Unmanned aerial vehicle dynamic path planning method considering marine environment change

Abstract

The invention discloses an unmanned aerial vehicle dynamic path planning method considering marine environment change, and belongs to the field of path planning. According to the method, the severe convection intensity, the storm moving track and the ship position are updated in real time through satellite communication, a replanning starting point is positioned in a millisecond level, a diversified new path population is generated, and a safe and feasible path is rapidly output through limited evolution; and the path danger cost is further dynamically calculated, so that the unmanned aerial vehicle actively bypasses the sudden high-risk area. Compared with the prior art, the method has the advantages that the problem of path failure caused by sudden environment change in a maritime search and rescue scene is effectively solved, and the task safety and efficiency can be remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of path planning, more particularly to a dynamic path planning method for unmanned aerial vehicles considering changes in marine environment. BACKGROUND

[0002] In recent years, unmanned aerial vehicles (UAVs) have been increasingly applied in marine search and rescue missions, and their path planning technology has become a research hotspot. In the prior art, some scholars have attempted to incorporate adverse weather factors (such as strong storms) into path planning models and achieve knowledge transfer through multi-task optimization to improve planning quality.

[0003] However, current mainstream methods are mostly for path planning in static environments, i.e., offline planning for fixed meteorological conditions and obstacle distribution, and cannot respond to dynamic changes in marine environments in real time, which leads to a real-time increase in the flight risk coefficient of UAVs as the environment changes.

[0004] Therefore, there is an urgent need for a planning scheme that can dynamically and quickly generate safe and efficient paths in real time to perceive changes in the environment, in order to cope with the high uncertainty of marine search and rescue scenarios. SUMMARY

[0005] In view of the above, the present application provides a dynamic path planning method for unmanned aerial vehicles considering changes in marine environment, to improve the safety and real-time performance of flight path planning for unmanned aerial vehicles in dynamic environments.

[0006] To achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0007] A dynamic path planning method for unmanned aerial vehicles considering changes in marine environment, comprising,

[0008] S1. Generating an initial population of waypoints based on a starting point and a target point;

[0009] S2. Obtaining an initial path set in a static environment through weight vector search;

[0010] S3. Performing multi-objective optimization based on the initial path set to obtain an optimal path;

[0011] S4. The unmanned aerial vehicle flies based on the optimal path and monitors changes in the navigation environment in real time, and when a change in the navigation environment is detected, step S5 is executed;

[0012] S5. Calculating the next waypoint based on the current position of the unmanned aerial vehicle, taking the next waypoint as a re-planning starting point, and repeating steps S1-S3.

[0013] Preferably, step S1 comprises:

[0014] recursively dividing the path formed by the starting point and the target point;

[0015] At the midpoint of the line connecting adjacent waypoints, a new waypoint is generated by applying a random offset along the vertical direction; the offset satisfies the following formula:

[0016]

[0017] In the formula, r d For [0,D max A uniformly distributed random number within a given range, D max Let θ be the maximum offset on the vertical median between p1 and p2, where p1 and p2 are the coordinates of two known waypoints, θ is the angle between the vertical median and the coordinate axes, and a and b are unit vectors.

[0018] Preferably, step S2 includes: dividing the target space into multiple sub-regions, generating feasible weight vectors and infeasible weight vectors respectively, and allocating the feasible weight vectors and infeasible weight vectors to each sub-region based on the angle between them and the center vector of the sub-region;

[0019] The feasible weight vector is uniformly distributed within the feasible region of the sub-region and is used to perform objective optimization under strict constraints.

[0020] The infeasibility weight vector is distributed outside the boundary of the infeasibility domain of the sub-region, and is used to retain potential solutions by adopting a lenient penalty strategy to enhance the global search capability.

[0021] Preferably, generating a feasible weight vector includes:

[0022] Uniform sampling is performed in the [0,π / 2] radian space to generate m radians, and the m radians are mapped to an m+1 dimensional weight vector;

[0023] The weight vectors are filtered using the Max-Min filtering method.

[0024] Preferably, mapping the m radians to an m+1 dimensional weight vector includes:

[0025]

[0026] θ m Let m represent the m-th radian, and ∈ be a small positive value.

