Unmanned surface vessel path planning methods, devices, systems, and storage media

Abstract

This invention discloses an unmanned surface vessel (USV) path planning method, apparatus, system, and storage medium. It constructs a grid map of marine obstacles and integrates adaptive functions based on shortest path, obstacle penalty, and smoothness. An initial population is generated through a Tent-Logistic-Cosine chaotic mapping, and initial fitness values ​​are calculated. A hybrid evolutionary strategy is incorporated to maintain population diversity during each mutation, preventing premature convergence. A precise hunting strategy is employed to enhance the algorithm's global search capability. A speed optimization strategy is combined to improve exploration efficiency during the dive phase. Using this invention, efficient path planning can be achieved in complex marine simulation environments, solving the problem of poor performance in path planning by the Red-tailed Eagle optimization algorithm.

Description

Technical Field

[0001] This invention belongs to the field of path planning technology, and particularly relates to an unmanned surface vessel (USV) path planning method, device, system, and storage medium. Background Technology

[0002] Unmanned surface vessel (USV) path planning algorithms are techniques that combine factors such as avoiding reefs, finding the shortest path, and optimizing path smoothness to plan the most suitable route. Effective path planning algorithms not only ensure the safety and efficiency of USVs during missions but also significantly improve their adaptability to the complex and ever-changing marine environment. Currently, path planning algorithms can be divided into two categories: traditional algorithms such as A* algorithm, fast search random tree, or ant colony optimization. However, traditional USV path planning techniques have limitations when facing complex environments, making it difficult to meet the requirements of efficient, safe, and flexible path planning. The other category consists of recently popular swarm intelligence algorithms, such as artificial fish swarm algorithms, slime mold algorithms, and dung beetle algorithms. These algorithms exhibit stronger adaptability and flexibility when dealing with complex environments, providing new solutions for USV path planning. However, they also suffer from slow convergence speeds and difficulty escaping local optima.

[0003] The Red-tailed Hawk Algorithm (RTH) is a swarm intelligence algorithm that simulates the hunting behavior of red-tailed hawks. Compared to other swarm intelligence algorithms, it has advantages such as high local search accuracy and fast convergence speed. However, this algorithm currently suffers from poor global exploration capabilities and is prone to getting trapped in local optima. When applied to path planning for unmanned surface vessels (USVs), it struggles to find optimal navigation paths. For example, during population initialization, the red-tailed hawk population positions are determined by random numbers, leading to a decrease in population diversity later in the algorithm's development. When population diversity decreases, individuals tend to converge to a certain position within the same solution space, easily getting trapped in local optima. Furthermore, when the RTH algorithm is in the low-altitude hovering phase, it uses a spiral approach to move towards the global optimum, which may cause the algorithm to prematurely concentrate on a certain area, neglecting other positions in the solution space, resulting in poor global exploration capabilities. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide an unmanned vessel path planning method, device, system, and storage medium to solve the problem of poor performance of the Red-tailed Eagle optimization algorithm in path planning.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] An unmanned surface vessel (USV) path planning method includes:

[0007] Step S1: Obtain a grid map of obstacles at sea based on the unmanned vessel and the obstacles;

[0008] Step S2: Design a path fitness function based on the shortest path, obstacle penalty, and smoothness of the marine obstacle grid map;

[0009] Step S3: The initial population of the Red-tailed Eagle algorithm is generated using the Tent-Logistic-Cosine chaotic mapping strategy, and the initial optimal individual and optimal fitness value are calculated using the fitness function.

[0010] Step S4: Use the high-altitude soaring, low-altitude hovering, and diving sprint phases to make an initial update of the population location;

[0011] Step S5: Update the existing population position using a hybrid evolution strategy, and determine the optimal solution after the update by boundary condition constraints;

[0012] Step S6: Update the population through the precise hunting phase, and adjust the red-tailed eagle population by comparing the optimal solution and optimal position before and after the update.

[0013] Step S7: Use a speed optimization strategy to adjust the exploration speed of each candidate solution to update the population position;

[0014] Step S8: After each strategy or stage is executed, a boundary condition constraint and candidate solution update judgment must be performed; after reaching the maximum number of iterations, the optimal individual and the optimal fitness value are extracted to obtain the shortest path.

[0015] Preferably, in step S2, the shortest path between the two points is:

[0016]

[0017] Where N is the population size and t is the current iteration number.

