Multi-uav cooperative path planning method based on adaptive multi-population hybrid algorithm

Abstract

The application provides a multi-unmanned aerial vehicle cooperative path planning method based on an adaptive multi-population hybrid algorithm, and comprises the following steps: 1, a multi-unmanned aerial vehicle cooperative path planning model is constructed, which comprises path length limitation, flight height limitation, unmanned aerial vehicle maneuvering constraint and obstacle avoidance limitation, unmanned aerial vehicle cooperation constraint and a target function; 2, the adaptive multi-population hybrid algorithm is applied to the multi-unmanned aerial vehicle cooperative path planning problem in a three-dimensional complex scene, and a multi-unmanned aerial vehicle cooperative path planning scheme is obtained. The application can solve the multi-unmanned aerial vehicle cooperative path planning problem in a three-dimensional scene, and obtain an optimal multi-unmanned aerial vehicle flight path, so that the path planning cost is reduced, and the unmanned aerial vehicle flight path is improved.

Description

Technical Field

[0001] The present invention belongs to the field of multi-UAV collaborative path planning, and specifically provides a multi-UAV collaborative path planning method based on an adaptive multi-population hybrid algorithm. Background Art

[0002] Due to their low cost, high flexibility, and excellent performance, drones have been widely used in military and civilian fields, including military offensives, smart agriculture, environmental monitoring, and search and rescue missions in disaster scenarios. With the development of the times and technological advancements, the tasks required of drones are becoming increasingly complex. A single drone cannot complete all tasks, so multiple drones are needed to complete the task together. Drone path planning is crucial for drones to complete their missions, especially in military combat environments. Drones must be able to intelligently plan paths to avoid threats such as enemy radar, missiles, and artillery, and complete the mission as quickly as possible while ensuring their own safety. As a result, collaborative path planning among multiple drones in complex environments has become a hot topic in drone research.

[0003] Collaborative path planning for multiple UAVs is an NP-hard problem and a critical component of military attack missions. The goal of path planning is to safely reach the mission point at the lowest cost. During flight, UAVs must not only avoid obstacles but also maintain a safe distance between them to avoid collisions. Approaches to solving UAV path planning include classical algorithms and intelligent optimization algorithms. Among early path planning algorithms, classical algorithms such as the rapidly exploring random tree (RRT) method, Voronoi diagrams, probabilistic roadmaps (PRMs), the A* algorithm, and artificial potential fields (APFs) were popular. These methods are relatively simple to implement and can solve path planning problems in simple scenarios, particularly in two-dimensional environments. However, when the planning space becomes complex, particularly in three-dimensional environments, the complexity of the solution becomes extremely high and they are prone to falling into local optima.

[0004] As the problem becomes more complex, classic algorithms are no longer suitable for solving UAV path planning problems. Intelligent optimization methods have gradually been used to solve multi-UAV path planning problems in three-dimensional scenarios, such as particle swarm optimization algorithm (PSO), genetic algorithm (GA), ant colony optimization algorithm (ACO), artificial bee colony algorithm (ABC), fruit fly optimization algorithm (FOA), differential evolution algorithm (DE), pigeon-inspired optimization algorithm (PIO), gray wolf optimization algorithm (GWO), etc. However, a single intelligent optimization method still has problems such as slow convergence, easy to fall into local optimality, and poor global search performance when solving complex multi-constrained optimization multi-UAV collaborative path planning problems. Summary of the Invention

[0005] In order to address the shortcomings of the above-mentioned existing technologies, the present invention proposes a multi-UAV collaborative path planning method based on an adaptive multi-population hybrid algorithm, in order to solve the multi-UAV collaborative path planning problem in a three-dimensional scene, thereby ensuring that multiple UAVs safely reach the target point and complete the flight mission, and reducing the cost of multi-UAV collaborative path planning, thereby obtaining the optimal multi-UAV path planning solution.

[0006] In order to achieve the above-mentioned object, the present invention adopts the following technical solutions:

[0007] The invention discloses a multi-UAV collaborative path planning method based on an adaptive multi-population hybrid algorithm. The method is applied to a multi-UAV collaborative path planning method in which N UAVs start from a known starting position. Fly to their respective target locations In a three-dimensional flight scene, obstacles with known positions are set in the three-dimensional flight scene, wherein, Indicates the starting position of the nth drone, The target point position of the nth UAV is represented by the multi-UAV collaborative path planning method according to the following steps:

[0008] Step 1: Construct the objective function of the multi-UAV collaborative path planning model;

[0009] Step 1.1: Let the flight path of the nth drone be ,in, is the number of trajectory points, represents the kth trajectory point of the nth UAV, and The coordinates of , and use formula (1) to establish the flight length model of the nth UAV :

[0010] (1)

[0011] In formula (1), Indicates that the nth drone starts from Track points To Track Points Flight distance;

[0012] Step 1.2: Use equations (2) and (3) to construct the flight altitude model of the nth UAV :

[0013] (2)

[0014] (3)

[0015] In formula (2) and formula (3), is the nth drone at k trajectory point The flight altitude model, is the minimum altitude at which the drone can fly. is the maximum altitude at which the drone can fly. Indicates that the nth drone is in the Track points The flight altitude, Indicates that the nth drone is in the -1 track point Flight altitude;

[0016] Step 1.3: Construct the maneuvering physical constraint model of the nth UAV, including the deflection angle constraint model of the nth UAV and the climb angle constraint model of the nth UAV ;

[0017] Step 1.4: Build the obstacle threat model for the nth UAV, including the radar threat model for the nth UAV and the missile threat model of the nth drone ;

[0018] Step 1.5: Construct the coordination model of the nth UAV, including: time coordination model and spatial collaboration models ;

[0019] Step 1.6: Use Equation (16) to construct the objective function of the multi-UAV collaborative path planning model :

[0020] (16)

[0021] In formula (16), , , , , are 5 weight factors;

[0022] Step 2: Use the adaptive multi-population hybrid algorithm to solve the multi-UAV collaborative path model and obtain the multi-UAV path planning solution with the lowest path planning cost.

