Unmanned aerial vehicle coverage path planning method

Abstract

The invention discloses an unmanned aerial vehicle coverage path planning method, and belongs to the technical field of unmanned aerial vehicle path planning, and the method comprises the steps: selecting the optimal direction of route point generation through calculating the number of grid points generated in different directions; constructing a waypoint generation mechanism based on energy consumption optimization, quantifying the weight of waypoints, and converting a waypoint optimization problem into a weighted vertex coverage problem; deducing a group of optimized waypoints by solving a weighted vertex coverage problem, and minimizing the total weight of the coverage waypoints while ensuring that the coverage task is completed; and solving an optimization problem, performing path planning based on the generated waypoints, and further optimizing the flight path. According to the method, the path planning efficiency can be remarkably improved, the energy consumption can be greatly reduced, the solving capability of a complex coverage path planning problem is improved, an efficient solution is provided for an energy sensitive task, and the potential of a path planning algorithm in practical application is further expanded.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) path planning, and more particularly to a UAV coverage path planning method based on an alternating hierarchical genetic algorithm based on game learning. Background Art

[0002] With the rapid development of drone technology, coverage path planning has garnered widespread attention across multiple fields. This problem is particularly crucial for remote environmental monitoring and visual inspection of large, three-dimensional structures. These missions often involve complex terrain and structures, requiring drones to efficiently and comprehensively cover the target area. In certain specific application scenarios, such as large-scale searches or patrols in mountainous areas at night, drones need to be equipped with photographic sensors and searchlights to ensure adequate illumination during the search. This configuration not only improves search accuracy but also enhances the drone's ability to operate in low-light environments and its perception capabilities in complex environments, thereby solving specific practical problems. LiDAR can help rescuers better understand the disaster area. Furthermore, the introduction of LiDAR sensors can significantly enhance the practical application of drones in search and rescue operations in urban or mountainous areas. Using drones for monitoring and search in uninhabited areas has always been a significant practical issue. The application of unmanned platforms for visual coverage can enable dynamic monitoring and rescue operations across large areas shortly after a disaster. Drones offer advantages such as autonomous takeoff and landing, a bird's-eye view, agility, and high maneuverability, enabling them to excel in a variety of complex environments. Given the autonomous nature of drones, mission planning becomes crucial. Reasonable planning ensures that drones can complete their missions efficiently. Consequently, drones are increasingly being used to monitor unmanned areas, and coverage path planning has become a key technology for achieving complete coverage of the surveillance area.

[0003] Coverage path planning typically consists of two main steps: waypoint generation and path planning. Its core goal is to plan an optimal path that satisfies multiple constraints, ensuring that the UAV's sensors fully cover the mission area. Appropriate trajectory planning not only shortens the UAV's flight distance but also effectively mitigates potential threats encountered during flight. For example, in mountainous or urban environments, UAVs need to avoid obstacles and adapt to changing weather conditions. In recent years, coverage path planning has been successfully applied in various fields, such as regional exploration and agricultural operations. In regional exploration, UAVs can efficiently cover large areas and provide detailed geographic information. In agricultural operations, UAVs can achieve efficient crop monitoring and pesticide spraying through precise path planning. These applications fully demonstrate the important role of coverage path planning in improving the efficiency and effectiveness of UAV missions. Although most studies have successfully implemented coverage path planning for UAVs, they focus on ensuring complete coverage of the mission area and provide little insight into how waypoint generation affects path planning and optimizes UAV energy consumption.

[0004] Through the above analysis, the problems and defects of the existing technology mainly include:

[0005] (1) System efficiency: Current coverage path planning systems often ignore the impact of waypoint generation on path planning and drone energy consumption while ensuring complete coverage of the mission area. Although existing methods have achieved certain results in two-dimensional and three-dimensional coverage tasks, they still have shortcomings in energy management in complex environments. How to further improve the efficiency of the system while ensuring coverage is an urgent problem to be solved.

[0006] (2) Flexibility of path planning: Most existing coverage path planning methods use a "zigzag scanning" approach to generate fixed waypoints, lacking the foresight to consider the drone's energy consumption in advance when selecting waypoints. In practical applications, drones need to adjust waypoints according to the environment, and existing methods are often unable to cope with this. Therefore, designing a planning method that can prioritize the selection of better waypoints is the key to improving the flexibility of drone planning.

[0007] (3) Energy consumption optimization: UAV energy management is an important part of coverage path planning. Existing methods have shortcomings in energy consumption optimization. Although some studies have proposed path planning methods to reduce energy consumption, these methods often ignore the impact of waypoint generation on energy consumption. How to further optimize the energy consumption of UAVs while ensuring coverage is an important research direction.

[0008] Due to the complex environments and real-time mission requirements, existing coverage path planning methods often struggle to balance efficiency and the dual optimization of path length and energy consumption in practical applications. Addressing this issue requires a new coverage path planning method that comprehensively considers minimizing path length and optimizing energy consumption, which is the focus and difficulty of current research. Summary of the Invention

[0009] In view of this, the present invention provides a method for UAV coverage path planning that solves at least some of the above-mentioned technical problems. First, the coverage path planning is decomposed into: waypoint generation and path planning. Specifically, in the waypoint generation stage, the optimal direction for waypoint generation is selected by calculating the number of grid points generated in different directions; a waypoint generation mechanism based on energy consumption optimization is constructed, the weights of the waypoints are quantified, and the waypoint optimization problem is converted into a weighted vertex cover problem; a set of optimized waypoints is derived by solving the weighted vertex cover problem, ensuring that the total weight of the coverage waypoints is minimized while the coverage task is completed; in the path planning stage, the optimization problem is solved, and path planning is performed based on the generated waypoints to further optimize the flight path. The present invention can significantly improve the efficiency of path planning and significantly reduce energy consumption, improve the ability to solve complex coverage path planning problems, and provide an efficient solution for energy-sensitive tasks.

