: starting point → task point 3 → task point 1 → task point 4 → task point 5 → task point 2 → destination. (Velocity gene segment) Vj ( j =1, 2, ..., N The range of values ​​for ) is [ V min , V max ],in V min and V max These represent the lower and upper limits of the drone's flight speed multiplier (relative to cruising speed), respectively, after decoding. V j The first in the corresponding access order j Flight speed multiplier for segment of flight, actual flight speed = V j × Cruise speed. This random key encoding method transforms the discrete waypoint ordering problem into a continuous sorting numerical optimization problem, allowing the standard leapfrog difference update formula to be directly applied without the need to design a dedicated discrete operator.

[0029] During population evolution, after an individual's position vector X is updated using a difference formula, new solutions are generated for various dimensions, which may then exceed their range. This method uniformly employs a boundary reflection strategy to handle this, rather than simple truncation. For the updated value of a certain dimension... x Its value range is [ Lower , Upper The update method for boundary reflections is as follows: If the value still exceeds the boundary after reflection, the dimension value is randomly reset to a random value within the boundary of that dimension. The boundary reflection strategy allows individuals searched near the boundary to be "bounced" back into the feasible region, rather than accumulating on the boundary, thus maintaining population diversity. Especially for ordination gene segments, the reflection mechanism allows the priority values ​​of each dimension to continuously and uniformly fluctuate within the range of [0, 1], avoiding the loss of ordination diversity due to a large number of individuals accumulating near 0 or 1.

[0030] Then, the system calculates the fitness value of each individual based on the optimization objective (minimizing total flight distance) and various constraints constructed in step S2. In this method, the fitness value is directly taken as the reciprocal (or negative) of the total flight distance, meaning the shorter the distance, the higher the fitness. For individuals that violate hard constraints (such as exceeding energy consumption limits), their fitness is set to a very small value (or infinity), causing them to be eliminated in subsequent selections. Next, the system sorts the initial population according to fitness and divides it into multiple meme groups (e.g., 10 groups, 20 individuals per group). Each meme group represents an independent search subspace, and parallel searches are conducted within different subspaces to maintain population diversity.

[0031] Within each meme group, the system employs an improved individual learning strategy for local search—that is, the worst individual in the group simultaneously learns from both the "shortest distance individual" and the "shortest time individual" within the group to update its own position. Specifically, for the worst frog XwX in the subgroup... w Its update formula is: After the update is complete, the newly generated X new The aforementioned boundary reflection strategy was used for repair.

[0032] During the local search process, this method also introduces a physical radius search mechanism that "expands outwards" to prevent the population from prematurely falling into local optima. Specifically, for the worst frog X in the current subpopulation... w With a certain probability P e，i (This probability is an adaptive parameter; see below for specific adjustment methods.) Exploration mode is triggered. When exploration mode is triggered, the above gravitational jump update formula is no longer used; instead, a new solution is generated using an adaptive expansion radius based on time deviation. Expansion radius R The calculation formula is: in, β T is the preset radius scaling factor (recommended value is 0.5 to 1.0). w This is the current flight time of the frog. This represents the average flight time of all frogs in the current subgroup. ϵ For extremely small positive numbers (such as 10) -10 (), used to prevent the denominator from being zero. When the flight time T of a frog is... w Much larger than the average value of the subgroup When T increases, the expansion radius R automatically increases, and the algorithm forces the generation of new random individuals over a large area around it to escape the current local optimum; when T... w When the solution approaches the subgroup mean, the expansion radius R automatically decreases, only performing minor perturbations. The formula for generating new solutions through the expansion search is: Among them, X max and X min These are the global upper and lower bound vectors for each dimension of the position vector, respectively, and N(0,1) is a standard normally distributed random number (mean 0, variance 1). This extended search mechanism enables the algorithm to adaptively adjust the exploration range based on the current individual's time performance deviation, avoiding the blindness of random exploration in traditional methods.