[0023] The multi-UAV collaborative path planning method based on the adaptive multi-population hybrid algorithm of the present invention is also characterized in that step 1.3 includes:

[0024] Step 1.3.1: Use equations (4) and (5) to construct the deflection angle model and climb angle model of the nth UAV respectively:

[0025] (4)

[0026] (5)

[0027] In formula (4) and formula (5), Indicates that the nth drone is at the kth trajectory point The deflection angle, Indicates that the nth drone is at the kth trajectory point The climbing angle, Represents the k+1th trajectory point of the nth UAV coordinates of

[0028] Step 1.3.2: Use equations (6) and (7) to construct the deflection angle constraint model of the nth UAV ;

[0029] (6)

[0030] (7)

[0031] In formula (6) and formula (7), is the nth drone at the kth trajectory point The deflection angle constraint, is the deflection angle of the nth UAV at the kth trajectory point, is the maximum deflection angle of the UAV;

[0032] Step 1.3.3: Use equations (8) and (9) to construct the climb angle constraint model of the nth UAV :

[0033] (8)

[0034] (9)

[0035] In formula (8) and formula (9), is the nth drone at the kth trajectory point The climb angle constraint, is the climbing angle of the nth UAV at the kth trajectory point, is the maximum climb angle of the UAV.

[0036] Furthermore, the step 1.4 includes:

[0037] Step 1.4.1: Use equation (10) to construct the radar model:

[0038] (10)

[0039] In formula (10), Represents any three-dimensional coordinate in a three-dimensional flight scene, It refers to the area that the radar can cover. is the location of the radar area center, is the radius of the radar;

[0040] Step 1.4.2: Use Equation (11) to construct the radar threat model of the nth UAV :

[0041] (11)

[0042] In formula (11), Indicates the nth drone Track points Hedi Track points The distance between Indicates the nth drone Track points With the Track points The trajectory segment Two intersection points with the radar area and The distance between Indicates the nth drone Track points Second intersection with the radar area The distance between Indicates the nth drone Track points First intersection with the radar area The distance between Represents a trajectory segment The distance from the midpoint of to the center of the radar area;

[0043] Step 1.4.3: Use Equation (12) to construct the missile model :

[0044] (12)

[0045] In formula (12), It is the area that the missile can cover. is the coordinate of the center of the bottom surface of the missile, Indicates the radius of the missile's bottom surface, Indicates the missile's altitude;

[0046] Step 1.4.4: Use equation (13) to construct a missile threat model:

[0047] (13)

[0048] In formula (13), Indicates the nth drone Track points Hedi Track points The distance between Indicates the nth drone Track points Hedi Track points The trajectory segment Two intersection points with the missile area and The distance between Indicates the nth drone Track points Second intersection with the missile area The distance between Indicates the nth drone Track points First intersection with the missile area The distance between express The distance from the midpoint of the missile to the center of the missile bottom circle.

[0049] Furthermore, the step 1.5 includes:

[0050] Step 1.5.1: Use Equation (14) to construct the time coordination model of the nth UAV :

[0051] (14)

[0052] In formula (14), is the flight time of the nth drone, Indicates the time it takes for the first drone to reach the corresponding target point, Indicates the buffer time for the remaining drones after the first one completes its flight mission;

[0053] Step 1.5.2: Use Equation (15) to construct the spatial coordination model of the nth UAV :

[0054] (15)

[0055] In formula (15), Indicates the The kth path point of the UAV With the The first drone Waypoints distance, is the penalty coefficient for drones violating the space safety distance, The safe distance between drones.

[0056] Furthermore, the step 2 includes:

[0057] Step 2.1: Initialize the starting and target locations of the mission, the number of drones N, the locations of radar obstacles, the locations of missiles, and the number of path points K;

[0058] Step 2.2: Initialize the population size to NP, the maximum number of iterations to itermax, the dimension of individuals in the population to D = N × K, define the current number of iterations to t, and initialize t = 1;

[0059] Step 2.3: Move N drones from the starting position Fly to their respective target locations The path planning scheme is encoded as an individual, so that the i-th individual in the t-th generation population is obtained ,in, represents the i-th individual The position of the jth dimension, indicating the The first drone Waypoint locations; Indicates rounding up, % indicates remainder;

[0060] Step 2.4: Calculate the i-th individual in the t-th generation population using formula (16) The fitness value of And find the top three wolves with the smallest fitness value from the t-th generation population and record them as the t-th generation population Wolf 、 Wolf and Wolf ;in, represents the population in the tth generation The j-th dimension position of the wolf individual, represents the population in the tth generation The j-th dimension position of the wolf individual, represents the population in the tth generation The j-th dimension position of the wolf individual;