[0010] To achieve the above object, the technical solution adopted by the present invention is:

[0011] In a first aspect, the present invention provides a method for planning a UAV coverage path. In this method, UAV coverage path planning is decomposed into: waypoint generation and path planning; wherein waypoint generation includes the following steps 1 to 3, and path planning includes the following steps 4 to 5:

[0012] Step 1: Select the best direction for generating waypoints by calculating the number of grid points generated in different directions;

[0013] Step 2: Use the energy consumption during waypoint transfer to get the average mapping to obtain the waypoint weight, and convert the waypoint optimization into weighted vertex cover;

[0014] Step 3: Solve the near-optimal solution of weighted vertex cover through a game algorithm based on temporal difference to obtain the optimized waypoint sequence;

[0015] Step 4: Construct the optimization objectives for the UAV coverage path planning, which include minimizing the path length and minimizing the energy consumption;

[0016] Step 5: According to the optimization goal of UAV coverage path planning, the alternating layered genetic algorithm is used to solve the optimized waypoint sequence to obtain the optimal coverage path.

[0017] Furthermore, in this method, sensing constraints and assumptions are first set to analyze and apply to actual situations, where:

[0018] Constraints include:

[0019] Equidistant: Set a normal vector perpendicular to the surface for each point on the surface, and keep a constant distance along the normal vector;

[0020] Front: Point the drone's sensor device toward the terrain surface to be covered, and keep it perpendicular to the terrain surface at all times;

[0021] Assumptions include:

[0022] The yaw angle of the sensing device is consistent with that of the drone;

[0023] The radius of the searchlight is less than half the width of the photographic picture.

[0024] Furthermore, in step 1, for the convex polygonal surface, the optimal direction for generating waypoints is selected by calculating the number of grid points generated in different directions. The specific process includes:

[0025] Arrange all edges in descending order of length, then traverse each edge and calculate its maximum distance to all vertices, and determine the number of vertical and horizontal dividing lines based on this distance and the length of the edge;

[0026] Calculate the number of waypoints for each edge and update the minimum number of waypoints during traversal;

[0027] The direction that can generate the least number of grid points is determined as the optimal direction for generating waypoints.

[0028] Furthermore, in step 2, the weight of the waypoint is obtained by averaging the energy consumption during the waypoint transfer, and the waypoint optimization is converted into a weighted vertex cover, which specifically includes:

[0029] (1) For points a and b on the terrain surface, calculate the slope angle between points a and b. The calculation formula is:

[0030]

[0031] Where, represents the slope angle between point a and point b, represents the absolute height difference between point a and point b, Represents the distance between point a and point b;

[0032] (2) Calculate the energy from point a to point b using the following formula:

[0033]

[0034] Where, Represents the energy value from point a to point b, represents the coefficients measured by linear fitting;

[0035] (3) Let the neighboring waypoint of waypoint i be j, and calculate the energy consumption from neighboring waypoint j to i. The calculation formula is:

[0036]

[0037] Where, represents the energy consumption of neighbor waypoint j to i, represents the coefficients measured by linear fitting, represents the path length from neighbor waypoint i to j, represents the slope angle from neighbor waypoint j to i;

[0038] (4) Take the average energy consumption of all neighboring waypoints j to i, and map this value to the weight of waypoint i.

[0039] Furthermore, in step 3, the game algorithm based on temporal difference is an algorithm constructed by combining the temporal difference method with the asymmetric snowdrift game. The algorithm is as follows:

[0040] 1) If , then:

[0041]

[0042]

[0043] Where, represents the payoff when the player chooses to cooperate at time t, represents the payoff when the player chooses to cooperate at time t+1, represents the benefit of waypoint i at time t, represents the profit transition variable at time t+1 when the player chooses to cooperate, is the learning rate, represents the payoff when the player chooses to cooperate at the waypoint i at time k+1, C represents cooperation, and D represents betrayal. and denote the probability of cooperation and defection, represents the difference in returns between time k+1 and time k, γ represents the discount factor, represents the benefit of waypoint i at time k;

[0044] 2) If , then:

[0045]

[0046]

[0047] Where, represents the payoff when the player chooses to betray at time t, represents the payoff transition variable at time t+1 when the player chooses to defect, represents the payoff when the player chooses to betray at time k+1 at waypoint i, represents the benefit of waypoint i when it chooses to defect at time k+1, It represents the difference in returns between time t+1 and time t.

[0048] Furthermore, in step 4, the expression for minimizing the path length is:

[0049]

[0050]

[0051] Where L represents the path length; represents the path length of the UAV from waypoint i to j; is the total number of waypoints. If the UAV travels from i to j, then ;otherwise, ; i and j represent waypoints;

[0052] The expression for minimizing energy consumption is:

[0053]

[0054]

[0055] Where, E represents the energy consumption; Indicates the turning angle from point i to point k through point j; represents the path length from waypoint j to k, such as Indicates the path length from waypoint 1 to waypoint 2; similarly, Indicates from waypoint To waypoint Path length; Represents the decision variable. If the drone flies from i to j and then to k, then ;otherwise, .