[0033] To facilitate the effective operation of the aforementioned local search and extended search mechanisms, this method also introduces a dual adaptive parameter control system to dynamically adjust key control parameters during the search process, as detailed below: (a) Adaptive shrinkage coefficient α i This coefficient controls the size of the frog's jump stride. It defines the frog's relative time lag. η i for: in, The shortest flight time in the current subgroup. This represents the longest flight time in the current subgroup. When T... w When approaching the optimal time for the subgroup η i →0; when T w When it is far worse than the optimal time for subgroups, η i →1. Adaptive contraction coefficient α i The calculation formula is: Where, α min and α max The lower and upper limits of the shrinkage coefficient are preset (recommended values ​​are 0.3 to 1.2), t t T is the current iteration number. max_iter The maximum number of iterations preset for the algorithm. This formula contains two dynamic factors: on the one hand, (1-t / T) max_iter The term decreases with the number of iterations, reflecting the idea of ​​simulated annealing—the algorithm explores extensively in the early stages and refines its extraction in the later stages; on the other hand, (1+ η i The item allows individuals who lag behind in time to receive a larger stride bonus (up to twice the base stride), thus achieving a "catch-up mechanism," while the stride of individuals who are ahead in time is compressed, with only minor adjustments.

[0034] (ii) Adaptive exploration probability P e，i This probability controls the likelihood of an individual entering an "expanding outwards" exploration mode. Define the energy urgency coefficient ψ for the current individual. i for: Among them, E remaining_i E represents the remaining battery power of the drone swarm corresponding to the current individual (if the solution corresponds to multiple drones, the average battery power of the swarm or the weighted sum of the remaining battery power of each drone can be used). max_i This represents the sum of the maximum battery capacities corresponding to the drone. When the battery is fully charged, ψ i→0 (Not urgent); When the battery is low, ψ i →1 (Extremely urgent). Adaptive exploration probability P e,i The calculation formula is: Among them, P min and P max To explore the preset lower and upper limits of probability (suggested values ​​are 0.05 to 0.45), σ(t) is the global stagnation trigger factor. When a drone (or swarm) has sufficient remaining power (ψ... i (small), its exploration probability P e,i Automatically adjusts to higher settings, tending to boldly expand the search to the surrounding area; when the battery is almost depleted (ψ i If the value is large, its exploration probability is automatically reduced to the minimum, forcing it to move closer to the current optimal solution to ensure safety. The global stagnation trigger factor σ(t) is defined as: Where, ΔD gbest The change in the global optimal distance over consecutive iterations is ε, where ε is a preset minimum positive threshold (e.g., 10). -4 K is the preset stagnation tolerance generation (e.g., 20 to 30 generations). When the population is detected to be stagnant (the global optimum has not improved significantly for K consecutive generations), σ(t) automatically increases to 1.0, forcibly increasing the exploration probability of the entire population and forcibly "exploding" the population to escape the local optimum.

[0035] (iii) α i With P e,i Coordination of Exploration Mode and Large Step Size: To prevent severe oscillations in the solution caused by the simultaneous activation of exploration mode and large step size, this method introduces mutually exclusive coupling logic. When a frog triggers exploration mode (i.e., rand()), <P e,i When rand() is a random number uniformly distributed in the interval [0,1], the effective shrinkage coefficient α of this iteration is forcibly increased. i Reduced to 50% of the original value: This mechanism ensures that the search in exploration mode adopts a "small steps but large directional changes" approach, while the search in development mode adopts a "large steps" approach (large steps to converge to the optimal point), with the two working together without interfering with each other.

[0036] After local search and adaptive parameter adjustment, the system reshuffles all individuals in the meme groups (i.e., a "mixing" step), disrupting the original groupings to facilitate information exchange between different subspaces. Then, the system employs a global search strategy (e.g., fine-grained search near the global optimum or introducing mutation operations) to further optimize the population, increasing its diversity and preventing premature convergence. Finally, the system determines whether the termination condition is met (e.g., reaching the maximum number of iterations or the global optimum showing no improvement for multiple consecutive generations). If met, the algorithm terminates, outputting the currently found optimal solution as the initial inspection path planning scheme; otherwise, it returns to continue iterating.

[0037] In one embodiment of the present invention, the step of dynamically adjusting and optimizing the initial inspection path planning scheme by combining real-time environmental data includes: The system uses sensors to monitor environmental data in real time and transmits the data to the control center via wireless communication technology. The control center determines whether to trigger the replanning mechanism based on the monitored environmental data. When the replanning mechanism is triggered, the local environment model is reconstructed based on the current environmental data, and combined with the unfinished task information in the initial inspection path planning scheme, a fast optimization algorithm is used to replan the remaining inspection paths. The replanned path is integrated with the completed path to obtain the final path planning scheme for the mixed target of multiple UAV inspection points and lines.

