A multi-UAV inspection path planning method considering wind influence and heterogeneous takeoff and landing

Abstract

This invention discloses a multi-UAV inspection route planning method that considers wind impact and heterogeneous takeoffs and landings. The method comprises the following steps: obtaining geographic information data, the weight of the inspection UAVs, the battery energy data of the inspection UAVs, and wind direction and speed forecast data for the inspection period; setting the horizontal airspeed and the ascent and descent flight speeds of the UAVs during the inspection; establishing an estimation matrix for the UAVs' direct flight energy consumption under wind impact; and solving a three-stage model for multi-UAV inspection route planning that considers wind impact and heterogeneous takeoffs and landings. This method models and measures the impact of wind on UAV flight energy consumption, taking into account the heterogeneous takeoff and landing patterns of UAVs. This method can fully utilize the advantages of multi-UAV applications in forest inspections, reduce forest fire prevention inspection costs, and minimize human and material resource consumption.

Description

Technical Field

[0001] The present invention belongs to the technical field of UAV path planning, and in particular relates to a multi-UAV inspection path planning method that takes into account wind influence and heterogeneous take-off and landing. Background Art

[0002] Compared to traditional productivity tools, drones offer the advantage of being able to move freely in low-altitude, three-dimensional space. With the continuous improvement of drone automation and reliability, civilian drones have found widespread application in aerial photography, inspections, agricultural and forestry plant protection, meteorological monitoring, and emergency rescue, leading to explosive growth in the drone industry.

[0003] Unmanned aerial vehicle (UAV) path planning problems are primarily categorized as point traversal and coverage. The primary difference between these two problems is that point traversal requires the UAV's path to pass through all target points, while coverage requires the UAV's trajectory to cover the entire mission area. Currently, scholars both domestically and internationally have achieved significant success in addressing point traversal, employing a variety of approaches, including pure mathematical optimization algorithms such as mixed integer programming and dynamic programming; graph-based planning methods such as Voronoi diagrams and fast random number search; heuristic optimization algorithms such as the A* algorithm; and swarm intelligence optimization algorithms such as the ant colony algorithm, genetic algorithm, and gray wolf algorithm. Numerous improvements have been made to these original algorithms, making them suitable for UAV path planning. The coverage path planning problem, primarily derived from robot vacuums and ground coverage search, primarily employs subspanning trees or pre-partitioning the mission area to balance the workload among multiple UAVs. Trajectories are then directly generated using coverage trajectory planning templates such as parallel reciprocating, spiral, and extended square paths. For drone forest fire inspection scenarios with high elevation fluctuations, these two-dimensional path planning methods lack precision in concave terrain and present a greater risk of collision in convex terrain. Therefore, this scenario requires pre-planning three-dimensional waypoints that visually cover the entire inspection area and addressing the traversal of the large number of planned aerial waypoints. Inspection tasks are routine and long-term in nature, so the cost of each inspection should be minimized when planning routes. Considering the impact of wind can help save energy through wind power, and turning, a major energy consumption factor for drones, should also be considered in the model. For large-scale mountain inspections where a single drone cannot complete them due to endurance limitations, multi-drone path planning is necessary, and the fairness of task allocation must also be considered. Existing research has significant shortcomings in this area and requires further improvement. Summary of the Invention

[0004] To address the shortcomings of the aforementioned prior art, the present invention aims to provide a multi-UAV inspection path planning method that considers wind impacts and heterogeneous takeoff and landing patterns. This method addresses the prior art's inadequate consideration of environmental wind impacts, poor flexibility in same-nest takeoff and landing, and high UAV flight energy consumption. Forest inspections cover a wide area and experience significant altitude fluctuations, necessitating three-dimensional path planning for multiple UAVs. This method models and measures the impact of wind on UAV flight energy consumption, taking into account heterogeneous takeoff and landing patterns. This method can fully leverage the advantages of multiple UAVs in forest inspections, reduce forest fire prevention inspection costs, and minimize both human and material resources.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] The present invention provides a multi-UAV inspection path planning method that considers wind influence and heterogeneous takeoff and landing, and the steps are as follows:

[0007] 1) Data acquisition: Obtain geographic information data, the weight of the inspection drone, the battery energy data of the inspection drone, and the wind direction and speed forecast data during the inspection period;

[0008] 2) Set parameters: Set the horizontal airspeed and the ascent and descent speeds of the drone during inspection;