[0018] The barrier-free path is calculated using a penalty formula.

[0019] f2 ( t)=E width ·E height ·n B

[0020] Combining the two formulas above, we obtain the adaptive function used in path planning.

[0021]

[0022] Where, n B E is an obstacle that is collided with on the path. width It is the width of the map and E height It is the height of the map.

[0023] Use the path smoothness function to eliminate unnecessary turning points and bends in the path.

[0024]

[0025] Among them, f i '(s) represents the first derivative of the path, f represents the slope at a point on the path, and f'(s) represents the slope at a point on the path. i ″(s) ​​represents the second derivative of the path, ω1 and ω2 are weighting coefficients used to adjust the weights of slope and curvature in the path function, M represents the total number of line segments into which the path is divided, and i represents different line segments in the path.

[0026] As a preferred option, the population update process in step S6 is as follows:

[0027]

[0028] RD = randn(1, dim)

[0029] X newpost =X best +HuntingForce·RD

[0030] Among them, t max The maximum number of iterations is set, t is the current iteration number, RD represents the random exploration hunting direction generated for each candidate solution, and X is the maximum number of iterations. new For the updated candidate solutions of this strategy, HuntingForce is the hunting force;

[0031] During the update process, the hunting intensity is adjusted based on the current iteration count.

[0032]

[0033] Here, mod is a function used to calculate the remainder when two numbers are divided; ifmod(t,10)=0 indicates the case where the remainder when the current iteration number t is divided by 10 is 0. When the remainder is equal to 0, the hunting intensity is adjusted once, that is, the hunting intensity is adjusted once every 10 iterations. In other cases, the hunting intensity remains unchanged.

[0034] As a preferred option, in step S7, the formula for calculating the candidate solution exploration speed is as follows:

[0035] V=ω·V+C1·r1·(X best -X)+C2·r2·(X best -X)

[0036] Where V is the exploration speed, ω is the inertia weight, C1 is the individual speed update factor, C2 is the group speed update factor, and r1 and r2 are random numbers between [0,1].

[0037] Position update formula:

[0038] X new =X pos +V

[0039] Among them, X new This is the updated candidate solution, X. pos This is the current candidate solution.

[0040] Preferably, the boundary conditions in step S8 are as follows:

[0041]

[0042] If the value of X is greater than the upper bound ub, then X new Take the upper bound ub,

[0043] If the value of X is between the upper bound ub and the lower bound lb, then X new Keep the value of X unchanged.

[0044] If the value of X is less than the lower bound lb, then X new Take the lower bound lb,

[0045] Candidate solution update judgment:

[0046] X newcost =fitness(X) newpost )

[0047]

[0048] Among them, X cost X is the fitness value of the current candidate solution. newcost X is the fitness value of the new candidate solution obtained through policy update. post X is the current candidate solution. newpost These are new candidate solutions obtained through policy updates.

[0049] The present invention also provides an unmanned vessel path planning device, comprising:

[0050] The first processing module is used to obtain a grid map of obstacles at sea based on the unmanned vessel and the obstacles.

[0051] The second processing module is used to design a path fitness function based on the shortest path, obstacle penalty, and smoothness of the marine obstacle grid map.

[0052] The third processing module is used to generate the initial population of the Red-tailed Eagle algorithm using the Tent-Logistic-Cosine chaotic mapping strategy, and to calculate the initial optimal individual and optimal fitness value through the fitness function.

[0053] The fourth processing module is used to perform preliminary updates on the population location using the high-altitude soaring, low-altitude hovering, and diving sprint phases.

[0054] The fifth processing module is used to update the existing population position using a hybrid evolution strategy and determine the optimal solution after the update by limiting boundary conditions.

[0055] The sixth processing module is used to update the population through the precise hunting phase and compare the optimal solution and optimal position before and after the update to adjust the red-tailed eagle population.

[0056] The seventh processing module is used to update the population position by adjusting the exploration speed of each candidate solution using a speed optimization strategy;

[0057] The eighth processing module is used to perform boundary condition constraints and candidate solution update judgment after each strategy or stage is executed; after reaching the maximum number of iterations, it extracts the optimal individual and the optimal fitness value to obtain the shortest path.