[0061] Step 2.5: If the individuals in the t-th generation population belong to the top three levels of wolves, use formula (17) to update , and , get the updated Wolf individual , updated Wolf individual , updated Wolf individual ;

[0062] (17)

[0063] In formula (17), represents the gray wolf parameter of the tth iteration, and , and represents the j-th dimension position of the r1-th individual and the r2-th individual randomly selected from the t-th generation population, and ; represents the first random number of the tth generation between 0 and 1;

[0064] If the i-th individual in the t-th generation population does not belong to the top three levels of wolves, then the updated i-th individual is calculated using formula (18): ; Thus we get the tth generation updated population ;

[0065] (18)

[0066] In formula (18), 、 and They are respectively the affected populations in the t generation and the Wolf individual, Wolf wolf individuals and The j-th dimension element of the i-th individual affected by the wolf individual, 、 and are the tth generation populations Wolf individuals, Wolf individuals and The weight of the individual wolf;

[0067] Step 2.6: Use formula (16) to calculate the updated i-th individual The fitness value of ,like < , then As the i-th individual in the intermediate population of generation t Otherwise, As the i-th individual in the intermediate population of generation t ; Thus we get the intermediate population of generation t;

[0068] Step 2.7: If t ≤ (2 / 3) × itermax, then divide the t-th generation renewal population according to the flight altitude of the individuals in the t-th generation intermediate population, and divide the first Δ individuals with the highest flight altitude into the t-th generation high-altitude sub-population, the last Δ individuals into the t-th generation low-altitude sub-population, and the middle Δ individuals into the t-th generation medium-altitude sub-population. Otherwise, divide the t-th generation renewal population according to the fitness values ​​of the individuals in the t-th generation intermediate population, and divide the first Δ individuals with the smallest fitness values ​​into the t-th generation good sub-population, the middle Δ individuals into the t-th generation medium sub-population, and the last Δ individuals into the t-th generation poor sub-population.

[0069] Step 2.8: When t≤(2 / 3)itermax, if the updated i-th individual belongs to the high-altitude subpopulation, then according to formula (19) we can get the ith mutant individual of the tth generation: ,in, The j-th dimension position of the i-th mutant individual in the t-generation population represented by ;

[0070] (19)

[0071] In formula (19), and are randomly selected from the hollow subpopulation of the t-th generation update population. Individuals and The j-th dimension position of each individual, and , is the first randomly selected from the high-altitude population of the t-th generation update population. The j-th dimension position of each individual; represents the second random number of the tth generation between 0 and 1;

[0072] like Belongs to the low-altitude subpopulation, then according to formula (20) we can get the ith mutant individual of the tth generation ;

[0073] (20)

[0074] In formula (20), , ,and is the first population randomly selected from the empty subpopulation of the t-th generation update population. Individual, Individual, The j-th dimension position of each individual, ; represents the third random number of the tth generation between 0 and 1;

[0075] like Belongs to the hollow subpopulation, then according to formula (21) we can get the ith mutant individual of the tth generation :

[0076] (twenty one)

[0077] In formula (21), , ,and is the first randomly selected from the empty subpopulation of the t-th generation update population. Individual, Individual, The j-th dimension position of each individual, and is the first randomly selected population from the low-space subpopulation of the t-th generation renewal population. Individual, The j-th dimension position of each individual, ; represents the fourth random number of the tth generation between 0 and 1;

[0078] When t>(2 / 3)itermax, if the updated i-th individual Belongs to a better subpopulation, then according to formula (22) we can get the ith mutant individual of the tth generation ;

[0079] (twenty two)

[0080] In formula (22), is the j-th dimension position of the individual with the best fitness value in the t-th generation updated population, and are randomly selected from the better subpopulation of the t-th generation update population. Individual, The j-th dimension position of each individual, , F is the mutation operator;

[0081] like Belongs to the medium subpopulation, then according to formula (23) we can get the ith mutant individual of the tth generation ;

[0082] (twenty three)

[0083] In formula (23), Indicates the j-th dimension position of the individual with the best fitness value in the medium subpopulation of the t-th generation updated population, and are randomly selected from the medium subpopulation of the t-th generation update population. Individual, The j-th dimension position of each individual, ;

[0084] like Belonging to the inferior subpopulation, according to formula (24), the ith mutant individual of the tth generation is obtained :

[0085] (twenty four)

[0086] In formula (24), and are randomly selected from the inferior subpopulation of the t-th generation update population. Individual, The j-th dimension position of each individual, represents the first randomly selected from the tth generation update population The j-th dimension position of each individual, ;

[0087] Step 2.9: Generate the i-th crossover individual of the t-th generation using formula (26) :

[0088] (25)

[0089] In formula (25), is the crossover operator, is a random integer between [1,D], represents the fifth random number of the tth generation between 0 and 1, yes The j-th dimension position in ;

[0090] Step 2.10: Calculate the i-th cross-individual vector according to formula (16) The fitness value of , and thus use formula (26) to get the i-th individual in the t+1 generation population :

[0091] (26)

[0092] Step 2.11: Assign t+1 to t. If t>itermax, it means that the itermax-th generation population is obtained, and the individual with the smallest fitness value is selected from it as the multi-UAV collaborative path planning solution. Otherwise, return to step 2.4 and execute sequentially.