[0027] Preferably, generating an infeasible weight vector includes generating a weight vector in radian space and dynamically updating it in the following manner to guide the search direction;

[0028]

[0029] In the formula, rand is a random number uniformly distributed in (0, 1), gen and maxgen are the current iteration number and the maximum iteration number, respectively, and λ is a control variable.

[0030] Preferably, in step S3, the optimization objectives in the multi-objective optimization include minimizing the total flight distance and minimizing the dynamic hazard cost; the location intensity of the strong convection area and the radius of influence of the strong convection; and the coordinate set of the sailing vessel and the radius of the safety exclusion zone.

[0031] The minimum total flight distance is:

[0032]

[0033] In the formula, P i+1 P i These are the i-th and i+1-th waypoints, respectively, and M is the total number of waypoints;

[0034] The minimum dynamic hazard cost is:

[0035] R=λ1R ship +λ2R storm

[0036] In the formula, R ship For the dangerous costs of drone-ship collisions, R storm λ1 represents the total cost of dynamic flight risks caused by strong convection on the sea surface for the UAV; λ2 represents the weight of the risk cost of collision between the UAV and the ship; λ2 represents the weight of the total cost of dynamic flight risks caused by strong convection on the sea surface for the UAV. The higher λ2 is, the more attention should be paid to the control of dynamic risks such as strong convection in this mission scenario, and λ1+λ2=1.

[0037]

[0038] d ij For drones and ships j The Euclidean distance between them, G i For the set of threatened waypoints, r ship The radius of the circular safety exclusion zone;

[0039]

[0040] R i (t) represents the expected value of the danger encountered by the drone when passing waypoint i, R i (t)=N·c i (t), where N is the scaling factor, and c i (t) represents the probability that the UAV will lose control or crash when it crosses the strong convection at waypoint i at time t. s(x i ,y i (t) represents the waypoint P at time t. i (x i ,y iThe intensity of strong convection on ) and r storm The radius of influence of strong convection.

[0041] Preferably, in step S3, the multi-objective optimization further includes path constraints, which include:

[0042] Turning angle constraint: Turning angle ψ at waypoint i i ≤ψ max The number of violations is Δψ i =max(ψ i -ψ max ,0);

[0043] Segment spacing constraint: distance d between adjacent waypoints i ∈[d min ,d max ], d min ,d max These are the minimum and maximum constraint values ​​for the waypoint spacing, respectively, and the number of violations is calculated using a piecewise function.

[0044] Preferably, in step S4, changes in the navigation environment are monitored in real time, including relying on satellite communication to update the following risk data in real time: the location and intensity of strong convection areas and the radius of influence of strong convection, as well as the coordinate set of the navigation vessels and the radius of the safe no-go zone.

[0045] Preferably, in step S5, calculating the next waypoint based on the current position of the UAV includes:

[0046] Calculate the theoretical next waypoint of the UAV after t seconds of flight based on the original path;

[0047] Candidate points that satisfy safety and route constraints are selected within the neighborhood of the next waypoint in the theory.

[0048] The candidate point with the optimal objective function is selected as the actual next waypoint.

[0049] As can be seen from the above technical solutions, the present invention discloses a dynamic path planning method for unmanned aerial vehicles (UAVs) that takes into account changes in the marine environment. It updates the intensity of strong convection, the movement trajectory of storms and the position of ships in real time through satellite communication, and locates the replanning starting point within milliseconds. It generates a diverse population of new paths and outputs safe and feasible paths through finite evolution. Furthermore, it dynamically calculates the risk cost of the path, thereby enabling the UAV to actively detour around sudden high-risk areas.

[0050] Compared with existing technologies, this invention effectively solves the problem of path failure caused by sudden environmental changes in maritime search and rescue scenarios, thereby significantly improving mission safety and efficiency.

[0051] Furthermore, the feasible / infeasible weighting mechanism and adaptive search strategy in this application ensure that local optima are avoided while circumventing sudden storms and ship restricted areas, thereby improving the success rate of UAV search and rescue in typhoon-affected sea areas. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0053] Figure 1 Flowchart of the UAV dynamic path planning method considering changes in the marine environment according to the present invention;

[0054] Figure 2 This is a flowchart illustrating the UAV dynamic path planning method for considering changes in the marine environment, as described in this invention. Detailed Implementation

[0055] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0056] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0057] This invention proposes a novel dynamic constraint multi-objective optimization model and a response dynamic constraint multi-objective evolutionary algorithm, which can better adapt to real-time and complex environmental changes at sea.