[0058] This invention also provides an unmanned vessel path planning system, comprising: a memory and a processor, wherein the memory stores a computer program executed by the processor, and the computer program executes an unmanned vessel path planning method when run by the processor.

[0059] This invention also provides a storage medium storing a computer program that executes an unmanned vessel path planning method during runtime.

[0060] This invention employs a multi-strategy improvement to the Red-tailed Eagle optimization algorithm for unmanned surface vessel (USV) path planning, which can significantly enhance the performance of the Red-tailed Eagle algorithm in USV path planning, effectively reduce the detours in the path, and thus achieve accurate planning of the USV's navigation path. Attached Figure Description

[0061] 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.

[0062] Figure 1 This is a flowchart of the unmanned vessel path planning method according to an embodiment of the present invention;

[0063] Figure 2 This is a comparison chart of the optimization iteration curves of the improved Red-tailed Hawk (IRTH) algorithm and other algorithms in 15 benchmark functions according to embodiments of the present invention;

[0064] Figure 3 For the embodiments of the present invention in the local area Figure 1 The path planning effect in the process;

[0065] Figure 4 For the embodiments of the present invention in the local area Figure 1 A schematic diagram of the fitness values ​​of the path planning function in the diagram;

[0066] Figure 5 For the embodiments of the present invention in the local area Figure 2 The path planning effect in the process;

[0067] Figure 6 For the embodiments of the present invention in the local area Figure 2 A schematic diagram of the fitness value of the path planning function. Detailed Implementation

[0068] 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.

[0069] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0070] Example 1:

[0071] like Figure 1 As shown, this embodiment of the invention provides an unmanned surface vessel (USV) path planning method, comprising the following steps:

[0072] S1. Acquire data related to the unmanned surface vessel and obstacles, and construct a two-dimensional grid map based on this data. The map designates whirlpools, reefs, and other obstacles at sea as impassable black squares, while passable areas are represented by white squares.

[0073] S2. Initialize all strategy parameters, setting the maximum number of iterations and population size. Design the objective function by considering constraints such as path shortest distance, path smoothness, obstacles, and environmental boundaries. Path smoothness: The generated navigation path should minimize invalid bends and be as smooth as possible. Environmental boundary and obstacle constraints: The generated movement path must strictly adhere to the boundary grid area set by the map, and each node on the path must not cross any grid marked as an obstacle. Path shortest distance condition: While satisfying the above constraints, the path selection should choose the shortest route between two points.

[0074] The path fitness functions used specifically include:

[0075] The shortest path between two points uses Euclidean theorem:

[0076]

[0077] Where N is the population size and t is the current iteration number.

[0078] To help unmanned boats avoid obstacles, a penalty formula is used to calculate an obstacle-free path.

[0079] f2(t)=E width ·E height ·n B

[0080] Combining the two formulas above yields the adaptive function used in path planning.

[0081]

[0082] Where, n B E is an obstacle that is collided with on the path. width It is the width of the map and E height It is the height of the map.

[0083] Once the initial path planning is complete, the path smoothness function is used to eliminate unnecessary turning points and bends in the path.

[0084]

[0085] Among them, f i (s) represents the position of a point on the path, used to describe the geometry of the path, f i '(s) represents the first derivative of the path, f represents the slope at a point on the path, and f'(s) represents the slope at a point on the path. i ″(s) ​​represents the second derivative of the path, ω1 and ω2 are weighting coefficients used to adjust the weights of the slope and curvature in the path function, M represents the total number of line segments into which the path is divided, and i represents the different line segments in the path.

[0086] S3. Improve the initial population generation stage of the Red-tailed Eagle algorithm using the Tent-Logistic-Cosine chaotic mapping strategy. This chaotic mapping exhibits complex and unpredictable behavior under different parameters and initial conditions, resulting in chaotic sequences with good randomness and uniformity. Applying this chaotic mapping to the initialization of the Red-tailed Eagle algorithm effectively improves the diversity of the algorithm's population, thus overcoming the original algorithm's deficiency in population diversity in the later stages. The formula for this chaotic mapping is as follows:

[0087]

[0088] X(t)=lb+(ub-lb)·Tent-Logistic-Cosine

[0089] Where r is a random number, X is the initial population of the Redtail Eagle algorithm, and lb and ub represent the lower and upper boundaries of the solution space.