[0056] Furthermore, in step 5, an alternating layered genetic algorithm is used to solve the optimal coverage path based on the optimized waypoint sequence. The process includes:

[0057] (1) Initialize the waypoint sequence population;

[0058] (2) In the first layer of operations, the fitness value is calculated using the path length. The tournament selection, sequential crossover, and simple mutation are run multiple times to obtain a waypoint sequence with the minimum relative path length, and this sequence is put into the population.

[0059] (3) In the second layer of operations, the fitness value is calculated using energy consumption, and the tournament selection, sequential crossover, and simple mutation are run multiple times to obtain a waypoint sequence with the minimum relative path length, which is then put into the population;

[0060] (4) Return to step (2) and perform the calculation again to obtain the waypoint sequence with the minimum path length and energy consumption, and obtain the optimal coverage path.

[0061] In a second aspect, the present invention further provides a UAV coverage path planning system, which is applied to the above-mentioned UAV coverage path planning method to perform UAV coverage path planning, and the system includes:

[0062] The waypoint generation module is used to select the optimal direction for waypoint generation by calculating the number of grid points generated in different directions. The module also uses the average energy consumption during waypoint transfer to obtain the waypoint weights, transforming the waypoint optimization into a weighted vertex cover problem. A near-optimal solution to the weighted vertex cover problem is solved using a temporal difference-based game algorithm to obtain an optimized waypoint sequence.

[0063] The path planning module is used to construct the optimization objectives of the UAV coverage path planning. The optimization objectives include minimizing the path length and minimizing the energy consumption. According to the optimization objectives of the UAV coverage path planning, an alternating layered genetic algorithm is used to solve the problem based on the optimized waypoint sequence to obtain the optimal coverage path.

[0064] In a third aspect, an embodiment of the present invention further provides an electronic device comprising a processor and a memory, wherein the memory stores machine-executable instructions that can be executed by the processor, and the processor executes the machine-executable instructions to implement the above-mentioned drone coverage path planning method.

[0065] As can be seen from the above technical solutions, the present invention provides a method and system for planning a UAV coverage path. Compared with the prior art, the present invention has at least the following beneficial effects:

[0066] 1. This invention employs an alternating hierarchical genetic algorithm (AGA) based on game learning for UAV coverage path planning. By decomposing the coverage path planning problem into two subproblems: waypoint generation and path planning, the AGA is optimized and solved using a temporal difference-based game algorithm and an AGA, respectively. This improves the efficiency of UAV coverage path planning in complex terrain (e.g., mountainous and hilly areas). This invention effectively addresses the issue of optimizing path length and energy consumption during UAV coverage path planning, significantly improving the efficiency of UAV mission execution in complex terrain. This invention facilitates adaptive generation of near-optimal waypoints and planning of coverage paths, dynamically integrating the optimization objectives of path length and energy consumption, thereby improving the performance and practicality of UAV coverage path planning.

[0067] 2. Existing UAV coverage path planning technologies often use a single algorithm (such as a genetic algorithm or particle swarm algorithm) for path optimization, making it difficult to simultaneously optimize path length and energy consumption, and they are prone to falling into local optimal solutions in complex terrain. The technical solution of the present invention decomposes the coverage path planning problem into two sub-problems: waypoint generation and path planning. These sub-problems are optimized and solved using a temporal difference-based game algorithm and an alternating hierarchical genetic algorithm, respectively. By combining the coupling of these two sub-problems, the efficiency and performance of UAV coverage path planning in complex terrain are significantly improved, while also significantly enhancing the effectiveness of UAV coverage path planning. Furthermore, by combining the advantages of game learning and genetic algorithms, the present invention not only effectively generates waypoints and plans coverage paths, but also further enriches the application of game learning in path planning, filling a technological gap in this field both domestically and internationally.

[0068] 3. The method of the present invention is simple in design and flexible in implementation. It can be widely used in the coverage path planning tasks of UAVs in complex terrains such as mountainous areas and hilly areas, and has high commercial value and market prospects.

[0069] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0070] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0072] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0073] Figure 1 A schematic flow chart of a method for planning a UAV coverage path according to an embodiment of the present invention;

[0074] Figure 2A A schematic diagram of the overlapping area of the coverage scanning field and the corresponding central angle of the drone provided in an embodiment of the present invention;

[0075] Figure 2B A schematic diagram illustrating the relationship between the searchlight radius and the constructed grid point spacing provided in an embodiment of the present invention;

[0076] Figure 3A A schematic diagram of energy information for a drone flying during a two-point turn provided by an embodiment of the present invention;

[0077] Figure 3B A schematic diagram of energy information for a drone flying during a three-point turn provided by an embodiment of the present invention;

[0078] Figure 4 A schematic diagram of cluster points on a valley terrain surface provided by an embodiment of the present invention;

[0079] Figure 5A A schematic diagram of the terrain of four mountains provided in an embodiment of the present invention;

[0080] Figure 5B A topographical diagram of eight peaks provided in an embodiment of the present invention;

[0081] Figure 6A A schematic diagram of a bird's-eye view of the paths ultimately generated by the algorithms on the four mountain peaks provided by an embodiment of the present invention;

[0082] Figure 6B A schematic diagram of a bird's-eye view of the paths ultimately generated by the algorithms on the eight mountain peaks provided by an embodiment of the present invention;

[0083] Figure 7 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0084] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments.