[0093] The electronic device of the present invention includes a memory and a processor, and is characterized in that the memory is used to store a program that supports the processor to execute the multi-UAV collaborative path planning method, and the processor is configured to execute the program stored in the memory.

[0094] The present invention provides a computer-readable storage medium, and the computer-readable storage medium stores a computer program, which is characterized in that when the computer program is run by a processor, the steps of the multi-UAV collaborative path planning method are executed.

[0095] Compared with the prior art, the present invention has the following beneficial effects:

[0096] 1. This paper proposes a multi-UAV collaborative path planning model. In addition to considering constraints such as path length cost, flight altitude limit, UAV maneuvering constraints and obstacle avoidance, this model also considers collaborative constraints between UAVs. A comprehensive multi-constraint objective function is then established. This allows the model to solve path planning solutions for multi-UAV path planning problems in complex three-dimensional scenarios, thus completing the multi-UAV path planning task.

[0097] 2. This invention applies an adaptive multi-population hybrid algorithm to the collaborative path planning problem for multiple UAVs. This algorithm improves the position update formula for individual gray wolves to enhance their exploitability. In the early stages of the algorithm, the algorithm divides the solution into multiple populations based on their characteristics, employing different global search strategies. A cooperative pairing strategy coordinates information exchange between subpopulations, improving their diversity and exploratory potential. In the later stages of the algorithm, the populations are divided based on fitness values, allowing them to learn from optimal solutions and accelerate convergence. This allows for faster solution of collaborative path planning solutions for multiple UAVs, reduces flight path costs, and ensures the safe completion of UAV missions. BRIEF DESCRIPTION OF THE DRAWINGS

[0098] Figure 1 Flowchart of the method of the present invention. DETAILED DESCRIPTION

[0099] In an embodiment, a multi-UAV collaborative path planning method based on an adaptive multi-population hybrid algorithm is applied to a multi-UAV collaborative path planning method consisting of N UAVs starting from a known starting position. Fly to their respective target locations In a three-dimensional flight scene with known obstacles, we need to find a path planning solution with the lowest path planning cost to ensure that the drone can reach the target point efficiently and avoid all obstacles. Indicates the starting position of the nth drone, Indicates the target point position of the nth UAV, such as Figure 1 As shown in the figure, the multi-UAV collaborative path planning method is carried out in the following steps:

[0100] Step 1: Construct the objective function of the multi-UAV collaborative path planning model;

[0101] Step 1.1: Let the flight path of the nth drone be ,in, is the number of trajectory points, represents the kth trajectory point of the nth UAV, and The coordinates of The shorter the flight length of a UAV is, the faster it can complete its flight mission and consume less energy. The flight length model of the nth UAV is established using formula (1): :

[0102] (1)

[0103] In formula (1), Indicates that the nth drone starts from Track points To Track Points flight distance.

[0104] Step 1.2: When a drone violates the altitude limit, it increases the risk of completing the mission. Use equations (2) and (3) to build the flight altitude model of the nth drone. :

[0105] (2)

[0106] (3)

[0107] In formula (2) and formula (3), is the nth drone at k trajectory point The flight altitude model, is the minimum altitude at which the drone can fly. is the maximum altitude at which the drone can fly. Indicates that the nth drone is in the Track points The flight altitude, Indicates that the nth drone is in the -1 track point flight altitude.

[0108] Step 1.3: The UAV will be subject to its own maneuverability constraints during flight. Construct the maneuverability physical constraint model of the nth UAV, including: the deflection angle constraint model of the nth UAV and the climb angle constraint model of the nth UAV :

[0109] Step 1.3.1: Use equations (4) and (5) to construct the deflection angle model and climb angle model of the nth UAV respectively:

[0110] (4)

[0111] (5)

[0112] In formula (4) and formula (5), Indicates that the nth drone is at the kth trajectory point The deflection angle, Indicates that the nth drone is at the kth trajectory point The climbing angle, Represents the k+1th trajectory point of the nth UAV 's coordinates.

[0113] Step 1.3.2: Use equations (6) and (7) to construct the deflection angle constraint model of the nth UAV ;

[0114] (6)

[0115] (7)

[0116] In formula (6) and formula (7), is the nth drone at the kth trajectory point The deflection angle constraint, is the deflection angle of the nth UAV at the kth trajectory point, is the maximum deflection angle of the UAV.

[0117] Step 1.3.3: Use equations (8) and (9) to construct the climb angle constraint model of the nth UAV :

[0118] (8)

[0119] (9)

[0120] In formula (8) and formula (9), is the nth drone at the kth trajectory point The climb angle constraint, is the climbing angle of the nth UAV at the kth trajectory point, is the maximum climb angle of the UAV.

[0121] Step 1.4: The UAV should avoid encountering radar and missile obstacles during flight to ensure that it can reach the target point safely. Construct the obstacle threat model of the nth UAV, including: the radar threat model of the nth UAV and the missile threat model of the nth drone :

[0122] Step 1.4.1: Use equation (10) to construct the radar model:

[0123] (10)

[0124] In formula (10), Represents any three-dimensional coordinate in a three-dimensional flight scene, It refers to the area that the radar can cover. is the location of the radar area center, is the radius of the radar.