[0090] S4, High-altitude soaring stage

[0091] The red-tailed hawk flies high into the sky to find the place with the most abundant food. This algorithm designs its exploration phase by simulating this behavior, considering the place with the most abundant food as the optimal solution, and the high altitude as the entire solution space. The Levy flight function is used to simulate the red-tailed hawk's flight. The specific position update formula is as follows.

[0092] X(t)=X best +(X mean -X(t-1))·Levy·TF(t)

[0093]

[0094] Among them, X best X represents the optimal position in the current iteration. mean Let be the average value at each position, Levy be the red-tailed eagle flight function, and TF(t) be the transition factor function. t represents the current iteration number. max To set the maximum number of iterations, this invention sets it to 1000. μ and ν are random numbers between [0,1]. Γ is the standard gamma function. The value range is usually 100. This invention takes s is a constant of 0.01, and β is a constant of 1.5.

[0095] S5, Low-altitude hovering phase:

[0096] Once the red-tailed eagle has located its prey, it descends and hovers above it, observing and gradually approaching. The red-tailed eagle algorithm simulates this behavior, approaching the optimal solution in the solution space through a spiral approximation method. The specific steps are as follows:

[0097] X(t)=X best +(x(t)+y(t))·StepSize(t)

[0098] StepSize(t) = X(t) - X mean

[0099]

[0100] Where x and y represent direction coordinates; R0∈[0.5,3] is the initial spiral radius; A∈[5,15] is the amplitude factor used to control the scaling of the angle. r∈[1,2] is the final spiral radius, and rand is a random number between [0,1]. R(t) is the spiral radius for each turn, and θ(t) is the spiral angle for each turn.

[0101] After the position is updated, the direction coordinates are normalized to ensure that their range is between [-1, 1].

[0102]

[0103] S6, Crouching Sprint Phase:

[0104] During this phase, the red-tailed hawk will sprint towards its target. The algorithm draws inspiration from simulating the hawk's diving sprint, rapidly approximating the optimal position obtained during the low-altitude soaring phase. The specific steps are as follows:

[0105] X(t) = α(t)·X best +x(t)·StepSize1(t)+y(t)·StepSize2(t)

[0106] StepSizel(t) = X(t) - TF(t)·X mean

[0107] StepSize2(t)=G(t)·X(t)-TF(t)·X best

[0108]

[0109] Where StepSizel1 and StepSizel2 are step size factors; a(t) is an acceleration factor that increases with t; G(t) is a gravity factor, which decreases when the solution is closer to the optimal solution, thus narrowing the optimization range. mean It is the average value of the candidate solutions.

[0110] S7, Hybrid Evolutionary Strategy

[0111] A hybrid evolutionary strategy is used to enhance the global search capability of the Red-tailed Eagle algorithm. The hybrid evolutionary strategy is specifically divided into two stages: mutation and crossover.

[0112] During the mutation phase, the mutation vector is calculated first.

[0113] mutant=X a +F·X b -X c

[0114] Where a, b, and c are three random numbers, each ranging from 1 to N, where N is the population size. These three random numbers are used to randomly select three distinct individuals from the population to determine the mutated individuals that will participate in generating candidate solutions. X a X b X c This represents the candidate solutions selected at positions a, b, and c. F is the mutation weight, which controls the intensity of the mutation. In this invention, F = 0.5.

[0115] Cross-phase:

[0116]

[0117] Where trial is the candidate solution updated after the crossover phase, CR is the crossover probability of the hybrid evolutionary strategy in the crossover phase, and the condition for performing crossover is that crossover is performed when the randomly obtained probability is less than the crossover probability. In this invention, CR is set to 0.9.

[0118] S8, Precision Hunting Phase

[0119] This invention adds a precise hunting stage to the Red-tailed Eagle algorithm, which enhances the local search capability near the global optimum in the later stages of the algorithm.

[0120] The population update process for this strategy is as follows:

[0121]

[0122] RD = randn(1, dim)

[0123] X newpost =X best +HuntingForce·RD

[0124] Among them, t max The maximum number of iterations is set, t is the current iteration number, RD represents the random exploration hunting direction generated for each candidate solution, and X is the maximum number of iterations. new These are candidate solutions for the updated strategy. HuntingForce is the hunting force.

[0125] The precision hunting strategy also adjusts the hunting intensity based on the current iteration number during the update process.