[0009] 3) Establish an energy consumption estimation matrix for direct flights of drones under wind influences: Develop equations for energy consumption per meter of flight for drones in different states, both in a windless environment and under wind influences; determine the flight pattern of the drone between the starting and ending points of each flight segment; calculate the energy consumption of direct flights between any two points within the set of aerial visual coverage inspection waypoints, and establish an energy consumption estimation matrix for direct flights of drones under wind influences;

[0010] 4) Solve the three-stage model for multi-UAV inspection path planning considering wind influence and heterogeneous takeoff and landing: in the first stage, the TSP problem is established to solve the Hamiltonian circuit that minimizes energy consumption under the influence of wind; in the second stage, the path optimization is performed while minimizing the total turning angle; in the third stage, multi-UAV task allocation is performed considering the operation mode of heterogeneous takeoff and landing, and the optimal solution for the take-off and landing positions of UAVs in each nest is calculated with the goal of task balance, so as to obtain the multi-UAV inspection path that takes into account the direct flight energy consumption, turning angle and task allocation fairness under the influence of the set wind.

[0011] Furthermore, the geographic location information data in step 1) includes the latitude and longitude range of the inspection area, digital elevation, three-dimensional coordinates of the machine nest and aerial visual coverage inspection waypoints.

[0012] Furthermore, the step 2) specifically includes: setting the horizontal airspeed v of the drone during inspection a For lab test speed v test , the value is 10m / s, the vertical flight speed vv It is divided into rising speed and falling speed. When rising, v v The value is positive, when v decreases v If the value is negative, the ascending speed is set to a certain value (m / s) in the interval [1,4], and the descending speed is set to a certain value (m / s) in the interval [-4,-1].

[0013] Furthermore, the step 3) specifically includes:

[0014] 31) Establish the energy consumption equation per meter of flight for UAVs under different conditions in a windless environment;

[0015] The energy consumption per meter of flight in the horizontal direction is calculated using the following formula:

[0016]

[0017] Where Epm represents the energy consumption per meter of flight in the horizontal direction in a windless environment, E M is the battery energy of the drone, v test The laboratory test speed is 10 m / s. During the test, the drone flies at this speed until the energy is exhausted. M With speed v test The maximum flight time of the drone measured in horizontal flight; set the flight speed v of the drone a =v test ;

[0018] The energy consumption per meter of flight in the vertical direction is calculated as follows:

[0019]

[0020] Where, Epm v Indicates the energy consumption per meter of the UAV in vertical flight, P v Indicates the flight power of the UAV in the vertical direction, v v Indicates the vertical flight speed of the drone, v when rising v The value is positive, when v decreases v The value is negative;

[0021] 32) Establish the energy consumption equation per meter of a UAV flying horizontally under the influence of wind;

[0022] The two attributes of the wind vector w are recorded as wind speed v w and wind direction w ; Under a given wind vector w, the ground speed of the drone is calculated as follows:

[0023] v g =v a +v w cos(d a -dw ) (3)

[0024] Where, v g is the ground speed of the UAV, v a is the flight speed of the UAV, d a is the flight direction, d w For wind direction;

[0025] Considering the influence of the wind and headwind components on energy consumption, the energy consumption per meter of horizontal flight of the UAV under wind vector w is Epm w It can be expressed as:

[0026]

[0027] 33) Determine the flight pattern of the UAV between the starting point and the end point of each flight segment;

[0028] When the drone performs an inspection mission, the flight types between the starting point and the end point of each flight include flight between the machine nest and the aerial visual coverage waypoint and flight between two waypoints. For the flight between the machine nest and the waypoint, the vertical movement distance is long. When taking off, the drone is set to first rise vertically to the height of the target waypoint, and then fly horizontally towards the target waypoint. When landing, it first flies from the waypoint to the air above the target machine nest, and then lands vertically. For the flight between two waypoints, the drone is set to move simultaneously at specified speeds in both the horizontal and vertical directions. If the specified position in one direction is reached first, the subsequent speed in that direction is reset to 0.

[0029] 34) Establish an energy consumption estimation matrix for direct UAV flight under wind influence;

[0030] For two aerial visual coverage waypoints i and j in the waypoint set P, let their coordinates be (x i ,y i ,z i ),(x j ,y j ,z j ), calculate the flight direction d of the drone according to the coordinates a , and calculate Epm according to step 31) v , calculate Epm according to step 32) w , then the energy consumption of the UAV flying from point i to point j can be calculated by formula (5):

[0031]

[0032] Where, Represents the energy consumption of the drone flying from point i to point j under wind vector w; calculates the energy consumption of forward and reverse flights between any two aerial visual coverage waypoints under wind vector w, and thus establishes the direct flight energy consumption estimation matrix E between two aerial visual coverage waypoints under wind vector w w , matrix E w The element in the i-th row and j-th column is the energy consumption of the drone flying from point i to point j under the wind vector w.