[0125] Step 1.4.2: Use Equation (11) to construct the radar threat model of the nth UAV :

[0126] (11)

[0127] In formula (11), Indicates the nth drone Track points Hedi Track points The distance between Indicates the nth drone Track points With the Track points The trajectory segment Two intersection points with the radar area and The distance between Indicates the nth drone Track points Second intersection with the radar area The distance between Indicates the nth drone Track points First intersection with the radar area The distance between Represents a trajectory segment The distance from the midpoint of the radar to the center of the radar area.

[0128] Step 1.4.3: Use Equation (12) to construct the missile model :

[0129] (12)

[0130] In formula (12), It is the area that the missile can cover. is the coordinate of the center of the bottom surface of the missile, Indicates the radius of the missile's bottom surface, Indicates the missile's altitude.

[0131] Step 1.4.4: Use equation (13) to construct a missile threat model:

[0132] (13)

[0133] In formula (13), Indicates the nth drone Track points Hedi Track points The distance between Indicates the nth drone Track points Hedi Track points The trajectory segment Two intersection points with the missile area and The distance between Indicates the nth drone Track points Second intersection with the missile area The distance between Indicates the nth drone Track points First intersection with the missile area The distance between express The distance from the midpoint of the missile to the center of the missile bottom circle.

[0134] Step 1.5: In multi-UAV path planning, it is necessary to consider the coordination between UAVs, including time coordination and space coordination. Time coordination means that the time for multiple UAVs to reach the corresponding target point should meet certain requirements. When this constraint is violated, time coordination costs will be generated. Space coordination refers to the constraints between multiple UAVs in the flight space, which means that UAVs must maintain a safe distance. When this constraint is violated, space coordination costs will be generated. Construct a coordination model for the nth UAV, including: time coordination model and spatial collaborative models .

[0135] Step 1.5.1: Use Equation (14) to construct the time coordination model of the nth UAV :

[0136] (14)

[0137] In formula (14), is the flight time of the nth drone, Indicates the time it takes for the first drone to reach the corresponding target point, Indicates the buffer time for the remaining drones after the first one completes its flight mission;

[0138] Step 1.5.2: Use Equation (15) to construct the spatial coordination model of the nth UAV :

[0139] (15)

[0140] In formula (15), Indicates the The kth path point of the UAV With the The first drone Waypoints distance, is the penalty coefficient for drones violating the space safety distance, The safe distance between drones.

[0141] Step 1.6: Use Equation (16) to construct the objective function of the multi-UAV collaborative path planning model :

[0142] (16)

[0143] In formula (16), , , , , There are 5 weighting factors.

[0144] Step 2: Solve the multi-UAV collaborative path model using a species-based adaptive multi-population hybrid algorithm. The adaptive multi-population hybrid algorithm includes an improved gray wolf optimization algorithm and an adaptive multi-population differential evolution algorithm. It improves the position update formula of gray wolf individuals and enhances development. In the early stage of the algorithm, the solution is divided into multiple populations based on the characteristics, different global search strategies are adopted, and information exchange between populations is coordinated through cooperative pairing strategies, which improves the diversity and exploratory nature of the population. In the later stage of the algorithm, the population is divided based on the fitness value, allowing it to learn from the excellent solution and accelerate convergence. In this way, the multi-UAV collaborative path planning problem can be better solved. The adaptive multi-population hybrid algorithm is carried out in the following steps:

[0145] Step 2.1: Initialize the starting and target locations of the mission, the number of drones N, the locations of radar obstacles, the locations of missiles, and the number of path points K;

[0146] Step 2.2: Initialize the population size to NP, the maximum number of iterations to itermax, the dimension of individuals in the population to D = N × K, define the current number of iterations to t, and initialize t = 1;

[0147] Step 2.3: Move N drones from the starting position Fly to their respective target locations The path planning scheme is encoded as an individual, so that the i-th individual in the t-th generation population is obtained ,in, represents the i-th individual The position of the jth dimension, indicating the The first drone Waypoint locations; Indicates rounding up, and % indicates remainder.

[0148] Step 2.4: Calculate the i-th individual in the t-th generation population using formula (16) The fitness value of And find the top three wolves with the smallest fitness value from the t-th generation population and record them as the t-th generation population Wolf 、 Wolf and Wolf ;in, represents the population in the tth generation The j-th dimension position of the wolf individual, represents the population in the tth generation The j-th dimension position of the wolf individual, represents the population in the tth generation The j-th dimension position of the wolf individual.

[0149] Step 2.5: If the individuals in the t-th generation population belong to the top three levels of wolves, use formula (17) to update , and , get the updated Wolf individual , updated Wolf individual , updated Wolf individual ;

[0150] (17)

[0151] In formula (17), represents the gray wolf parameter of the tth iteration, and , and represents the j-th dimension position of the r1-th individual and the r2-th individual randomly selected from the t-th generation population, and ; Represents the first random number of the tth generation between 0 and 1.

[0152] If the i-th individual in the t-th generation population does not belong to the top three levels of wolves, then the updated i-th individual is calculated using formula (18): ; Thus we get the tth generation updated population ;

[0153] (18)

[0154] In formula (18), 、 and They are respectively the affected populations in the t generation and the Wolf individual, Wolf wolf individuals and The j-th dimension element of the i-th individual affected by the wolf individual, 、 and are the tth generation populations Wolf individuals, Wolf individuals and The weight of the individual wolf.

[0155] Step 2.6: Use formula (16) to calculate the updated i-th individual The fitness value of ,like < , then As the i-th individual in the intermediate population of generation t Otherwise, As the i-th individual in the intermediate population of generation t ; Thus we get the intermediate population of the tth generation.