[0085] In describing the present invention, it should be noted that some processes described in this specification and accompanying drawings include multiple operations that appear in a specific order. However, it should be understood that these operations may be performed in a different order than the order in which they appear, or may be performed in parallel. Furthermore, the use of various sequence numbers is for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0086] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort shall fall within the scope of protection of the present invention.

[0087] See also Figure 1 As shown, an embodiment of the present invention provides a method for UAV coverage path planning. First, the coverage path planning problem is decomposed into two sub-problems: waypoint generation and path planning. Specifically, in the waypoint generation stage, the optimal direction for waypoint generation is selected by calculating the number of grid points generated in different directions; a waypoint generation mechanism based on energy consumption optimization is constructed, the weights of the waypoints are quantified, and the waypoint optimization problem is converted into a weighted vertex cover problem; a set of optimized waypoints are derived by solving the weighted vertex cover problem to ensure that the total weight of the coverage waypoints is minimized while the coverage task is completed; in the path planning stage, the optimization problem is solved, and path planning is performed based on the generated waypoints to further optimize the flight path. Specifically, in this method, waypoint generation includes the following steps S1 to S3, and path planning includes the following steps S4 to S5:

[0088] S1, determine the best direction for generating waypoints, that is, select the best direction for generating waypoints by calculating the number of grid points generated in different directions;

[0089] S2, the waypoint optimization problem is modeled as a weighted vertex cover problem, that is, the weight of the waypoint is obtained by taking the average mapping in the way of energy consumption when the waypoint is transferred;

[0090] S3, after converting the problem into a minimum weighted vertex cover problem, a near-optimal solution to the weighted vertex cover problem and an optimized waypoint sequence are obtained through a game algorithm based on temporal differences;

[0091] S4, based on the goal of coverage path planning, i.e., minimizing path length and energy consumption simultaneously, a multi-objective problem is constructed;

[0092] In step S5, an alternating layered genetic algorithm is used to solve the multi-objective optimization problem, and a near-optimal coverage path is obtained based on the waypoints generated in step S3.

[0093] The method of the present invention can significantly improve path planning efficiency and greatly reduce energy consumption, improve the ability to solve complex coverage path planning problems, and provide an efficient solution for energy-sensitive tasks.

[0094] 1. The specific implementation of the method of the present invention is described in detail below with reference to the accompanying drawings:

[0095] In the embodiment of the present invention, it is necessary to prioritize the setting of sensing constraints (1) (2) and assumptions (3)-(5) in order to analyze and apply them according to the actual situation:

[0096] (1) Equidistant: In order to keep the photos taken by the photographic sensor in the same proportion and prevent the size of the picture from being inconsistent, a normal vector perpendicular to the surface U is set for each point on the surface G, and a constant distance is maintained along the normal vector, that is, the three-dimensional terrain surface (smooth surface) A is obtained;

[0097] (2) Frontal: The drone must point its photographic sensor and searchlight (or radar sensor) towards the terrain surface to be covered and maintain a vertical position with the terrain surface at all times;

[0098] (3) The three-dimensional terrain surface can be perfectly fitted, that is, B-spline surface fitting;

[0099] (4) The photographic sensor and searchlight (or additional infrared scanner) are fixed on a gimbal stabilizer, that is, the camera itself does not adjust the angle, and is used to keep the camera's yaw angle consistent with the drone;

[0100] (5) The radius R of the searchlight needs to be less than half the width of the photograph to ensure that the full field of view within the searchlight is captured.

[0101] Furthermore, in step S1, the embodiment of the present invention first selects the optimal direction for waypoint generation by calculating the number of grid points generated in different directions. Specifically, all edges are sorted in descending order of length. Then, each edge is traversed and its maximum distance from all vertices is calculated. The number of vertical and horizontal dividing lines is determined based on this distance and the length of the edge. Next, the number of waypoints for each edge is calculated, and the minimum number of waypoints is updated during the traversal process. Ultimately, the optimal number of waypoints generated is determined to obtain the optimal direction for waypoint generation.

[0102] The calculation of the number of different grid points in step S1 of the present embodiment is for convex polygonal surfaces. This is because concave polygons can be divided into multiple convex polygons for calculation. In practice, the region of interest may be concave or convex. However, since concave polygons can be solved by decomposing them into convex polygons, when encountered in practice, it is only necessary to prioritize decomposing small convex regions of interest and repeat the steps several times.

[0103] Furthermore, in step S2 provided by the embodiment of the present invention, the waypoint generation sub-problem is transformed into a weighted vertex cover problem, the core of which is to use the energy consumption when the waypoint is transferred to map the weight information of the waypoint. Figure 2A and Figure 2B The relationship between the radius, grid spacing and other values is shown. Step S2 includes the following steps:

[0104] S21, calculate the slope angle between point a and point b :

[0105] like Figure 3A As shown, let the distance between points a and b be d ab , the absolute height difference is h ab Then the slope angle It can be expressed as: .

[0106] like Figure 3B As shown, the turning angle of point a through point b to point c can be calculated. If the coordinates of a, b and c are , , , and the distance between points a and b is , the distance between points b and c is , the distance between points a and c is Then according to the law of cosines, the turning angle is It can be expressed as: .

[0107] S22, get the specific quantification of the energy from point a to point b:

[0108] Generally, the specific energy value from point a to point b is different from the specific energy value from point b to point a. Here, the energy value from point a to point b is calculated as: ,in It represents the coefficient measured by linear fitting. At the same time, the energy consumption caused by the turning angle from point a to point c through point b can also be calculated as: .