[0058] In one embodiment, the UAV dynamic path planning method provided in this application, such as Figure 1 The steps include,

[0059] S1. Generate an initial waypoint population based on the starting point and the target point;

[0060] S2. Obtain the initial path set in the static environment through weight vector search;

[0061] S3. Perform multi-objective optimization based on the initial path set to obtain the optimal path;

[0062] S4. The UAV flies based on the optimal path and monitors changes in the navigation environment in real time. When a change in the navigation environment is detected, step S5 is executed.

[0063] S5. Calculate the next waypoint based on the current location of the UAV, and use the next waypoint as the starting point for replanning. Repeat steps S1-S3.

[0064] The UAV path planning method of this application includes two key stages: the preparation stage of static path planning and the flight stage of dynamic path replanning.

[0065] For static path planning, traditional random initialization of the population struggles to quickly generate feasible solutions after environmental changes, often leading to suboptimal or infeasible paths in dynamic path planning. To address this issue, this application first proposes a novel population generation mechanism focused on constructing a high-quality initial population for the algorithm to optimize planning performance.

[0066] In some implementation schemes, static path planning includes:

[0067] Step S1: First, generate an initial waypoint population based on the starting point and the target point; including:

[0068] Given a starting point (start) and a target point (end), recursively split the path formed by the starting point and the target point;

[0069] A new waypoint is generated by applying a random offset along the vertical direction at the midpoint of the line connecting adjacent waypoints; in this application, the offset satisfies the following formula:

[0070]

[0071] In the formula, r d For [0,D max A uniformly distributed random number within a given range, D max Let θ be the maximum offset on the vertical median between p1 and p2, where p1 and p2 are the coordinates of two known waypoints, θ is the angle between the vertical median and the y-axis, and a and b are unit vectors parallel to the y-axis and x-axis, respectively.

[0072] In one embodiment, step S2 involves obtaining an initial path set in a static environment through weight vector search.

[0073] In this embodiment, the target space is divided into multiple sub-regions to decompose the complex multi-objective problem into several sub-problems.

[0074] Furthermore, N1 feasible weight vectors and N2 infeasible weight vectors are generated respectively, and these vectors are assigned to each sub-region based on the angle between them and the center vector of the sub-region, forming a subpopulation. The center vector of each sub-region is the bisector of that sub-region. For example, if a subspace exists with an radian range of [0, π / 2], then v = π / 4 is its center vector. If the angle between a weight vector and the center vector is less than a specific angle, it is assigned to the sub-region corresponding to that center vector.

[0075] In this application, the feasible weight vector is uniformly distributed within the feasible region of the sub-region and is used to perform objective optimization under strict constraints, focusing on optimizing objectives such as path distance and safety, thereby generating diverse Pareto optimal solutions;

[0076] To generate feasible weight vectors for feasible regions, in one embodiment, m radians θ = {θ1, θ2, ..., θ} are uniformly generated in a two-dimensional space. m}, where θ∈[0,π / 2], and θ m =0, and then mapped to an m+1 dimensional weight vector according to the following formula;

[0077]

[0078] θ m This represents the m-th radian, where ∈ is a small positive value, preferably ∈ = 10. -10 .

[0079] Furthermore, the Max-Min screening method is used to select a uniformly distributed feasible weight vector from the m weights.

[0080] The infeasible weight vector is distributed outside the boundary of the infeasible region of the sub-region. It is used to retain potential solutions near the boundary of the feasible region by adopting a relaxed penalty strategy to significantly enhance the global search capability and prevent the feasible weight vector from getting trapped in local optima.

[0081] For generating infeasible weight vectors for infeasible regions, this application employs a generation method similar to that used for feasible weight vectors, but without θ. m =0 is not a constraint, and mapping according to the above formula is not required.