[0033] Furthermore, the calculation of the vertical flight power of the UAV in step 31) is as follows:

[0034]

[0035] k1∈[0,1] (7)

[0036]

[0037]

[0038] Where, P v It represents the flight power of the UAV in the vertical direction, m is the weight of the UAV, g is the acceleration of gravity, N is the total number of blades of a single propeller, c is the d is the drag coefficient of the UAV blade, c is the chord width of the UAV blade, ρ is the air density, and R is the length of the UAV blade (the radius of the circle formed during rotation); k1 is a constant, k2 is the deviation factor between the ideal uniform airflow and the actual airflow, k3 is the proportional factor that converts the rotor angular velocity into thrust, and c2 is the intermediate coefficient.

[0039] Furthermore, the step 4) specifically includes:

[0040] 41) Stage 1: Solving the Hamiltonian circuit for minimizing energy consumption under the influence of wind;

[0041] The cost matrix is ​​established using the direct flight energy consumption between each aerial visual coverage waypoint under the influence of wind vector w as an element, and a TSP problem model is established to solve the Hamiltonian circuit with the minimum cost; the details are as follows:

[0042]

[0043]

[0044]

[0045]

[0046]

[0047] Where E represents the total energy consumption of the UAV that traverses the Hamiltonian circuit of all aerial visual coverage waypoints; represents the energy consumption of the UAV flying from point i to point j under wind vector w; X ij is a decision variable. When it is 1, it means the UAV flies from point i to point j. When it is 0, it means the UAV does not fly from point i to point j. i 、u j represents the order in which the UAV visits points i and j, and P is the set of waypoints;

[0048] The branch and bound method is used to solve the model of equations (10)-(14). The branch and bound method initializes the problem and creates a root node, then continuously expands the nodes and uses the bound function to evaluate the possibility of the optimal solution of each subproblem until the optimal solution is found or no better solution can be found.

[0049] 42) Stage 2: Path optimization with the goal of minimizing the total turning angle;

[0050] Further considering the turning minimization objective, a multi-objective TSP model is established; using A ijq Indicates that the continuous path points are p i ,p j ,p q When the drone is in p j The turning angle at is calculated as follows:

[0051] A ijq =180°-∠p i p j p q (15)

[0052] The objective function of minimizing the turning angle is established as follows:

[0053]

[0054] Where A is the total turning angle of the inspection loop. The multi-objective TSP model considering the minimization of total energy consumption and total turning amount is established by equations (10) to (16). The Hamiltonian loop obtained in stage 1 is used as the initial solution to input into the NSGA-II algorithm. A solution is selected from the Pareto front solution in which the turning amount is reduced while the direct flight energy consumption does not increase. The inspection loop obtained in stage 2 is set as the waypoint sequence {x0, x1, x2, ..., x |P| ,x0};

[0055] 43) Phase 3: Multi-UAV task allocation considering heterogeneous take-off and landing operation modes;

[0056] Insert |V| nests into the inspection loop as the landing point of the preceding UAV and the take-off point of the following UAV, so as to segment the inspection path and assign it to each UAV. V is the nest set. For a nest v, its candidate insertion position set is recorded as Candi(v). Define the n nearest two-dimensional distances around it. Near The aerial visual coverage waypoints are the neighboring waypoints of the aircraft nest, denoted as Near(v); if there are two consecutive waypoints x in the path obtained in stage 2 i 、x i+1 are all in Near(v), then x i Add Candi(v) as the candidate insertion position of the nest v; the details are as follows:

[0057]

[0058] Decision variable X vi The details are as follows:

[0059]

[0060] After the nest is inserted, the inspection loop is divided into |V| segments. The starting and ending points of each segment are both nests. The segment path is inspected by the drone in the nest at the starting point of the segment. Suppose the flight path S of the drone taking off from nest v is v for Then the energy consumption E(v) of the UAV is:

[0061]

[0062] Where, Indicates that the drone is from point Flight to point The energy consumed is the energy consumed by the drone taking off from the nest v to fly from the i-th point to the i+1-th point on the corresponding inspection path; a minimum-maximum objective function is established to find an inspection task allocation scheme that minimizes the maximum inspection energy consumption of all drones and ensures that all drones have a relatively fair workload; the specific objective function is as follows:

[0063]

[0064] The constraints of the stage three model are as follows:

[0065]

[0066]

[0067] Where, E(v) represents the inspection energy consumption of the UAV taking off from the nest v, E M represents the battery energy of the drone, λ Eis the proportional coefficient; after the nest is inserted into the flight path of the UAV, it becomes the landing nest of the preceding UAV and the take-off nest of the following UAV. Formula (21) indicates that each nest can only be inserted once; Formula (22) indicates that the flight energy consumption of each UAV shall not exceed the maximum energy consumption set for normal inspection;

[0068] By solving the mathematical model composed of equations (17)-(22), we can obtain the multi-UAV inspection path that takes into account the direct flight energy consumption, turning angle and task allocation fairness under the influence of the set wind.