[0109] S23, calculate the energy consumption from neighboring waypoint j to i:

[0110] The grid points are defined in order as , denote i’s neighbor waypoint as j, and the energy consumption of each j to i is based on the path length and slope angle To express it, it can be written as:

[0111]

[0112] in, The value of can be measured according to the actual situation. In the embodiment here, the value of the quad-rotor drone is About 0.1164kJ / m, The value is about 0.0173kJ / deg.

[0113] S24, take the average energy consumption of all neighboring waypoints j to i, and map this value to the weight of waypoint i:

[0114] In this embodiment, the average energy consumption of all neighbors j to reach i is calculated and mapped to the weight of i. Therefore, the energy consumption is mapped to the weight information of each waypoint.

[0115] In step S3 provided by the embodiment of the present invention, since players have stronger or weaker desires to go home, they exhibit significantly different action trends and benefits during the game. Therefore, a payoff matrix for the asymmetric snowdrift game can be constructed. This game can then be combined with the temporal difference method to develop a new algorithm for solving the weighted vertex cover problem. The main contents of the algorithm include:

[0116] S31, construct an asymmetric snowdrift game model:

[0117] Based on the most primitive game model—the snowdrift game and the asymmetric game, the present invention can define the asymmetric snowdrift game as follows:

[0118] Because players with a strong desire to go home have a stronger tendency to go home, they will gain more benefits. Such players are called strong players, represented by SP. Conversely, weak players are represented by WP, and their benefits are relatively low. The factor that controls benefits is called an asymmetric factor. The benefit matrix in this case is as follows:

[0119]

[0120] Among them, C represents cooperation, D represents betrayal, Indicates that strong players choose to cooperate, Indicates that the strong player chooses to betray, indicates that the weak player chooses to cooperate, Indicates that the weak player chooses to defect; r indicates the profit and loss ratio, represents the asymmetric factor.

[0121] S32, combine TD(λ) in the temporal difference method with the asymmetric snowdrift game to construct an algorithm TD(λ)G for solving the weighted vertex cover problem:

[0122] At this point, by combining the self-iteration of TD(λ) and the calculation of the game's benefits, the algorithm can be constructed as follows:

[0123] (1) If , then:

[0124] and ;

[0125] Where, represents the payoff when the player chooses to cooperate at time t, represents the payoff when the player chooses to cooperate at time t+1, represents the benefit of waypoint i at time t, represents the profit transition variable at time t+1 when the player chooses to cooperate, is the learning rate, represents the payoff when the player chooses to cooperate at the waypoint i at time k+1, C represents cooperation, and D represents betrayal. and denote the probability of cooperation and defection, represents the difference in returns between time k+1 and time k, γ represents the discount factor, represents the benefit of waypoint i at time k.

[0126] (2) If , then:

[0127] and .

[0128] Where, represents the payoff when the player chooses to betray at time t, represents the payoff transition variable at time t+1 when the player chooses to betray, represents the payoff when the player chooses to betray at the waypoint i at time k+1, represents the benefit of waypoint i when it chooses to defect at time k+1, It represents the difference in returns between time t+1 and time t, that is, , and represent the probability of cooperation and defection, , , and γ is the discount factor, set to 0.95.

[0129] Furthermore, the specific steps of the TD(λ)G algorithm can be expressed as follows:

[0130] (1) Initialization: All players are randomly assigned action strategies A = {C, D}, where C represents cooperation and D represents defection;

[0131] (2) Find the waypoints with the largest degree in the graph and sort them, calculate the benefits obtained after i and all neighbors play the game, and use weighted sum to calculate and ;

[0132] (3) Each player makes decisions based on the decision-making rule (DMR);

[0133] (4) If , then:

[0134] Otherwise, If the strategies of all players do not change, the iteration stops, otherwise it returns to step (2) until the iteration stops;

[0135] (5) Modify the strategy at the end of the iteration. When the strategy of i is C, if the neighbor set of i If the strategy of all waypoints in is C, then the strategy of i is modified to D; at the same time, if the strategy of i is C, There is only one point where j’s strategy is D and when the weight of i is greater than the weight of j, the strategies of i and j are changed.

[0136] The specific representation of DMR is as follows:

[0137] (1) If , the player will adopt the cooperative C strategy at the next moment, and the probability of cooperation increases, that is, ; At the same time, the probability of betrayal decreases, that is .

[0138] (2) If , the player will adopt the betrayal D strategy at the next moment, and the probability of cooperation decreases, that is, ; At the same time, the probability of betrayal increases, that is .

[0139] (3) If , the player will adopt the betrayal D strategy at the next moment, and the cooperation probability remains unchanged, while the betrayal probability increases, that is, and .function is the modified Fermi rule, expressed as follows:

[0140]

[0141] in represent or , k is the Boltzmann constant times, set to .

[0142] In this embodiment, during the model initialization phase, the initial probability distribution and discount factor values need to be appropriately set based on the actual waypoint characteristics and mission requirements. This refined parameter setting not only accelerates the model's convergence speed but also significantly improves the model's prediction accuracy and generalization performance, resulting in better experimental results.

[0143] The multi-objective optimization problem in step S4 provided in the embodiment of the present invention can be decomposed into two main objectives, namely, minimizing the total path length and minimizing energy consumption. The main objective problem formula is as follows:

[0144] ① Minimize path length goal:

[0145]

[0146]

[0147] Where L represents the path length; represents the length of the path that the UAV travels from i to j; is the total number of waypoints. If the UAV travels from i to j, then ;otherwise, .