[0082] Meanwhile, the lenient penalty strategy refers to a strategy for maintaining infeasible solutions, aiming to preserve potentially infeasible solutions. During evolution, this is achieved by adjusting the radian θ. m Dynamically update weights to guide search direction:

[0083]

[0084] In the formula, rand is a random number uniformly distributed in (0, 1), gen and maxgen are the current iteration number and the maximum iteration number, respectively, and λ is a control variable, which is generally taken as an integer in [2, 5].

[0085] In one embodiment, S3, multi-objective optimization is performed based on the initial path set to obtain the optimal path; in this application, the total distance of the UAV flight and the safety of the UAV during flight are taken as optimization objectives, and a path update based on real-time strong convection and ship movement data is proposed, so that the UAV can adjust its flight path in time to reduce the risk of crashing due to strong winds or collisions with ships.

[0086] Specifically, in modeling dynamic flight problems, the sea surface environment is discretized into two-dimensional terrain mesh elements r. cell The intensity of strong convection within the grid is updated in real time; and the coordinate set of the sailing vessels is defined as C = c. j (j=1,2,…,n), with ship coordinates c j =(x j ,y j Centered on a radius R ship A circular safety restricted area is defined, and the set of threatened waypoints G is dynamically marked. i .

[0087] The optimization objectives of this modeling include minimizing the total flight distance and minimizing the dynamic hazard cost;

[0088] The total flight distance, i.e., the sum of the Euclidean distances between waypoints, is:

[0089]

[0090] In the formula, P i+1 P i These are the i-th and i+1-th waypoints, respectively, and M is the total number of waypoints;

[0091] The dynamic hazard cost is:

[0092] R=λ1R ship +λ2R storm

[0093] In the formula, λ1 represents the weight of the risk cost of a collision between the UAV and a ship, and λ2 represents the weight of the total dynamic flight risk cost of the UAV caused by strong convection on the sea surface. The higher λ2 is, the more attention should be paid to the management of dynamic risks such as strong convection in this mission scenario, and λ1 + λ2 = 1, R ship The dangerous costs of drone-ship collisions,

[0094]

[0095] d ijFor drones and ships j The Euclidean distance between them

[0096] R storm The total dynamic flight risk cost of the UAV caused by strong convection over the sea is related to time t; each grid cell represents the risk of the UAV encountering strong convection, and the formula for calculating the total risk cost is:

[0097]

[0098] R i (t) represents the expected value of the danger encountered by the UAV when passing through waypoint i.

[0099] R i (t)=N·c i (t)

[0100] N is the scaling factor, c i (t) represents the probability of the UAV losing control or crashing when it crosses strong convection at waypoint i at time t. The calculation formula is:

[0101]

[0102] s(x i ,y i (t) represents the waypoint P at time t. i (x i ,y i The intensity of strong convection on ) and r storm The radius of influence of strong convection.

[0103] Furthermore, this application considers constraints such as the turning angle limit and waypoint spacing limit of the UAV, thereby enabling the UAV to calculate a feasible and safe flight path plan, wherein,

[0104] Turning angle constraint: Turning angle ψ at waypoint i i ≤ψ max The number of violations is Δψ i =max(ψ i -ψ max ,0); In one specific embodiment, the turning angle of the UAV flight path is limited to:

[0105]

[0106] Among them, s 1,i s1 represents the constraint violation of the maximum turning angle limit when the UAV passes through the i-th waypoint, and s1 is the sum of constraint violations of the turning angle at all waypoints, meaning that all waypoints should have no constraint violations; Δψ i =ψ i -ψ max , Δψi For the number of violations, ψ i For the turning angle constraint of the UAV at waypoint i, ψ max Maximum turning angle limit;

[0107] Segment spacing constraint: distance d between adjacent waypoints i ∈[d min ,d max ], d min ,d max These are the minimum and maximum constraint values ​​for the waypoint spacing, respectively, and the number of violations is calculated using a piecewise function.

[0108] In one exemplary embodiment, the minimum constraint value is set to 1 and the maximum constraint value to 8. The constraints are then:

[0109]

[0110] d i =||P i+1 -P i ||

[0111] Where s 2,i s1 represents the constraint violation of the maximum waypoint spacing limit on the current path of the UAV, and s2 is the sum of constraint violations of the segment spacing, that is, all waypoints should have no constraint violations.