[0126]

[0127] Here, `mod` is a function used to calculate the remainder when two numbers are divided. `ifmod(t,10)=0` indicates the case where the remainder when the current iteration number `t` is divided by 10 is 0. When the remainder is 0, the hunting intensity is adjusted once, that is, the hunting intensity is adjusted once every 10 iterations. In other cases, the hunting intensity remains unchanged.

[0128] S9, Speed ​​Optimization Strategy

[0129] The location is updated by changing the exploration speed of each candidate solution. The specific steps are as follows:

[0130] Formula for calculating candidate solution exploration speed:

[0131] V=ω·V+C1·r1·(X best -X)+C2·r2·(X best -X)

[0132] Where V is the exploration speed, ω is the inertia weight (set to 0.5 in this invention), C1 is the individual speed update factor (set to 1.5), and C2 is the population speed update factor (set to 1.5). r1 and r2 are random numbers between [0,1].

[0133] Position update formula:

[0134] X new =X pos +V

[0135] Among them, X new This is the updated candidate solution, X. pos This is the current candidate solution.

[0136] S10. Boundary conditions and candidate solution update conditions

[0137] After each strategy or phase is executed, a boundary condition constraint and candidate solution update check must be performed. Boundary condition constraints ensure that the updated candidate solutions do not exceed the solution space. The candidate solution update check ensures the effectiveness of the update strategy and avoids blind updates.

[0138] The boundary conditions are as follows:

[0139]

[0140] If the value of X is greater than the upper bound ub, then X new Take the upper bound ub,

[0141] If the value of X is between the upper bound ub and the lower bound lb, then X new Keep the value of X unchanged.

[0142] If the value of X is less than the lower bound lb, then X newTake the lower bound lb,

[0143] Candidate solution update judgment:

[0144] X newcost =fitness(X) newpost )

[0145]

[0146] Among them, X cost X is the fitness value of the current candidate solution. newcost X is the fitness value of the new candidate solution obtained through policy update. post X is the current candidate solution. newpost These are new candidate solutions obtained through policy updates.

[0147] S11. After reaching the maximum number of iterations, extract the optimal individual and the optimal fitness value to obtain the shortest path.

[0148] To verify the effectiveness of the improved algorithm, benchmark function tests and path planning tests were conducted. In the benchmark function tests, the improved algorithm was compared with the original Red-tailed Eagle algorithm, Beluga Whale optimization algorithm, Genetic Algorithm, Gray Wolf optimization algorithm, and Star Crow optimization algorithm across 15 benchmark functions. The formulas or names of the 15 benchmark functions are shown in Table 1. The results of each algorithm's execution in the 15 benchmark function tests are as follows: Figure 2 As shown. In path planning tests, this invention compares the improved algorithm with the other algorithms mentioned above to find the optimal path on two maps with different obstacle distributions. The specific paths are as follows: Figure 3 , Figure 5 As shown, the fitness values ​​of the path planning functions of each algorithm are as follows: Figure 4 , Figure 6 As shown.

[0149] Table 1

[0150]

[0151]

[0152] In the algorithm performance test, the population size of each algorithm was set to 30, the maximum number of iterations was 1000, and each test function was tested 30 times. The final test data of each function is shown in Table 2. In the test under different environments, the maximum number of iterations was set to 200, and the shortest path values ​​of each algorithm are shown in Table 3.

[0153] Table 2

[0154]

[0155]

[0156]

[0157] Table 3

[0158]

[0159] The beneficial effects of the embodiments of the present invention are as follows:

[0160] 1. The Tent-Logistic-Cosine chaotic mapping strategy is used to improve the initial population generation stage of the Red-tailed Eagle algorithm. This chaotic mapping is known for its complex and unpredictable behavior under diverse parameters and initial conditions, and can generate chaotic sequences with strong randomness and uniform distribution characteristics. Applying this chaotic mapping technique to the initialization process of the Red-tailed Eagle algorithm not only enriches the diversity of the population, but also effectively optimizes the problem of insufficient population diversity that may occur in the later stages of the original algorithm, thereby significantly improving the algorithm's global search capability and convergence performance.