[0148] ②Minimize energy consumption goals:

[0149]

[0150]

[0151] Where E represents the energy; Indicates the turning angle from point i to point k via point j; β is approximately 0.1164 kJ / m, and δ is approximately 0.0173 kJ / deg; Represents the distance between waypoint i and waypoint j along the waypoint fitting surface; represents the path length from waypoint j to k, such as It can be expressed as the path length from waypoint 1 to waypoint 2; similarly, Indicates from waypoint To waypoint Path length; Represents the decision variable. If the drone flies from i to j and then to k, then ;otherwise, .

[0152] In step S5 provided by the embodiment of the present invention, an alternating layered genetic algorithm is used to solve the multi-objective optimization problem, and an optimal coverage path is obtained based on the waypoints generated in step S3. The specific process includes:

[0153] S51, clustering valley lowland clusters:

[0154] like Figure 4 As shown in the figure, when the UAV passes through the valley lowlands, there may be a relative overlap (i.e., very close distance) of the aerial waypoints due to the existence of sensing constraints. Therefore, in this module called clustered waypoints, the cluster points can be clustered, and the center point is taken as the coverage point, while the other points in the cluster are uncovered points.

[0155] Using a clustering algorithm to identify cluster points for drone path planning offers significant advantages. First, by clustering nearby target points, the number of path waypoints required for calculation is significantly reduced, effectively conserving computing resources and improving path planning efficiency. Second, using cluster centers as required waypoints optimizes flight paths and avoids frequent, narrow-angle turns. This not only reduces energy consumption but also improves flight stability.

[0156] S52, using alternating hierarchical genetic algorithm (AHG) to solve multi-objective optimization problems:

[0157] Considering the complexity of waypoint selection in UAV path planning and the limited computing resources, this paper designs a waypoint optimization selection method based on genetic algorithm. The specific process is as follows:

[0158] (1) Initialize the waypoint sequence population AirW;

[0159] (2) In the first layer of operations, the path length is considered as the main factor to calculate the fitness value. The tournament selection, sequential crossover, and simple mutation are run 5000 times to obtain a waypoint sequence with the minimum relative path length, and this sequence is put into the population AirW;

[0160] (3) In the second layer of operation, energy consumption is considered as the main factor to calculate the fitness value. The tournament selection, sequential crossover, and simple mutation are run 3000 times to obtain a waypoint sequence with the minimum relative path length, and the sequence is put into the population AirW;

[0161] (4) Return to step (2) and perform the calculation again to obtain the waypoint sequence that minimizes both the final path length and energy consumption, and use this to calculate the values of the path length and energy consumption.

[0162] 2. Experimental evaluation:

[0163] like Figure 5A and Figure 5B To demonstrate the performance of the method presented herein, simulation experiments were conducted on a 100m×100m terrain with four peaks and a 200m×100m terrain with eight peaks, respectively. The drone's flight altitude was set to 10m. The code for the alternating hierarchical genetic algorithm based on game learning was compiled in Python 3.12.2 using Visual Studio Code 1.97.2. The simulations were conducted on an Inspiron 3568 laptop with an Intel(R) Core(TM) i7-7500U CPU @ 2.70GHz. The computer used the alternating hierarchical genetic algorithm based on game learning. The present invention was evaluated in coverage path planning experiments conducted by drones in scenarios with four and eight peaks. To demonstrate the effectiveness of the TD(λ)G-AHG algorithm (the present method), Table 1 first compares the characteristics of the BF variant, Spiral variant, HO-CTP variant, and AAPMS variant with the TD(λ)G-AHG algorithm.

[0164] Table 1 Comparison of method characteristics

[0165]

[0166] In Table 1, it can be noted that the BF-variant, Spiral-variant, HO-CTP-variant, and AAPMS-variant simply consider the complete coverage of the mission area and ignore the impact of waypoint generation on path planning. However, the TD(λ)G-AHG algorithm not only considers the complete coverage of the mission area, but also realizes the coupling between the two and considers the impact of waypoint generation on path planning.

[0167] The present invention conducted simulation experiments on terrains with four peaks and eight peaks, and a direct comparison of performance was made in Table 2. Figure 6A and Figure 6B The flight paths are shown in Figure 1 (part (a) is the BF variant; part (b) is the Spiral variant; part (c) is the HO-CTP variant; part (d) is the APPMS variant; and part (e) is the TD(λ)G-AHG variant). Performance comparison is reflected in two indicators: and , What is the performance indicator of path length? Indicates an indicator to measure energy consumption performance. represents the path length of the BF-variant, Indicates the path length of the remaining methods; Indicates the energy consumption of BF-variant, Indicates the energy consumption of the remaining methods.

[0168] Table 2 Performance evaluation

[0169]

[0170] ①. Simulation on four peaks:

[0171] Table 2 shows that the TD(λ)G-AHG algorithm performs well in terms of both path length and energy consumption. This algorithm achieves the shortest path length, improving upon the BF-variant algorithm while simultaneously reducing energy consumption. In contrast, the other three algorithms require longer paths and consume more energy to complete their coverage tasks. While the APPMS-variant algorithm achieves a relatively balanced balance between path length and energy consumption, the TD(λ)G-AHG algorithm outperforms the APPMS-variant in both aspects, resulting in better overall performance. This result demonstrates that the TD(λ)G-AHG algorithm not only significantly shortens flight path length while saving energy, but also effectively reduces flight time and mitigates the risk of encountering uncertainties during flight. This demonstrates that, compared to other algorithms, the TD(λ)G-AHG algorithm, which considers the coupling of waypoint generation and path planning, has greater practical application value in UAV coverage path planning.