[0069] Beneficial effects of the present invention:

[0070] Aiming at the UAV forest inspection scenario, the present invention models and measures the impact of wind on the UAV flight energy consumption, proposes a multi-UAV inspection path planning model considering the UAV's heterogeneous take-off and landing modes, and decomposes the problem into three sub-problems for solution.

[0071] The method proposed in the present invention can give full play to the application advantages of drones in forest inspection, improve the efficiency of forest fire inspection, detect fires in a timely manner, reduce manpower and material resources, and reduce personnel safety risks.

[0072] Considering the influence of wind in the path planning stage can effectively utilize the influence of wind to reduce inspection energy consumption, so that the drone can retain more energy to deal with possible sudden abnormalities; the heterogeneous take-off and landing operation mode can make the inspection path more flexible, reduce energy consumption and improve efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 Flowchart of the method of the present invention.

[0074] Figure 2 This is a distribution map of the inspection area terrain, aerial waypoints, and aircraft nests in an embodiment of the present invention.

[0075] Figure 3 This is a path diagram for solving the first stage in an embodiment of the present invention.

[0076] Figure 4 This is a path diagram for solving the second stage in an embodiment of the present invention.

[0077] Figure 5 This is a diagram of the inspection plan solved in stage three of an embodiment of the present invention. DETAILED DESCRIPTION

[0078] In order to facilitate understanding by those skilled in the art, the present invention will be further described below with reference to embodiments and drawings. The contents mentioned in the embodiments are not intended to limit the present invention.

[0079] Reference Figure 1As shown, the present invention provides a multi-UAV inspection path planning method considering wind influence and heterogeneous take-off and landing, and the steps are as follows:

[0080] 1) Data acquisition: Obtain geographic information data, the weight of the inspection drone, the battery energy data of the inspection drone, and the wind direction and speed forecast data during the inspection period;

[0081] Some of the data used in the example are shown in Table 1. The terrain of the inspection area and the distribution of aerial waypoints and machine nests are shown in Table 1. Figure 2 shown.

[0082] Table 1

[0083]

[0084] 2) Set parameters: Set the horizontal airspeed, ascent speed, and descent speed of the drone during inspection. In this example, set the drone's cruising speed to 10m / s, the ascent speed to 3m / s, and the descent speed to -2m / s.

[0085] 3) Establish an energy consumption estimation matrix for direct flights of drones under wind influences: Develop equations for energy consumption per meter of flight for drones in different states, both in a windless environment and under wind influences; determine the flight pattern of the drone between the starting and ending points of each flight segment; calculate the energy consumption of direct flights between any two points within the set of aerial visual coverage inspection waypoints, and establish an energy consumption estimation matrix for direct flights of drones under wind influences;

[0086] 31) Establish the energy consumption equation per meter of flight for UAVs under different conditions in a windless environment;

[0087] The energy consumption per meter of flight in the horizontal direction is calculated using the following formula:

[0088]

[0089] Where Epm represents the energy consumption per meter of flight in the horizontal direction in a windless environment, E M is the battery energy of the drone, v test The laboratory test speed is 10 m / s. During the test, the drone flies at this speed until the energy is exhausted. M With speed v test The maximum flight time of the drone measured in horizontal flight; set the flight speed v of the drone a =v test ;

[0090] The energy consumption per meter of flight in the vertical direction is calculated as follows:

[0091]

[0092] Where, Epm vIndicates the energy consumption per meter of the UAV in vertical flight, P v Indicates the flight power of the UAV in the vertical direction, v v Indicates the vertical flight speed of the drone;

[0093] 32) Establish the energy consumption equation per meter of a UAV flying horizontally under the influence of wind;

[0094] The two attributes of the wind vector w are recorded as wind speed v w and wind direction w ; Under a given wind vector w, the ground speed of the drone is calculated as follows:

[0095] v g =v a +v w cos(d a -d w ) (3)

[0096] Where, v g is the ground speed of the UAV, v a is the flight speed of the UAV, d a is the flight direction, d w For wind direction;