[0156] Step 2.7: If t ≤ (2 / 3) × itermax, then divide the t-th generation renewal population according to the flight altitude of the individuals in the t-th generation intermediate population, and divide the first Δ individuals with the highest flight altitude into the t-th generation high-altitude sub-population, the last Δ individuals into the t-th generation low-altitude sub-population, and the middle Δ individuals into the t-th generation medium-altitude sub-population. Otherwise, divide the t-th generation renewal population according to the fitness values ​​of the individuals in the t-th generation intermediate population, and divide the first Δ individuals with the smallest fitness values ​​into the t-th generation good sub-population, the middle Δ individuals into the t-th generation medium sub-population, and the last Δ individuals into the t-th generation poor sub-population.

[0157] Step 2.8: When t≤(2 / 3)itermax, if the updated i-th individual belongs to the high-altitude subpopulation, then according to formula (19) we can get the ith mutant individual of the tth generation: ,in, The j-th dimension position of the i-th mutant individual in the t-generation population represented by ;

[0158] (19)

[0159] In formula (19), and are randomly selected from the hollow subpopulation of the t-th generation update population. Individuals and The j-th dimension position of each individual, and , is the first randomly selected from the high-altitude population of the t-th generation update population. The j-th dimension position of each individual; Represents the second random number of the tth generation between 0 and 1.

[0160] like Belongs to the low-altitude subpopulation, then according to formula (20) we can get the ith mutant individual of the tth generation ;

[0161] (20)

[0162] In formula (20), , ,and is the first population randomly selected from the empty subpopulation of the t-th generation update population. Individual, Individual, The j-th dimension position of each individual, ; represents the third random number of the tth generation between 0 and 1;

[0163] like Belongs to the hollow subpopulation, then according to formula (21) we can get the ith mutant individual of the tth generation :

[0164] (twenty one)

[0165] In formula (21), , ,and is the first randomly selected from the empty subpopulation of the t-th generation update population. Individual, Individual, The j-th dimension position of each individual, and is the first randomly selected population from the low-space subpopulation of the t-th generation renewal population. Individual, The j-th dimension position of each individual, ; Represents the fourth random number of generation t between 0 and 1.

[0166] When t>(2 / 3)itermax, if the updated i-th individual Belongs to a better subpopulation, then according to formula (22) we can get the ith mutant individual of the tth generation ;

[0167] (twenty two)

[0168] In formula (22), is the j-th dimension position of the individual with the best fitness value in the t-th generation updated population, and are randomly selected from the better subpopulation of the t-th generation update population. Individual, The j-th dimension position of each individual, , F is the mutation operator;

[0169] like Belongs to the medium subpopulation, then according to formula (23) we can get the ith mutant individual of the tth generation ;

[0170] (twenty three)

[0171] In formula (23), Indicates the j-th dimension position of the individual with the best fitness value in the medium subpopulation of the t-th generation updated population, and are randomly selected from the medium subpopulation of the t-th generation update population. Individual, The j-th dimension position of each individual, .

[0172] like Belonging to the inferior subpopulation, according to formula (24), the ith mutant individual of the tth generation is obtained :

[0173] (twenty four)

[0174] In formula (24), and are randomly selected from the inferior subpopulation of the t-th generation update population. Individual, The j-th dimension position of each individual, represents the first randomly selected from the tth generation update population The j-th dimension position of each individual, .

[0175] Step 2.9: Generate the i-th crossover individual of the t-th generation using formula (26) :

[0176] (25)

[0177] In formula (25), is the crossover operator, is a random integer between [1,D], represents the fifth random number of the tth generation between 0 and 1, yes The j-th dimension position in .

[0178] Step 2.10: Calculate the i-th cross-individual vector according to formula (16) The fitness value of , and thus use formula (26) to obtain the i-th individual in the t+1 generation population:

[0179] (26)

[0180] Step 2.11: Assign t+1 to t. If t>itermax, it means that the itermax-th generation population is obtained, and the individual with the smallest fitness value is selected from it as the multi-UAV collaborative path planning solution. Otherwise, return to step 2.4 and execute sequentially.

[0181] In this embodiment, an electronic device includes a memory and a processor, wherein the memory is used to store a program that supports the processor to execute the above method, and the processor is configured to execute the program stored in the memory.

[0182] In this embodiment, a computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above method are executed.