[0112] In one or more implementation schemes, the flight phase for dynamic path replanning includes:

[0113] 1) Environmental change handling: The UAV faces several environmental changes during flight; in each iteration, the algorithm selects the optimal path from the set of flight paths based on weights.

[0114] 2) Replanning trigger mechanism: If environmental changes are detected, the algorithm calculates the next waypoint and uses it as the starting point to reinitialize the waypoint population.

[0115] In one specific embodiment, the specific execution steps are as follows:

[0116] S4. The UAV flies based on the optimal path and monitors changes in the navigation environment in real time. When a change in the navigation environment is detected, step S5 is executed.

[0117] In this embodiment, the UAV flies from its base to a designated search and rescue area. During the flight, the UAV updates maritime risk data through communication with observation satellites and performs path replanning to obtain a safer flight path. The maritime risk data includes, via satellite communication, real-time updates of the location and intensity of strong convection areas and the radius of influence of strong convection, as well as the coordinate set of ships and the radius of safe no-go zones.

[0118] S5. Calculate the next waypoint based on the current location of the UAV, and use the next waypoint as the starting point for replanning. Repeat steps S1-S3.

[0119] The next waypoint refers to the waypoint that the UAV should reach t seconds after starting flight from the current waypoint. First, the next waypoint is calculated using data obtained during the preparation phase when the environment remains unchanged. Then, a set of possible waypoints near the original waypoint is selected and calculated using the objective function and constraint functions. The waypoint that satisfies both the objective function and the constraints is set as the new next waypoint.

[0120] Further, following steps S1-S3, evolutionary operations are performed within preset finite evolutionary constraints, and the set of non-dominated paths is updated in real time to ensure that each change generates an adaptive flight trajectory.

[0121] In one embodiment, the overall path planning method flow of this application refers to Figure 2 ;

[0122] First, an initial population pop, index start, and end are generated. Evolutionary calculations are performed within max_FES evaluation times to obtain the set of non-dominated paths, paths.

[0123] Determine whether k iterations have been performed. If so, end the path planning. Otherwise, select the optimal path pathi from the path set paths based on the weights w1 and w2 (optimization target weights) and fly on path pathi.

[0124] Further determine whether environmental changes are detected. If not, continue flying on path pathi. Otherwise, calculate the next waypoint pnext of the UAV after flying for t seconds, and regenerate the population pop with pnext as the flight starting point.

[0125] Then, within max_FES evaluation times, the evolution calculation is re-performed, the path set paths is updated, and the process returns to the second step to continue determining whether k iterations have been performed.

[0126] Through data comparison from 30 repeated experiments, this application demonstrates that the path planning method provided exhibits significant effectiveness. Specifically, its HV index (Hypervolume, the larger the better) is up to 4.1 times higher than the baseline method, fully demonstrating its significant advantages in the convergence efficiency and diversity of solution sets.

[0127] Further ablation experiments confirmed that its population generation mechanism is the core component of the technology—if this mechanism is removed, the algorithm will be completely unable to generate feasible solutions, which highlights its key role in quickly obtaining feasible solutions in dynamic environments.

[0128] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0129] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A dynamic path planning method for unmanned aerial vehicles (UAVs) that considers changes in the marine environment, characterized in that, include: S1. Generate an initial waypoint population based on the starting point and the target point; S2. Obtain the initial path set in the static environment through weight vector search; S3. Perform multi-objective optimization based on the initial path set to obtain the optimal path; S4. The UAV flies based on the optimal path and monitors changes in the navigation environment in real time. When a change in the navigation environment is detected, step S5 is executed. S5. Calculate the next waypoint based on the current location of the UAV, and use the next waypoint as the starting point for replanning. Repeat steps S1-S3.

2. The UAV dynamic path planning method according to claim 1, characterized in that, Step S1 includes: Recursively divide the path formed by the starting point and the target point; At the midpoint of the line connecting adjacent waypoints, a new waypoint is generated by applying a random offset along the vertical direction; the random offset satisfies the following formula: In the formula, r d For [0,D max A uniformly distributed random number within a given range, D max Let θ be the maximum offset on the vertical median between p1 and p2, where p1 and p2 are the coordinates of two known waypoints, θ is the angle between the vertical median and the coordinate axes, and a and b are unit vectors.