[0097] Considering the influence of the wind and headwind components on energy consumption, the energy consumption per meter of horizontal flight of the UAV under wind vector w is Epm w It can be expressed as:

[0098]

[0099] 33) Determine the flight pattern of the UAV between the starting point and the end point of each flight segment;

[0100] When the drone performs an inspection mission, the flight types between the starting point and the end point of each flight include flight between the machine nest and the aerial visual coverage waypoint and flight between two waypoints. For the flight between the machine nest and the waypoint, the vertical movement distance is long. When taking off, the drone is set to first rise vertically to the height of the target waypoint, and then fly horizontally towards the target waypoint. When landing, it first flies from the waypoint to the air above the target machine nest, and then lands vertically. For the flight between two waypoints, the drone is set to move simultaneously at specified speeds in both the horizontal and vertical directions. If the specified position in one direction is reached first, the subsequent speed in that direction is reset to 0.

[0101] 34) Establish an energy consumption estimation matrix for direct UAV flight under wind influence;

[0102] For two aerial visual coverage waypoints i and j in the waypoint set P, let their coordinates be (x i ,y i ,zi ),(x j ,y j ,z j ), calculate the flight direction d of the drone according to the coordinates a , and calculate Epm according to step 31) v , calculate Epm according to step 32) w , then the energy consumption of the UAV flying from point i to point j can be calculated by formula (5):

[0103]

[0104] Where, Represents the energy consumption of the drone flying from point i to point j under wind vector w; calculates the energy consumption of forward and reverse flights between any two aerial visual coverage waypoints under wind vector w, and thus establishes the direct flight energy consumption estimation matrix E between two aerial visual coverage waypoints under wind vector w w , matrix E w The element in the i-th row and j-th column is the energy consumption of the drone flying from point i to point j under the wind vector w.

[0105] The calculation of the vertical flight power of the UAV in step 31) is as follows:

[0106]

[0107] k1∈[0,1] (7)

[0108]

[0109]

[0110] Where, P v It represents the flight power of the UAV in the vertical direction, m is the weight of the UAV, g is the acceleration of gravity, N is the total number of blades of a single propeller, c is the d is the drag coefficient of the UAV blade, c is the chord width of the UAV blade, ρ is the air density, and R is the length of the UAV blade (the radius of the circle formed during rotation); k1 is a constant, k2 is the deviation factor between the ideal uniform airflow and the actual airflow, k3 is the proportional factor that converts the rotor angular velocity into thrust, and c2 is the intermediate coefficient.

[0111] In the example,

[0112] The calculated energy consumption matrix E of the direct flight of UAV under the influence of wind w As shown in Table 2:

[0113] Table 2

[0114]

[0115] 4) Solve a three-stage model for multi-UAV inspection path planning that considers wind effects and heterogeneous takeoffs and landings. In the first stage, a TSP problem is established to solve a Hamiltonian circuit that minimizes energy consumption under wind effects. In the second stage, path optimization is performed to minimize the total turning angle. In the third stage, multi-UAV task allocation is performed considering the heterogeneous takeoff and landing operation mode. With task balance as the goal, the optimal solution for the takeoff and landing positions of UAVs in each nest is calculated, and a multi-UAV inspection path is obtained that takes into account direct flight energy consumption, turning angles, and task allocation fairness under the given wind effects.

[0116] 41) Stage 1: Solving the Hamiltonian circuit for minimizing energy consumption under the influence of wind;

[0117] The cost matrix is ​​established using the direct flight energy consumption between each aerial visual coverage waypoint under the influence of wind vector w as an element, and a TSP problem model is established to solve the Hamiltonian circuit with the minimum cost; the details are as follows:

[0118]

[0119]

[0120]

[0121]

[0122]

[0123] Where E represents the total energy consumption of the UAV that traverses the Hamiltonian circuit of all aerial visual coverage waypoints; represents the energy consumption of the UAV flying from point i to point j under wind vector w; X ij is a decision variable. When it is 1, it means the UAV flies from point i to point j. When it is 0, it means the UAV does not fly from point i to point j. i 、u j represents the order in which the UAV visits points i and j, and P is the set of waypoints;

[0124] The branch and bound method is used to solve the model of formula (10)-(14); the branch and bound method initializes the problem and creates a root node, then continuously expands the nodes and uses the bound function to evaluate the possibility of the optimal solution of each subproblem until the optimal solution is found or no better solution can be found; in this example, the total energy consumption of the path obtained in stage 1 is 1105109.36J, and the path obtained is as follows Figure 3 shown.