[0161] 2. By combining a hybrid evolutionary strategy, the randomness and diversity brought about by the mutation and crossover phase effectively enhance the population's exploration ability and reduce the risk of the algorithm getting trapped in local optima. This allows the algorithm to continuously expand into new search domains while maintaining in-depth exploration of known solution regions. Consequently, when facing complex optimization problems, it can more quickly approach the global optimum or find a near-optimal solution, improving the convergence speed and global search capability of the Red-tailed Eagle algorithm.

[0162] 3. A precise hunting phase was added to the Red-tailed Eagle algorithm. This strategy, through the introduction of random directions, enhances the algorithm's adaptability to the search space, helps it escape local optima, and increases its robustness. The gradually decreasing hunting force allows the algorithm to search the region near the global optimum more precisely in later stages, increasing the probability of finding the exact solution. The proposed precise hunting strategy helps improve the Red-tailed Eagle algorithm's global search capability and enhances its local space optimization ability.

[0163] 4. The speed optimization strategy modifies the exploration speed of each candidate solution to achieve position updates. Utilizing the interaction of inertia weights, individual speed update factors, and group speed update factors, candidate solutions are guided towards better solutions. Inertia weights help candidate solutions maintain their current search direction, while individual and group speed update factors push them closer to their historical best and global best positions, respectively. The introduction of random numbers adds necessary randomness to the search process to prevent the algorithm from prematurely getting trapped in local optima. In this way, the algorithm can continuously explore other unfamiliar regions while maintaining a deep exploration of the known solution space. This strategy enhances the algorithm's exploration capability and adaptability, improving search efficiency and solution quality.

[0164] Example 2:

[0165] This invention also provides an unmanned surface vessel path planning device, comprising:

[0166] The first processing module is used to obtain a grid map of obstacles at sea based on the unmanned vessel and the obstacles.

[0167] The second processing module is used to design a path fitness function based on the shortest path, obstacle penalty, and smoothness of the marine obstacle grid map.

[0168] The third processing module is used to generate the initial population of the Red-tailed Eagle algorithm using the Tent-Logistic-Cosine chaotic mapping strategy, and to calculate the initial optimal individual and optimal fitness value through the fitness function.

[0169] The fourth processing module is used to perform preliminary updates on the population location using the high-altitude soaring, low-altitude hovering, and diving sprint phases.

[0170] The fifth processing module is used to update the existing population position using a hybrid evolution strategy and determine the optimal solution after the update by limiting boundary conditions.

[0171] The sixth processing module is used to update the population through the precise hunting phase and compare the optimal solution and optimal position before and after the update to adjust the red-tailed eagle population.

[0172] The seventh processing module is used to update the population position by adjusting the exploration speed of each candidate solution using a speed optimization strategy;

[0173] The eighth processing module is used to perform boundary condition constraints and candidate solution update judgment after each strategy or stage is executed; after reaching the maximum number of iterations, it extracts the optimal individual and the optimal fitness value to obtain the shortest path.

[0174] Example 3:

[0175] This invention also provides an unmanned vessel path planning system, comprising: a memory and a processor, wherein the memory stores a computer program executed by the processor, and the computer program executes an unmanned vessel path planning method when run by the processor.

[0176] Example 4:

[0177] This invention also provides a storage medium storing a computer program that executes an unmanned vessel path planning method during runtime.