Claims

1. A multi-UAV collaborative path planning method based on an adaptive multi-population hybrid algorithm, characterized in that: It is applied to N drones flying from a known starting position. Fly to their respective target locations In a three-dimensional flight scene, obstacles with known positions are set in the three-dimensional flight scene, wherein, Indicates the starting position of the nth drone, Represents the target point position of the nth UAV. The multi-UAV collaborative path planning method is performed according to the following steps: Step 1: Construct the objective function of the multi-UAV collaborative path planning model; Step 1.1: Let the flight path of the nth drone be ,in, is the number of trajectory points, represents the kth trajectory point of the nth UAV, and The coordinates of , and use formula (1) to establish the flight length model of the nth UAV : (1) In formula (1), Indicates that the nth drone starts from Track points To Track Points Flight distance; Step 1.2: Use equations (2) and (3) to construct the flight altitude model of the nth UAV : (2) (3) In formula (2) and formula (3), is the nth drone at k trajectory point The flight altitude model, is the minimum altitude at which the drone can fly. is the maximum altitude at which the drone can fly. Indicates that the nth drone is in the Track points The flight altitude, Indicates that the nth drone is in the -1 track point Flight altitude; Step 1.3: Construct the maneuvering physical constraint model of the nth UAV, including the deflection angle constraint model of the nth UAV and the climb angle constraint model of the nth UAV ; Step 1.4: Build the obstacle threat model for the nth UAV, including the radar threat model for the nth UAV and the missile threat model of the nth drone ; Step 1.5: Construct the coordination model of the nth UAV, including: time coordination model and spatial collaboration models ; Step 1.6: Use Equation (16) to construct the objective function of the multi-UAV collaborative path planning model : (16) In formula (16), , , , , are 5 weight factors; Step 2: Use the adaptive multi-population hybrid algorithm to solve the multi-UAV collaborative path model and obtain the multi-UAV path planning solution with the lowest path planning cost.

2. The multi-UAV collaborative path planning method based on the adaptive multi-population hybrid algorithm according to claim 1 is characterized in that: The step 1.3 includes: Step 1.3.1: Use equations (4) and (5) to construct the deflection angle model and climb angle model of the nth UAV respectively: (4) (5) In formula (4) and formula (5), Indicates that the nth drone is at the kth trajectory point The deflection angle, Indicates that the nth drone is at the kth trajectory point The climbing angle, Represents the k+1th trajectory point of the nth UAV coordinates of Step 1.3.2: Use equations (6) and (7) to construct the deflection angle constraint model of the nth UAV ; (6) (7) In formula (6) and formula (7), is the nth drone at the kth trajectory point The deflection angle constraint, is the deflection angle of the nth UAV at the kth trajectory point, is the maximum deflection angle of the UAV; Step 1.3.3: Use equations (8) and (9) to construct the climb angle constraint model of the nth UAV : (8) (9) In formula (8) and formula (9), is the nth drone at the kth trajectory point The climb angle constraint, is the climbing angle of the nth UAV at the kth trajectory point, is the maximum climb angle of the UAV.

3. The multi-UAV collaborative path planning method based on the adaptive multi-population hybrid algorithm according to claim 2 is characterized in that: The step 1.4 includes: Step 1.4.1: Use equation (10) to construct the radar model: (10) In formula (10), Represents any three-dimensional coordinate in a three-dimensional flight scene, It refers to the area that the radar can cover. is the location of the radar area center, is the radius of the radar; Step 1.4.2: Use Equation (11) to construct the radar threat model of the nth UAV : (11) In formula (11), Indicates the nth drone Track points Hedi Track points The distance between Indicates the nth drone Track points With the Track points The trajectory segment Two intersection points with the radar area and The distance between Indicates the nth drone Track points Second intersection with the radar area The distance between Indicates the nth drone Track points First intersection with the radar area The distance between Represents a trajectory segment The distance from the midpoint of to the center of the radar area; Step 1.4.3: Use Equation (12) to construct the missile model : (12) In formula (12), It is the area that the missile can cover. is the coordinate of the center of the bottom surface of the missile, Indicates the radius of the missile's bottom surface, Indicates the missile's altitude; Step 1.4.4: Use equation (13) to construct a missile threat model: (13) In formula (13), Indicates the nth drone Track points Hedi Track points The distance between Indicates the nth drone Track points Hedi Track points The trajectory segment Two intersection points with the missile area and The distance between Indicates the nth drone Track points Second intersection with the missile area The distance between Indicates the nth drone Track points First intersection with the missile area The distance between express The distance from the midpoint of the missile to the center of the missile bottom circle.

4. The multi-UAV collaborative path planning method based on the adaptive multi-population hybrid algorithm according to claim 3 is characterized in that: The step 1.5 includes: Step 1.5.1: Use Equation (14) to construct the time coordination model of the nth UAV : (14) In formula (14), is the flight time of the nth drone, Indicates the time it takes for the first drone to reach the corresponding target point, Indicates the buffer time for the remaining drones after the first one completes its flight mission; Step 1.5.2: Use Equation (15) to construct the spatial coordination model of the nth UAV : (15) In formula (15), Indicates the The kth path point of the UAV With the The first drone Waypoints distance, is the penalty coefficient for drones violating the space safety distance, The safe distance between drones.