[0125] 42) Stage 2: Path optimization with the goal of minimizing the total turning angle;

[0126] Further considering the turning minimization objective, a multi-objective TSP model is established; using A ijqIndicates that the continuous path points are p i ,p j ,p q When the drone is in p j The turning angle at is calculated as follows:

[0127] A ijq =180°-∠p i p j p q (15)

[0128] The objective function of minimizing the turning angle is established as follows:

[0129]

[0130] Where A is the total turning angle of the inspection loop. The multi-objective TSP model considering the minimization of total energy consumption and total turning amount is established by equations (10) to (16). The Hamiltonian loop obtained in stage 1 is used as the initial solution to input into the NSGA-II algorithm. A solution is selected from the Pareto front solution in which the turning amount is reduced while the direct flight energy consumption does not increase. The inspection loop obtained in stage 2 is set as the waypoint sequence {x0, x1, x2, ..., x |P| ,x0}; In the example, the total turning angle of the path obtained in stage 1 is 7959.15°, the energy consumption of the inspection loop obtained in stage 2 is 1107844.04J, and the turning angle is 7587.66°. Compared with the results obtained in stage 1, the energy consumption is increased by 0.39% and the turning angle is reduced by 3.15%. The obtained path is as follows Figure 4 shown.

[0131] 43) Phase 3: Multi-UAV task allocation considering heterogeneous take-off and landing operation modes;

[0132] Insert |V| nests into the inspection loop as the landing point of the preceding UAV and the take-off point of the following UAV, so as to segment the inspection path and assign it to each UAV. V is the nest set. For a nest v, its candidate insertion position set is recorded as Candi(v). Define the n nearest two-dimensional distances around it. Near The aerial visual coverage waypoints are the neighboring waypoints of the aircraft nest, denoted as Near(v); if there are two consecutive waypoints x in the path obtained in stage 2 i 、x i+1 are all in Near(v), then x i Add Candi(v) as the candidate insertion position of the nest v; the details are as follows:

[0133]

[0134] Decision variable X vi The details are as follows:

[0135]

[0136] After the nest is inserted, the inspection loop is divided into |V| segments. The starting and ending points of each segment are both nests. The segment path is inspected by the drone in the nest at the starting point of the segment. Suppose the flight path S of the drone taking off from nest v is v for Then the energy consumption E(v) of the UAV is:

[0137]

[0138] Where, Indicates that the drone is from point Flight to point The energy consumed is the energy consumed by the drone taking off from the nest v to fly from the i-th point to the i+1-th point on the corresponding inspection path; a minimum-maximum objective function is established to find an inspection task allocation scheme that minimizes the maximum inspection energy consumption of all drones and ensures that all drones have a relatively fair workload; the specific objective function is as follows:

[0139]

[0140] The constraints of the stage three model are as follows:

[0141]

[0142]

[0143] Where, E(v) represents the inspection energy consumption of the UAV taking off from the nest v, E M represents the battery energy of the drone, λ E is the proportional coefficient; after the nest is inserted into the flight path of the UAV, it becomes the landing nest of the preceding UAV and the take-off nest of the following UAV. Formula (21) indicates that each nest can only be inserted once; Formula (22) indicates that the flight energy consumption of each UAV shall not exceed the maximum energy consumption set for normal inspection;

[0144] By solving the mathematical model composed of equations (17)-(22), we can obtain the multi-UAV inspection path that takes into account the direct flight energy consumption, turning angle and task allocation fairness under the influence of the set wind.

[0145] Using genetic algorithms to solve the mathematical model composed of equations (16)-(21), a multi-UAV inspection path planning scheme that takes into account direct flight energy consumption, turning angles, and task allocation fairness under the influence of set wind can be obtained. In this example, the objective function value is 337263.83, and the energy consumption of the four UAVs is 296538.31, 333565.36, 335330.62, and 337263.83 respectively. The total energy consumption is 1302698.11; the resulting path is as follows Figure 5 shown.

[0146] The present invention has many specific application paths. The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements can be made without departing from the principles of the present invention. These improvements should also be considered as the scope of protection of the present invention.