[0038] In this embodiment, the overall process of "dynamic adjustment and optimization" in step S4 is defined, clarifying the complete closed loop from environmental monitoring to path replanning. First, various sensors deployed on the drone or within the inspection area (such as wind speed sensors, cameras, radar, etc.) monitor environmental data in real time, including meteorological conditions (wind speed, wind direction, temperature, etc.), obstacle information (suddenly appearing buildings, moving objects, etc.), and the drone's own status (remaining battery power, equipment health status, etc.). This data is transmitted in real time to the ground control center via wireless communication technologies (such as 4G / 5G, LoRa, Wi-Fi, etc.). Then, the control center analyzes and evaluates the received environmental data to determine whether to trigger the replanning mechanism. Triggering conditions may include: wind speed exceeding a safety threshold, discovery of new obstacles, a drone being unable to continue its mission for any reason, or a change in the mission objective. When the replanning mechanism is triggered, the control center reconstructs a local environmental model near the mission area that was not completed in the initial path planning scheme based on the latest environmental data (e.g., updating the obstacle distribution and meteorological conditions in that area). By combining the incomplete task information from the initial inspection path planning scheme (such as which point and line targets have not been inspected), a fast optimization algorithm (such as the dynamic window method or an improved A* algorithm) is used to replan the remaining inspection paths, generating updated path segments. Finally, the replanned paths are integrated with the paths that have already been flown to form a complete and updated multi-UAV inspection path planning scheme, which is then distributed to each UAV for execution. The purpose of this is to enable the path planning scheme to respond in real time to dynamic environments. When the environmental conditions on which the initial planning was based change, the system can quickly adjust to ensure that the inspection task can continue to be completed safely and efficiently in changing environments.

[0039] In one embodiment of the present invention, the step of replanning the remaining inspection path using a fast optimization algorithm includes: The Dynamic Window Method (DWA) is used to search for feasible combinations of velocity and angular velocity in the velocity space based on the current state of the UAV, generating a series of candidate trajectories. Evaluate each candidate trajectory; Select the candidate trajectory with the best evaluation results as the next flight path for the UAV; Repeat the above steps until the remaining inspection tasks are completed.

[0040] In this embodiment, a specific implementation method for "replanning the remaining inspection path using a fast optimization algorithm" in step S4 is specified, namely, the Dynamic Window Approach (DWA). DWA is a local path planning method suitable for drones and mobile robots, and its core idea is to search in velocity space. Specifically, the system first obtains the current state of the drone, including its current position, current speed, and current orientation. Then, within the feasible dynamic range of the drone's speed and angular velocity (limited by the drone's maximum acceleration and maximum deceleration), multiple combinations of speed and angular velocity are sampled. For each sampled (speed, angular velocity) combination, the system simulates the drone's trajectory over a future period, generating a series of candidate trajectories (e.g., generating 100 candidate trajectories). Next, the system evaluates each candidate trajectory. Evaluation metrics typically include: whether the trajectory is safe (whether it collides with obstacles), the smoothness of the trajectory, and the forward efficiency of the trajectory relative to the target point (i.e., whether it is moving towards the unfinished inspection target). The system calculates a comprehensive score for each candidate trajectory based on these metrics. Finally, the system selects the candidate trajectory with the highest overall score as the UAV's next actual flight path. After the UAV executes this path segment, the system repeats the above process—re-acquiring the current state, re-sampling the velocity combination, and re-generating and evaluating candidate trajectories—until all remaining inspection tasks are completed. The purpose of this is to provide a computationally efficient and fast local path planning solution when the environment dynamically changes, enabling the UAV to avoid sudden obstacles in real time and continuously move towards the target point.

[0041] In one embodiment of the present invention, the step of replanning the remaining inspection path using a fast optimization algorithm includes: Using the improved A* algorithm, in the reconstructed local environment model, the current position of the UAV is taken as the starting point, and the remaining inspection target points are taken as the target points; During the search process, a heuristic function is introduced, which comprehensively considers the distance from the UAV to the target point and environmental risk factors to guide the search direction; By continuously expanding the nodes and comparing the cost functions, the optimal path from the starting point to the target point is found, which serves as the remaining inspection path for the drone.