3. The UAV dynamic path planning method according to claim 1, characterized in that, Step S2 includes: dividing the target space into multiple sub-regions, generating feasible weight vectors and infeasible weight vectors respectively, and allocating the feasible weight vectors and infeasible weight vectors to each sub-region based on the angle between them and the center vector of the sub-region; The feasible weight vector is uniformly distributed within the feasible region of the sub-region and is used to perform objective optimization under strict constraints. The infeasibility weight vector is distributed outside the boundary of the infeasibility domain of the sub-region, and is used to retain potential solutions by adopting a lenient penalty strategy to enhance the global search capability.

4. The UAV dynamic path planning method according to claim 3, characterized in that, Generating feasible weight vectors includes: Uniform sampling is performed in the [0,π / 2] radian space to generate m radians, and the m radians are mapped to an m+1 dimensional weight vector; The weight vectors are filtered using the Max-Min filtering method.

5. The UAV dynamic path planning method according to claim 4, characterized in that, Mapping the m radians to an m+1 dimensional weight vector includes: θ m Let m represent the m-th radian, and ∈ be a small positive value.

6. The UAV dynamic path planning method according to claim 3, characterized in that, Generate infeasible weight vectors, including generating weight vectors in radian space, and dynamically updating them as follows to guide the search direction; In the formula, rand is a random number uniformly distributed in (0, 1), gen and maxgen are the current iteration number and the maximum iteration number, respectively, and λ is a control variable.

7. The UAV dynamic path planning method according to claim 1, characterized in that, In step S3, the optimization objectives in the multi-objective optimization include minimizing the total flight distance and minimizing the dynamic hazard cost; The total flight distance is: In the formula, P i+1 P i These are the i-th and i+1-th waypoints, respectively, and M is the total number of waypoints; The dynamic hazard cost is: R=λ1R ship +λ2R storm In the formula, R ship For the dangerous costs of drone-ship collisions, R storm The total cost of dynamic flight risks caused by strong convection on the sea surface for unmanned aerial vehicles; λ1 represents the weight of the hazardous cost of a collision between the drone and a ship, and λ2 represents the weight of the total dynamic flight risk cost of the drone caused by strong convection on the sea surface. d ij For drones and ships j The Euclidean distance between them, G i For the set of threatened waypoints, r ship The radius of the circular safety exclusion zone. R i (t) represents the expected value of the danger encountered by the drone when passing waypoint i, R i (t)=N·c i (t), where N is the scaling factor, and c i (t) represents the probability that the UAV will lose control or crash when it crosses the strong convection at waypoint i at time t. s(x i ,y i (t) represents the waypoint P at time t. i (x i ,y i The intensity of strong convection on ) and r storm The radius of influence of strong convection.

8. The UAV dynamic path planning method according to claim 1, characterized in that, In step S3, the multi-objective optimization further includes path constraints, which include: Turning angle constraint: Turning angle ψ at waypoint i i ≤ψ max The number of violations is Δψ i =max(ψ i -ψ max ,0); Segment spacing constraint: distance d between adjacent waypoints i ∈[d min ,d max ], d min ,d max These are the minimum and maximum constraint values ​​for the waypoint spacing, respectively, and the number of violations is calculated using a piecewise function.

9. The UAV dynamic path planning method according to claim 1, characterized in that, In step S4, real-time monitoring of changes in the navigation environment, including reliance on satellite communication, is conducted, and the following risk data is updated in real time: Location and intensity of strong convection areas and radius of influence of strong convection, as well as coordinate set of ships and radius of safe no-go zones.

10. The UAV dynamic path planning method according to claim 1, characterized in that, In step S5, the next waypoint is calculated based on the current position of the UAV, including: Calculate the theoretical next waypoint of the UAV after t seconds of flight based on the original path; Candidate points that satisfy safety and route constraints are selected within the neighborhood of the next waypoint in the theory. The candidate point with the optimal objective function is selected as the actual next waypoint.

Images