Claims

1. A multi-UAV inspection path planning method considering wind influence and heterogeneous takeoff and landing, characterized by: Here are the steps: 1) Data acquisition: Obtain geographic information data, the weight of the inspection drone, the battery energy data of the inspection drone, and the wind direction and speed forecast data during the inspection period; 2) Set parameters: Set the horizontal airspeed and the ascent and descent speeds of the drone during inspection; 3) Establish an energy consumption estimation matrix for direct flights of drones under wind influences: Develop equations for energy consumption per meter of flight for drones in different states, both in a windless environment and under wind influences; determine the flight pattern of the drone between the starting and ending points of each flight segment; calculate the energy consumption of direct flights between any two points within the set of aerial visual coverage inspection waypoints, and establish an energy consumption estimation matrix for direct flights of drones under wind influences; 4) Solve a three-stage model for multi-UAV inspection path planning that considers wind effects and heterogeneous takeoffs and landings. In the first stage, a TSP problem is established to solve a Hamiltonian circuit that minimizes energy consumption under wind effects. In the second stage, path optimization is performed to minimize the total turning angle. In the third stage, multi-UAV task allocation is performed considering the heterogeneous takeoff and landing operation mode. With task balance as the goal, the optimal solution for the takeoff and landing positions of UAVs in each nest is calculated, and a multi-UAV inspection path is obtained that takes into account direct flight energy consumption, turning angles, and task allocation fairness under the given wind effects. The step 3) specifically includes: 31) Establish the energy consumption equation per meter of flight for UAVs under different conditions in a windless environment; The energy consumption per meter of flight in the horizontal direction is calculated using the following formula: Where Epm represents the energy consumption per meter of flight in the horizontal direction in a windless environment, E M is the battery energy of the drone, v test For laboratory test speed; t M With speed v test The maximum flight time of the drone measured in horizontal flight; set the flight speed v of the drone a =v test ; The energy consumption per meter of flight in the vertical direction is calculated as follows: Where, Epm v Indicates the energy consumption per meter of the UAV in vertical flight, P v Indicates the flight power of the UAV in the vertical direction, v v Indicates the vertical flight speed of the drone; 32) Establish the energy consumption equation per meter of a UAV flying horizontally under the influence of wind; The two attributes of the wind vector w are recorded as wind speed v w and wind direction w ; Under a given wind vector w, the ground speed of the drone is calculated as follows: in g =in a +in w cos(d a -d w ) (3) Where, v g is the ground speed of the UAV, v a is the flight speed of the UAV, d a is the flight direction, d w For wind direction; Considering the influence of the wind and headwind components on energy consumption, the energy consumption per meter of horizontal flight of the UAV under wind vector w is Epm w It can be expressed as: 33) Determine the flight pattern of the UAV between the starting point and the end point of each flight segment; When the drone performs an inspection mission, the flight types between the starting point and the end point of each flight include flight between the machine nest and the aerial visual coverage waypoint and flight between two waypoints. For the flight between the machine nest and the waypoint, the vertical movement distance is long. When taking off, the drone is set to first rise vertically to the height of the target waypoint, and then fly horizontally towards the target waypoint. When landing, it first flies from the waypoint to the air above the target machine nest, and then lands vertically. For the flight between two waypoints, the drone is set to move simultaneously at specified speeds in both the horizontal and vertical directions. If the specified position in one direction is reached first, the subsequent speed in that direction is reset to 0. 34) Establish an energy consumption estimation matrix for direct UAV flight under wind influence; For two aerial visual coverage waypoints i and j in the waypoint set P, let their coordinates be (x i ,y i ,z i ),(x j ,y j ,z j ), calculate the flight direction d of the drone according to the coordinates a , and calculate Epm according to step 31) v , calculate Epm according to step 32) w , then the energy consumption of the UAV flying from point i to point j is calculated by formula (5): Where, Represents the energy consumption of the drone flying from point i to point j under wind vector w; calculates the energy consumption of forward and reverse flights between any two aerial visual coverage waypoints under wind vector w, and thus establishes the direct flight energy consumption estimation matrix E between two aerial visual coverage waypoints under wind vector w w , matrix E w The element in the i-th row and j-th column is the energy consumption of the drone flying from point i to point j under the wind vector w.

2. The multi-UAV inspection path planning method considering wind influence and heterogeneous takeoff and landing according to claim 1 is characterized in that: The geographic location information data in step 1) includes the latitude and longitude range of the inspection area, digital elevation, three-dimensional coordinates of the machine nest and aerial visual coverage inspection waypoints.

3. The multi-UAV inspection path planning method considering wind influence and heterogeneous takeoff and landing according to claim 1 is characterized in that: The step 2) specifically includes: setting the horizontal airspeed v of the drone during inspection a For lab test speed v test , vertical flight speed v v It is divided into rising speed and falling speed. When rising, v v The value is positive, when v decreases v The value is negative.