5. The multi-UAV collaborative path planning method based on the adaptive multi-population hybrid algorithm according to claim 4 is characterized in that: The step 2 includes: Step 2.1: Initialize the starting and target locations of the mission, the number of drones N, the locations of radar obstacles, the locations of missiles, and the number of path points K; Step 2.2: Initialize the population size to NP, the maximum number of iterations to itermax, the dimension of individuals in the population to D = N × K, define the current number of iterations to t, and initialize t = 1; Step 2.3: Move N drones from the starting position Fly to their respective target locations The path planning scheme is encoded as an individual, so that the i-th individual in the t-th generation population is obtained ,in, represents the i-th individual The position of the jth dimension indicates the The first drone Waypoint locations; Indicates rounding up, % indicates remainder; Step 2.4: Calculate the i-th individual in the t-th generation population using formula (16) The fitness value of And find the top three wolves with the smallest fitness value from the t-th generation population and record them as the t-th generation population Wolf 、 Wolf and Wolf ;in, represents the population in the tth generation The j-th dimension position of the wolf individual, represents the population in the tth generation The j-th dimension position of the wolf individual, represents the population in the tth generation The j-th dimension position of the wolf individual; Step 2.5: If the individuals in the t-th generation population belong to the top three levels of wolves, use formula (17) to update , and , get the updated Wolf individual , updated Wolf individual , updated Wolf individual ; (17) In formula (17), represents the gray wolf parameter of the tth iteration, and , and represents the j-th dimension position of the r1-th individual and the r2-th individual randomly selected from the t-th generation population, and ; represents the first random number of the tth generation between 0 and 1; If the i-th individual in the t-th generation population does not belong to the top three levels of wolves, then the updated i-th individual is calculated using formula (18): ; Thus we get the tth generation updated population ; (18) In formula (18), 、 and They are respectively the affected populations in the t generation and the Wolf individual, Wolf wolf individuals and The j-th dimension element of the i-th individual affected by the wolf individual, 、 and are the tth generation populations Wolf individuals, Wolf individuals and The weight of the individual wolf; Step 2.6: Use formula (16) to calculate the updated i-th individual The fitness value of ,like < , then As the i-th individual in the intermediate population of generation t Otherwise, As the i-th individual in the intermediate population of generation t ; Thus we get the intermediate population of generation t; Step 2.7: If t ≤ (2 / 3) × itermax, then divide the t-th generation renewal population according to the flight altitude of the individuals in the t-th generation intermediate population, and divide the first Δ individuals with the highest flight altitude into the t-th generation high-altitude sub-population, the last Δ individuals into the t-th generation low-altitude sub-population, and the middle Δ individuals into the t-th generation medium-altitude sub-population. Otherwise, divide the t-th generation renewal population according to the fitness values ​​of the individuals in the t-th generation intermediate population, and divide the first Δ individuals with the smallest fitness values ​​into the t-th generation good sub-population, the middle Δ individuals into the t-th generation medium sub-population, and the last Δ individuals into the t-th generation poor sub-population. Step 2.8: When t≤(2 / 3)itermax, if the updated i-th individual belongs to the high-altitude subpopulation, then according to formula (19) we can get the ith mutant individual of the tth generation: ,in, The j-th dimension position of the i-th mutant individual in the t-generation population represented by ; (19) In formula (19), and are randomly selected from the hollow subpopulation of the t-th generation update population. Individuals and The j-th dimension position of each individual, and , is the first randomly selected from the high-altitude population of the t-th generation update population. The j-th dimension position of each individual; represents the second random number of the tth generation between 0 and 1; like Belongs to the low-altitude subpopulation, then according to formula (20) we can get the ith mutant individual of the tth generation ; (20) In formula (20), , ,and is the first population randomly selected from the empty subpopulation of the t-th generation update population. Individual, Individual, The j-th dimension position of each individual, ; represents the third random number of the tth generation between 0 and 1; like Belongs to the hollow subpopulation, then according to formula (21) we can get the ith mutant individual of the tth generation : (21) In formula (21), , ,and is the first randomly selected from the empty subpopulation of the t-th generation update population. Individual, Individual, The j-th dimension position of each individual, and is the first randomly selected population from the low-space subpopulation of the t-th generation renewal population. Individual, The j-th dimension position of each individual, ; represents the fourth random number of the tth generation between 0 and 1; When t>(2 / 3)itermax, if the updated i-th individual Belongs to a better subpopulation, then according to formula (22) we can get the ith mutant individual of the tth generation ; (22) In formula (22), is the j-th dimension position of the individual with the best fitness value in the t-th generation updated population, and are randomly selected from the better subpopulation of the t-th generation update population. Individual, The j-th dimension position of each individual, , F is the mutation operator; like Belongs to the medium subpopulation, then according to formula (23) we can get the ith mutant individual of the tth generation ; (23) In formula (23), Indicates the j-th dimension position of the individual with the best fitness value in the medium subpopulation of the t-th generation updated population, and are randomly selected from the medium subpopulation of the t-th generation update population. Individual, The j-th dimension position of each individual, ; like Belonging to the inferior subpopulation, according to formula (24), the ith mutant individual of the tth generation is obtained : (24) In formula (24), and are randomly selected from the inferior subpopulation of the t-th generation update population. Individual, The j-th dimension position of each individual, represents the first randomly selected from the tth generation update population The j-th dimension position of each individual, ; Step 2.9: Generate the i-th crossover individual of the t-th generation using formula (26) : (25) In formula (25), is the crossover operator, is a random integer between [1,D], represents the fifth random number of the tth generation between 0 and 1, yes The j-th dimension position in ; Step 2.10: Calculate the i-th cross-individual vector according to formula (16) The fitness value of , and thus use formula (26) to get the i-th individual in the t+1 generation population : (26) Step 2.11: Assign t+1 to t. If t>itermax, it means that the itermax-th generation population is obtained, and the individual with the smallest fitness value is selected from it as the multi-UAV collaborative path planning solution. Otherwise, return to step 2.4 and execute sequentially.

6. An electronic device comprising a memory and a processor, characterized in that: The memory is used to store a program that supports the processor to execute the multi-UAV collaborative path planning method described in any one of claims 1-5, and the processor is configured to execute the program stored in the memory.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the multi-UAV collaborative path planning method according to any one of claims 1 to 5 are executed.

Images