[0042] In this embodiment, another specific implementation method for "replanning the remaining inspection path using a fast optimization algorithm" in step S4 is specified, namely, using the improved A* algorithm. 。The A* algorithm is a classic heuristic search algorithm widely used in path planning. Specifically, the system first sets the UAV's current position as the starting point of the search in a reconstructed local environment model (which includes updated obstacle distribution and terrain information), and sets the incomplete inspection target points (including endpoints of point and line targets) as the search target points. Then, the system initiates the A* algorithm's search process. During the search, the algorithm introduces a heuristic function to guide the search direction. This heuristic function considers multiple factors: the Euclidean distance from the UAV's current position to the target point (the closer the distance, the smaller the heuristic value), environmental risk factors along the path (such as passing through high-risk areas increasing the heuristic value), etc. Through this heuristic function, the algorithm can prioritize expanding nodes that are more likely to lead to the target point. The algorithm continuously expands nodes, calculating the cost function for each expanded node. f ( n )= g ( n )+ h ( n ),in g ( n () represents the distance from the starting point to the current node. n The actual cost, h ( n ) is from the current node n The algorithm uses a heuristic cost estimation method to reach the target point. By comparing the cost function values ​​of each node, it selects the node with the lowest cost for expansion until the optimal path from the starting point to the target point is found. The final output path is the remaining inspection path of the UAV. The purpose of this is to provide a replanning scheme that guarantees optimality when the environment undergoes local changes, and it is suitable for scenarios with relatively gentle environmental changes and sufficient computation time.

[0043] In one embodiment of the present invention, a path planning system for a mixed target of multiple UAV inspection points and lines is also provided. The path planning system for a mixed target of multiple UAV inspection points and lines includes a control center and a number of UAVs communicating with the control center. The control center is used to execute the path planning method for mixed targets of multiple UAV inspection points and lines as described in any of the above embodiments and send the path planning scheme to the UAVs; The drone is used to receive and execute the path planning scheme sent by the control center.

[0044] In this embodiment, the aforementioned method embodiments are concretized as a physical system. This system mainly consists of two core components: a control center and several drones. Each drone maintains a connection with the control center via wireless communication methods (such as 4G / 5G, LoRa, Wi-Fi, etc.). The control center can be a ground-based server or a high-performance computing device, internally containing computer program code implementing any of the above method embodiments. During operation, the control center executes the path planning method for multi-drone inspection of mixed targets (points and lines) as described in any embodiment of this invention, including constructing a target information model, constructing a path planning model, using intelligent optimization algorithms to generate an initial plan, and dynamically adjusting and optimizing based on real-time environmental data. After completing the path planning, the control center sends the generated path planning scheme (including the flight waypoint sequence of each drone, flight speed of each segment, inspection target order, etc.) to each drone via a wireless communication link. Each drone is equipped with a flight control system and a communication module to receive the path planning scheme sent by the control center and autonomously fly according to the path and speed specified in the scheme to complete its respective inspection task. During flight, the drone can also collect environmental data through onboard sensors and transmit it back to the control center, providing data support for dynamic adjustments. Through this close integration of hardware and software, the path planning method is implemented into a deployable and operational drone inspection system, enabling multiple drones to conduct efficient, safe, and adaptive collaborative inspections in complex scenarios with mixed point and line targets.

[0045] Figure 2 An internal structural diagram of a computer device in one embodiment is shown, which can serve as the control center in a path planning system for mixed targets (multiple UAV inspection points and lines). Figure 2 As shown, the computer device includes a processor, memory, network interface, input device, and display screen connected via a system bus. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and may also store a computer program. When executed by the processor, this computer program enables the processor to implement the path planning method for mixed targets of multiple UAV inspection points and lines provided in this embodiment of the invention. The internal memory may also store a computer program. When executed by the processor, this computer program enables the processor to implement the path planning method for mixed targets of multiple UAV inspection points and lines provided in this embodiment of the invention. The display screen of the computer device can be a liquid crystal display screen or an e-ink display screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, a trackball, or a touchpad provided on the casing of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0046] Those skilled in the art will understand that Figure 2The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer device to which the present invention is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0047] In one embodiment, a computer device is provided, the computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, performs the following steps: Based on the point target information and line target information within the area to be inspected, construct a target information model that includes the coordinates of the point targets, the starting and ending coordinates of the line targets, and the topological relationship between the three. Acquire the number of drones, performance parameters, and target information model, and construct a point-line hybrid target path planning model for multi-drone collaborative inspection based on the total flight distance, total flight time, and energy consumption of all drones completing the inspection task. The path planning model is solved using an intelligent optimization algorithm to generate an initial inspection path planning scheme; By combining real-time environmental data, the initial inspection path planning scheme is dynamically adjusted and optimized, and multiple drones are controlled to complete collaborative inspection tasks based on the adjustment and optimization results.