[0172] ②. Simulation on eight peaks:

[0173] As shown in Table 2, while the AAPMS-variant algorithm achieves relative minimization of path length and energy consumption, outperforming the BF-variant algorithm in both respects, the proposed TD(λ)G-AHG algorithm demonstrates significant improvement. After simplifying the cluster points, the TD(λ)G-AHG algorithm achieves superior performance, significantly outperforming the other algorithms by reducing both path length and energy consumption compared to the BF-variant algorithm. This has significant practical value in the real world, allowing drones to conserve energy for search, patrol, and other coverage missions while also storing excess energy for emergency situations.

[0174] In summary, the proposed method excels in optimizing path length and energy consumption. From a practical application perspective, the TD(λ)G-AHG algorithm employed in this paper significantly improves the efficiency and reliability of drones for search, patrol, and other coverage tasks. This demonstrates the effectiveness of TD(λ)G-AHG-based drone coverage path planning in terms of practicality and performance optimization.

[0175] From the description of the above embodiments, those skilled in the art will know that: the present invention provides a method for UAV coverage path planning based on an alternating hierarchical genetic algorithm using game learning, which selects the optimal direction for generating waypoints by calculating the number of grid points generated in different directions; constructs a waypoint generation mechanism based on energy consumption optimization, quantifies the weights of waypoints, and transforms the waypoint optimization problem into a weighted vertex cover problem; derives a set of optimized waypoints by solving the weighted vertex cover problem, ensuring that the total weight of the covered waypoints is minimized while the coverage task is completed; solves the optimization problem, performs path planning based on the generated waypoints, and further optimizes the flight path. The present invention can significantly improve path planning efficiency and significantly reduce energy consumption, improve the ability to solve complex coverage path planning problems, provide an efficient solution for energy-sensitive tasks, and further expand the potential of path planning algorithms in practical applications.

[0176] Furthermore, an embodiment of the present invention also provides a drone coverage path planning system, which is applied to the drone coverage path planning method described in the above embodiment to perform drone coverage path planning. The system mainly includes:

[0177] The waypoint generation module is used to select the optimal direction for waypoint generation by calculating the number of grid points generated in different directions. The module also uses the average energy consumption during waypoint transfer to obtain the waypoint weights, transforming the waypoint optimization into a weighted vertex cover problem. A near-optimal solution to the weighted vertex cover problem is solved using a temporal difference-based game algorithm to obtain an optimized waypoint sequence.

[0178] The path planning module is used to: construct the optimization objectives of the UAV coverage path planning, the optimization objectives include: minimizing the path length and minimizing the energy consumption; and according to the optimization objectives of the UAV coverage path planning, use the alternating layered genetic algorithm based on the optimized waypoint sequence to solve and obtain the optimal coverage path.

[0179] In a specific embodiment, preferably, the system also includes a sensor constraint module for ensuring that the UAV maintains a perpendicular state to the terrain surface during the coverage path planning process according to the set sensor constraints and assumptions, and meets the coverage requirements of the photographic sensor and the searchlight.

[0180] An embodiment of the present invention provides a drone coverage path planning system, the implementation principle and technical effects of which are the same as those of the aforementioned method embodiment. For the sake of brief description, for parts not mentioned in this embodiment, reference can be made to the corresponding contents in the aforementioned method embodiment, which will not be repeated here.

[0181] Further, refer to Figure 7 As shown, an embodiment of the present invention also provides an electronic device that can execute the above-mentioned method and system. The electronic device may include a processor 10, a memory 11, a communication bus 12 and a communication interface 13, and may also include a computer program stored in the memory 11 and capable of running on the processor 10.

[0182] In some embodiments, the processor 10 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and a combination of various control chips. The processor 10 is the control core (Control Unit) of the electronic device, which uses various interfaces and lines to connect the various components of the entire electronic device, and executes or executes programs or modules stored in the memory 11, and calls the data stored in the memory 11 to perform various functions of the electronic device and process data.

[0183] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, devices, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0184] It should be noted that the word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention can be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer.

[0185] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0186] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art who can easily conceive of changes or substitutions within the technical scope disclosed in the present invention should be covered by the scope of protection of the present invention.

Claims

1. A method for planning a UAV coverage path, characterized in that: In this method, the UAV coverage path planning is decomposed into: waypoint generation and path planning; wherein, waypoint generation includes the following steps 1 to 3, and path planning includes the following steps 4 to 5: Step 1: Select the best direction for generating waypoints by calculating the number of grid points generated in different directions; Step 2: Use the energy consumption during waypoint transfer to get the average mapping to obtain the waypoint weight, and convert the waypoint optimization into weighted vertex cover; Step 3: Solve the near-optimal solution of weighted vertex cover through a game algorithm based on temporal difference to obtain the optimized waypoint sequence; Step 4: Construct the optimization objectives for the UAV coverage path planning, which include minimizing the path length and minimizing the energy consumption; Step 5: According to the optimization goal of UAV coverage path planning, the alternating layered genetic algorithm is used to solve the optimized waypoint sequence to obtain the optimal coverage path.

2. The method for planning a UAV coverage path according to claim 1, characterized in that: In this method, sensing constraints and assumptions are first set to analyze and apply to actual situations, where: Constraints include: Equidistant: Set a normal vector perpendicular to the surface for each point on the surface, and keep a constant distance along the normal vector; Front: Point the drone's sensor device toward the terrain surface to be covered, and keep it perpendicular to the terrain surface at all times; Assumptions include: The yaw angle of the sensing device is consistent with that of the drone; The radius of the searchlight is less than half the width of the photographic picture.