4. The multi-UAV inspection path planning method considering wind influence and heterogeneous takeoff and landing according to claim 1 is characterized in that: The calculation of the vertical flight power of the UAV in step 31) is as follows: k1∈[0,1](7) Where, P v It represents the flight power of the UAV in the vertical direction, m is the weight of the UAV, g is the acceleration of gravity, N is the total number of blades of a single propeller, c is the d is the drag coefficient of the UAV blade, c is the chord width of the UAV blade, ρ is the air density, and R is the length of the UAV blade; k1 is a constant, k2 is the deviation factor between the ideal uniform airflow and the actual airflow, k3 is the proportional factor that converts the rotor angular velocity into thrust, and c2 is the intermediate coefficient.

5. The multi-UAV inspection path planning method considering wind influence and heterogeneous takeoff and landing according to claim 1 is characterized in that: The step 4) specifically includes: 41) Stage 1: Solving the Hamiltonian circuit for minimizing energy consumption under the influence of wind; The cost matrix is ​​established using the direct flight energy consumption between each aerial visual coverage waypoint under the influence of wind vector w as an element, and a TSP problem model is established to solve the Hamiltonian circuit with the minimum cost; the details are as follows: Where E represents the total energy consumption of the UAV that traverses the Hamiltonian circuit of all aerial visual coverage waypoints; represents the energy consumption of the UAV flying from point i to point j under wind vector w; X ij is a decision variable. When it is 1, it means the UAV flies from point i to point j. When it is 0, it means the UAV does not fly from point i to point j. i 、u j represents the order in which the UAV visits points i and j, and P is the set of waypoints; The branch and bound method is used to solve the model of Equations (10)-(14). The branch and bound method initializes the problem and creates a root node. Then, the optimal solution possibility of each subproblem is evaluated by continuously expanding the nodes and using the bound function until the optimal solution is found or no better solution can be found. 42) Stage 2: Path optimization with the goal of minimizing the total turning angle; Further considering the turning minimization objective, a multi-objective TSP model is established; using A ijq Indicates that the continuous path points are p i ,p j ,p q When the drone is in p j The turning angle at is calculated as follows: A ijq =180°-∠p i p j p q (15) The objective function of minimizing the turning angle is established as follows: Where A is the total turning angle of the inspection loop. The multi-objective TSP model considering the minimization of total energy consumption and total turning amount is established by equations (10) to (16). The Hamiltonian loop obtained in stage 1 is used as the initial solution to input into the NSGA-II algorithm. A solution is selected from the Pareto front solution in which the turning amount is reduced while the direct flight energy consumption does not increase. The inspection loop obtained in stage 2 is set as the waypoint sequence {x0, x1, x2, ..., x |P| ,x0}; 43) Phase 3: Multi-UAV task allocation considering heterogeneous take-off and landing operation modes; Insert |V| nests into the inspection loop as the landing point of the preceding UAV and the take-off point of the following UAV, so as to segment the inspection path and assign it to each UAV. V is the nest set. For a nest v, its candidate insertion position set is recorded as Candi(v). Define the n nearest two-dimensional distances around it. Near The aerial visual coverage waypoints are the neighboring waypoints of the aircraft nest, denoted as Near(v); if there are two consecutive waypoints x in the path obtained in stage 2 i 、x i+1 are all in Near(v), then x i Add Candi(v) as the candidate insertion position of the nest v; the details are as follows: Decision variable X vi The details are as follows: After the nest is inserted, the inspection loop is divided into |V| segments. The starting and ending points of each segment are both nests. The segment path is inspected by the drone in the nest at the starting point of the segment. Suppose the flight path S of the drone taking off from nest v is v for Then the energy consumption E(v) of the UAV is: Where, Indicates that the drone is from point Flight to point The energy consumed is the energy consumed by the drone taking off from the nest v to fly from the i-th point to the i+1-th point on the corresponding inspection path; a minimum-maximum objective function is established to find an inspection task allocation scheme that minimizes the maximum inspection energy consumption of all drones and ensures that all drones have a relatively fair workload; the objective function is as follows: min max{E(v),v∈V}(20) The constraints of the stage three model are as follows: Where, E(v) represents the inspection energy consumption of the UAV taking off from the nest v, E M represents the battery energy of the drone, λ E is the proportional coefficient; after the nest is inserted into the flight path of the UAV, it becomes the landing nest of the preceding UAV and the take-off nest of the following UAV. Formula (21) indicates that each nest can only be inserted once; Formula (22) indicates that the flight energy consumption of each UAV shall not exceed the maximum energy consumption set for normal inspection; By solving the mathematical model composed of equations (17)-(22), we can obtain the multi-UAV inspection path that takes into account the direct flight energy consumption, turning angle and task allocation fairness under the influence of the set wind.

Images