[0048] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, causes the processor to perform the following steps: Based on the point target information and line target information within the area to be inspected, construct a target information model that includes the coordinates of the point targets, the starting and ending coordinates of the line targets, and the topological relationship between the three. Acquire the number of drones, performance parameters, and target information model, and construct a point-line hybrid target path planning model for multi-drone collaborative inspection based on the total flight distance, total flight time, and energy consumption of all drones completing the inspection task. The path planning model is solved using an intelligent optimization algorithm to generate an initial inspection path planning scheme; By combining real-time environmental data, the initial inspection path planning scheme is dynamically adjusted and optimized, and multiple drones are controlled to complete collaborative inspection tasks based on the adjustment and optimization results.

[0049] It should be understood that although the steps in the flowcharts of the various embodiments of the present invention are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the various embodiments may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of the sub-steps or stages of other steps. Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0050] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0051] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A path planning method for mixed targets involving multiple UAV inspection points and lines, characterized in that, The path planning method for mixed targets involving multiple UAV inspection points and lines includes: Based on the point target information and line target information within the area to be inspected, construct a target information model that includes the coordinates of the point targets, the starting and ending coordinates of the line targets, and the topological relationship between the three. Acquire the number of drones, performance parameters, and target information model, and construct a point-line hybrid target path planning model for multi-drone collaborative inspection based on the total flight distance, total flight time, and energy consumption of all drones completing the inspection task. The path planning model is solved using an intelligent optimization algorithm to generate an initial inspection path planning scheme; By combining real-time environmental data, the initial inspection path planning scheme is dynamically adjusted and optimized, and multiple drones are controlled to complete collaborative inspection tasks based on the adjustment and optimization results.

2. The path planning method for mixed targets of multiple UAV inspection points and lines according to claim 1, characterized in that, The construction of the target information model, which includes the coordinates of point targets, the starting and ending coordinates of line targets, and the topological relationship between the three, includes: Using Geographic Information System (GIS) technology or manual measurement methods, obtain the coordinates of all point targets within the area to be inspected, and represent them in two-dimensional or three-dimensional coordinate form; Identify all line targets that can be formed from point targets, and record the coordinates of the starting and ending points. Also record one or more attribute information of the line target, such as its type, length, and direction. Add topological relationship descriptions between point targets and line targets to construct a target information model. The topological relationships include whether a point target is located on a line target and whether a line target passes near a point target.

3. The path planning method for mixed targets of multiple UAV inspection points and lines according to claim 1, characterized in that, The model for acquiring the number of drones, performance parameters, and target information, and constructing a point-line hybrid target path planning model for multi-drone collaborative inspection based on the total flight distance, total flight time, and energy consumption of all drones completing inspection tasks, includes: Set the performance parameters of the drone, which include one or more of the following: maximum flight distance, maximum flight speed, minimum turning radius, endurance, and communication range; Based on the target information model, the inspection task is divided into point target tasks and line target tasks, and each UAV is assigned a number of point target tasks and / or a number of line target tasks. Using the total flight distance of all UAVs completing inspection tasks as the optimization objective, the total flight time as the dynamic search guide, and the total energy consumption as the absolute hard constraint, a complete point-line hybrid target path planning model for multi-UAV collaborative inspection is constructed.

4. The path planning method for mixed targets of multiple UAV inspection points and lines according to claim 3, characterized in that, The total flight distance of all drones completing inspection tasks is used as the optimization objective, i.e. , where n is the number of drones. Let be the flight distance of the i-th drone; Using the total flight time of all drones completing the inspection mission as a dynamic search guide, the frog is determined by the following formula: Among them, X w X: The position vector of the worst frog in the current subgroup; best_D X: The position of the frog with the shortest total distance in the current subgroup (spatial optimal solution); best_T λ: The position of the frog with the shortest total flight time in the current subgroup (optimal speed solution); D λ is the gravitational coefficient learned from the "shortest distance solution". T The gravitational coefficient learned from the solution in the shortest time; The total energy consumption of all drones completing the inspection task is used as a hard constraint, i.e. E ( X )≤ E sum_max And any single machine Where E(X) represents the total energy consumption of all drones completing the inspection task. E sum_max E represents the maximum energy consumption for all drones to complete the inspection mission. k Energy consumption for a single drone to complete an inspection task. This represents the maximum energy consumption required for a single drone to complete an inspection task.