[0178] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for unmanned surface vessel (USV) path planning, characterized in that, include: Step S1: Obtain a grid map of obstacles at sea based on the unmanned vessel and the obstacles; Step S2: Design a path fitness function based on the shortest path, obstacle penalty, and smoothness of the marine obstacle grid map; Step S3: The initial population of the Red-tailed Eagle algorithm is generated using the Tent-Logistic-Cosine chaotic mapping strategy, and the initial optimal individual and optimal fitness value are calculated using the fitness function. Step S4: Use the high-altitude soaring, low-altitude hovering, and diving sprint phases to make an initial update of the population location; Step S5: Update the existing population position using a hybrid evolution strategy, and determine the optimal solution after the update by boundary condition constraints; Step S6: Update the population through the precise hunting phase, and adjust the red-tailed eagle population by comparing the optimal solution and optimal position before and after the update. Step S7: Use a speed optimization strategy to adjust the exploration speed of each candidate solution to update the population position; Step S8: After each strategy or stage is executed, a boundary condition constraint and candidate solution update judgment must be performed; after reaching the maximum number of iterations, the optimal individual and the optimal fitness value are extracted to obtain the shortest path; In step S2, the shortest path between the two points is: Where N is the population size and t is the current iteration number. The barrier-free path is calculated using a penalty formula. f2(t)=E width ·E height ·n B Combining the two formulas above, we obtain the adaptive function used in path planning. Where, n B E is an obstacle that is collided with on the path. width It is the width of the map and E height It is the height of the map. Use the path smoothness function to eliminate unnecessary turning points and bends in the path. Among them, f i '(s) represents the first derivative of the path, f represents the slope at a point on the path, and f'(s) represents the slope at a point on the path. i "(s) represents the second derivative of the path, represents the curvature of the path, ω1 and ω2 are weight coefficients used to adjust the weights of the slope and curvature in the path function, M represents the total number of line segments into which the path is divided, and i represents the different line segments in the path. In step S6, the population update process is as follows: RD = randn(1, dim) X newpost =X best +HuntingForce·RD Among them, t max The maximum number of iterations is set, t is the current iteration number, RD represents the random exploration hunting direction generated for each candidate solution, and X is the maximum number of iterations. new For the updated candidate solutions of this strategy, HuntingForce is the hunting force; During the update process, the hunting intensity is adjusted based on the current iteration count. Here, mod is a function used to calculate the remainder when two numbers are divided; ifmod(t,10)=0 indicates the case where the remainder when the current iteration number t is divided by 10 is 0. When the remainder is equal to 0, the hunting intensity is adjusted once, that is, the hunting intensity is adjusted once every 10 iterations. In other cases, the hunting intensity remains unchanged.

2. The unmanned surface vessel path planning method as described in claim 1, characterized in that, In step S7, the formula for calculating the candidate solution exploration speed is as follows: V=ω·V+C1·r1·(X best -X)+C2·r2·(X best -X) Where V is the exploration speed, ω is the inertia weight, C1 is the individual speed update factor, C2 is the group speed update factor, and r1 and r2 are random numbers between [0,1]. Position update formula: X new =X pos +V Among them, X new This is the updated candidate solution, X. pos This is the current candidate solution.

3. The unmanned vessel path planning method as described in claim 2, characterized in that, In step S8, the boundary conditions are as follows: If the value of X is greater than the upper bound ub, then X new Take the upper bound ub, If the value of X is between the upper bound ub and the lower bound lb, then X new Keep the value of X unchanged. If the value of X is less than the lower bound lb, then X new Take the lower bound lb, Candidate solution update judgment: X newcost =fitness(X newpost ) Among them, X cost X is the fitness value of the current candidate solution. newcost X is the fitness value of the new candidate solution obtained through policy update. post X is the current candidate solution. newpost These are new candidate solutions obtained through policy updates.

4. An unmanned vessel path planning device for implementing the unmanned vessel path planning method of claim 1, characterized in that, include: The first processing module is used to obtain a grid map of obstacles at sea based on the unmanned vessel and the obstacles. The second processing module is used to design a path fitness function based on the shortest path, obstacle penalty, and smoothness of the marine obstacle grid map. The third processing module is used to generate the initial population of the Red-tailed Eagle algorithm using the Tent-Logistic-Cosine chaotic mapping strategy, and to calculate the initial optimal individual and optimal fitness value through the fitness function. The fourth processing module is used to perform preliminary updates on the population location using the high-altitude soaring, low-altitude hovering, and diving sprint phases. The fifth processing module is used to update the existing population position using a hybrid evolution strategy and determine the optimal solution after the update by limiting boundary conditions. The sixth processing module is used to update the population through the precise hunting phase and compare the optimal solution and optimal position before and after the update to adjust the red-tailed eagle population. The seventh processing module is used to update the population position by adjusting the exploration speed of each candidate solution using a speed optimization strategy; The eighth processing module is used to perform boundary condition constraints and candidate solution update judgment after each strategy or stage is executed; after reaching the maximum number of iterations, it extracts the optimal individual and the optimal fitness value to obtain the shortest path.

5. An unmanned surface vessel path planning system, characterized in that, include: A memory and a processor, wherein the memory stores a computer program executed by the processor, the computer program performing the unmanned vessel path planning method as described in any one of claims 1 to 3 when executed by the processor.

6. A storage medium, characterized in that, The storage medium stores a computer program, which executes the unmanned vessel path planning method as described in any one of claims 1 to 3 when it runs.

Images