3. The method for planning a UAV coverage path according to claim 2, wherein: In step 1, for a convex polygonal surface, the optimal direction for generating waypoints is selected by calculating the number of grid points generated in different directions. The specific process includes: Arrange all edges in descending order of length, then traverse each edge and calculate its maximum distance to all vertices, and determine the number of vertical and horizontal dividing lines based on this distance and the length of the edge; Calculate the number of waypoints for each edge and update the minimum number of waypoints during traversal; The direction that can generate the least number of grid points is determined as the optimal direction for generating waypoints.

4. The method for planning a UAV coverage path according to claim 3, wherein: In the second step, the energy consumption during the waypoint transfer is averaged and mapped to obtain the weight of the waypoint, and the waypoint optimization is converted into a weighted vertex cover, which specifically includes: (1) For points a and b on the terrain surface, calculate the slope angle between points a and b. The calculation formula is: ； Where, represents the slope angle between point a and point b, represents the absolute height difference between point a and point b, Represents the distance between point a and point b; (2) Calculate the energy from point a to point b using the following formula: ； Where, Represents the energy value from point a to point b, represents the coefficients measured by linear fitting; (3) Let the neighboring waypoint of waypoint i be j, and calculate the energy consumption from neighboring waypoint j to i. The calculation formula is: ； Where, represents the energy consumption of neighbor waypoint j to i, represents the coefficients measured by linear fitting, represents the path length from neighbor waypoint i to j, represents the slope angle from neighbor waypoint j to i; (4) Take the average energy consumption of all neighboring waypoints j to i, and map this value to the weight of waypoint i.

5. The method for planning a UAV coverage path according to claim 4, characterized in that: In step 3, the game algorithm based on temporal difference is an algorithm constructed by combining the temporal difference method with the asymmetric snowdrift game. The algorithm is as follows: 1) If , then: ； ； Where, represents the payoff when the player chooses to cooperate at time t, represents the payoff when the player chooses to cooperate at time t+1, represents the benefit of waypoint i at time t, represents the profit transition variable at time t+1 when the player chooses to cooperate, is the learning rate, represents the payoff when the player chooses to cooperate at the waypoint i at time k+1, C represents cooperation, and D represents betrayal. and denote the probability of cooperation and defection, represents the difference in returns between time k+1 and time k, γ represents the discount factor, represents the benefit of waypoint i at time k; 2) If , then: ； ； Where, represents the payoff when the player chooses to betray at time t, represents the payoff transition variable at time t+1 when the player chooses to betray, represents the payoff when the player chooses to betray at time k+1 at waypoint i, represents the benefit of waypoint i when it chooses to defect at time k+1, It represents the difference in returns between time t+1 and time t.

6. The method for planning a UAV coverage path according to claim 4, characterized in that: In step 4, the expression for minimizing the path length is: ； ； Where L represents the path length; represents the path length of the UAV from waypoint i to j; is the total number of waypoints. If the UAV travels from i to j, then ; otherwise, ; i and j represent waypoints; The expression for minimizing energy consumption is: ； ； Where, E represents the energy consumption; Indicates the turning angle from point i to point k through point j; represents the path length from waypoint j to k, Indicates the path length from waypoint 1 to waypoint 2; similarly, Indicates from waypoint To waypoint Path length; Represents the decision variable. If the drone flies from i to j and then to k, then ; otherwise, .

7. The method for planning a UAV coverage path according to claim 1, wherein: In step 5, an alternating layered genetic algorithm is used to solve the problem based on the optimized waypoint sequence to obtain the optimal coverage path. The process includes: (1) Initialize the waypoint sequence population; (2) In the first layer of operations, the fitness value is calculated using the path length. The tournament selection, sequential crossover, and simple mutation are run multiple times to obtain a waypoint sequence with the minimum relative path length, and this sequence is put into the population. (3) In the second layer of operations, the fitness value is calculated using energy consumption, and the tournament selection, sequential crossover, and simple mutation are run multiple times to obtain a waypoint sequence with the minimum relative path length, which is then put into the population; (4) Return to step (2) and perform the calculation again to obtain the waypoint sequence with the minimum path length and energy consumption, and obtain the optimal coverage path.

8. A UAV coverage path planning system, characterized in that: When applied, the method for planning a UAV coverage path according to any one of claims 1 to 7 is executed to plan a UAV coverage path, and the system includes: The waypoint generation module is used to select the optimal direction for waypoint generation by calculating the number of grid points generated in different directions. The module also uses the average energy consumption during waypoint transfer to obtain the waypoint weights, transforming the waypoint optimization into a weighted vertex cover problem. A near-optimal solution to the weighted vertex cover problem is solved using a temporal difference-based game algorithm to obtain an optimized waypoint sequence. The path planning module is used to construct the optimization objectives of the UAV coverage path planning. The optimization objectives include minimizing the path length and minimizing the energy consumption. According to the optimization objectives of the UAV coverage path planning, an alternating layered genetic algorithm is used to solve the problem based on the optimized waypoint sequence to obtain the optimal coverage path.

9. An electronic device, characterized in that: It includes a processor and a memory, the memory stores machine-executable instructions that can be executed by the processor, and the processor executes the machine-executable instructions to implement a drone coverage path planning method as described in any one of claims 1 to 7.

Images