5. The path planning method for mixed targets of multiple UAV inspection points and lines according to claim 3, characterized in that, The constraints include at least one of the following: Safe distance constraint: Ensure that the distance between any two drones is always greater than the set safe distance during flight. ,Right now: In the formula, and Let be the position coordinates of the i-th and j-th UAVs, respectively; Collision avoidance constraints: The drone must not collide with obstacles during flight. By establishing an obstacle model, it is ensured that the drone's flight path maintains a certain distance from the obstacles. Flight altitude constraint: The flight altitude h of the drone must be at the set minimum flight altitude and the maximum flight altitude , that is ; Speed ​​constraint: The flight speed of the drone It needs to be at the set minimum flight speed and maximum flight speed Between, that is ; Task completion time constraint: The time for each drone to complete its assigned task cannot exceed its flight time. ,Right now ,in Let be the mission completion time of the i-th drone.

6. The path planning method for mixed targets of multiple UAV inspection points and lines according to claim 1, characterized in that, The step of using an intelligent optimization algorithm to solve the path planning model and generate an initial inspection path planning scheme includes: Based on the number of drones and the complexity of the inspection task, a certain number of initial populations are randomly generated. Each individual in the population represents a multi-drone inspection path planning scheme. Based on the constructed optimization objective and constraints, the fitness value of each individual is calculated. The higher the fitness value, the better the path planning scheme represented by that individual. The initial population is divided into multiple meme groups, each meme group representing an independent search subspace; Within each meme group, an improved individual learning strategy is used for local search to update the individual's position; The individuals in all meme groups are reshuffled to facilitate information exchange; A global search strategy is used to optimize the population, further improving its diversity. The algorithm parameters are dynamically adjusted based on the evolutionary state of the population. Determine if the termination condition is met. If it is, terminate the algorithm and output the optimal solution as the initial inspection path planning scheme. Otherwise, return to continue iterating.

7. The path planning method for mixed targets of multiple UAV inspection points and lines according to claim 1, characterized in that, The dynamic adjustment and optimization of the initial inspection path planning scheme based on real-time environmental data includes: The system uses sensors to monitor environmental data in real time and transmits the data to the control center via wireless communication technology. The control center determines whether to trigger the replanning mechanism based on the monitored environmental data. When the replanning mechanism is triggered, the local environment model is reconstructed based on the current environmental data, and combined with the unfinished task information in the initial inspection path planning scheme, a fast optimization algorithm is used to replan the remaining inspection paths. The replanned path is integrated with the completed path to obtain the final path planning scheme for the mixed target of multiple UAV inspection points and lines.

8. The path planning method for mixed targets of multiple UAV inspection points and lines according to claim 7, characterized in that, The process of replanning the remaining inspection paths using a fast optimization algorithm includes: The Dynamic Window Method (DWA) is used to search for feasible combinations of velocity and angular velocity in the velocity space based on the current state of the UAV, generating a series of candidate trajectories. Evaluate each candidate trajectory; Select the candidate trajectory with the best evaluation results as the next flight path for the UAV; Repeat the above steps until the remaining inspection tasks are completed.

9. The path planning method for mixed targets of multiple UAV inspection points and lines according to claim 7, characterized in that, The process of replanning the remaining inspection paths using a fast optimization algorithm includes: Using the improved A* algorithm, in the reconstructed local environment model, the current position of the UAV is taken as the starting point, and the remaining inspection target points are taken as the target points; During the search process, a heuristic function is introduced, which comprehensively considers the distance from the UAV to the target point and environmental risk factors to guide the search direction; By continuously expanding the nodes and comparing the cost functions, the optimal path from the starting point to the target point is found, which serves as the remaining inspection path for the drone.

10. A path planning system for mixed targets involving multiple unmanned aerial vehicle (UAV) inspection points and lines, characterized in that, The path planning system for mixed targets of multiple UAV inspection points and lines includes a control center and several UAVs that communicate with the control center. The control center is used to execute the path planning method for mixed targets of multiple UAV inspection points and lines as described in any one of claims 1-9 and send the path planning scheme to the UAVs; The drone is used to receive and execute the path planning scheme sent by the control center.

Images