AUV (Autonomous Underwater Vehicle) multi-target global path planning method based on sparrow search algorithm

Abstract

The invention discloses an AUV (Autonomous Underwater Vehicle) multi-target global path planning method based on a sparrow search algorithm in the field of artificial intelligence and ocean engineering, and the method comprises the following steps: S1, constructing a multi-target optimization model with path length and path safety as the core, and reducing the total length of a path as much as possible while ensuring that the path avoids an obstacle, therefore, a short and safe path is found; s2, determining constraint conditions of path search of the AUV in a complex static environment formed by a plurality of obstacles; s3, solving the multi-objective optimization model established in the step S1 by using a sparrow search strategy; s4, after the algorithm converges, outputting an optimal solution, and drawing an obstacle avoidance 3D schematic diagram and a Pareto frontier map in AUV path planning according to the optimal solution; according to the invention, by improving the sparrow search algorithm and introducing an obstacle avoidance mechanism, the problem of path optimization in a complex underwater environment can be effectively solved, and the navigation efficiency of the AUV and the reliability of task completion are improved.

Description

Technical Field

[0001] The invention belongs to the field of artificial intelligence and marine engineering, and in particular relates to an AUV multi-objective global path planning method based on a sparrow search algorithm. Background Art

[0002] Autonomous Underwater Vehicle (AUV) is an unmanned system that performs tasks autonomously underwater. It is widely used in marine surveys, environmental monitoring, military reconnaissance, pipeline inspection and other fields. AUV has strong autonomous capabilities and can complete complex tasks in underwater environments without the need to maintain continuous communication with ground stations or mother ships. Its significant advantage is that it can perform tasks in deep sea or dangerous areas for a long time, which greatly improves the efficiency and safety of underwater operations.

[0003] However, the complexity and unpredictability of the underwater environment poses a huge challenge to the navigation and path planning of AUVs. When operating underwater, AUVs need to deal with various static and dynamic obstacles (such as ocean terrain, shipwrecks, coral reefs and other underwater moving objects), and must also consider multiple goals, such as minimizing energy consumption, minimizing mission completion time, and ensuring safety. Therefore, the path planning of AUVs becomes a multi-objective optimization problem, which requires finding an optimal path that is both safe and efficient.

[0004] When implementing AUV path planning, intelligent algorithms are often used to solve path optimization problems in complex environments. When faced with underwater path planning, traditional optimization algorithms often show shortcomings such as being prone to falling into local optimality, limited search efficiency, and lack of environmental adaptability. For example, in a complex non-convex environment, the particle swarm optimization algorithm is prone to converge to a local optimal solution, especially in the presence of multiple obstacles. Such defects will seriously affect the path quality and mission success rate of the AUV. Although the crossover and mutation process of the genetic algorithm can increase the diversity of solutions, it also increases the complexity of the algorithm, resulting in insufficient efficiency in real-time path planning. When the AUV needs to respond quickly to complex underwater environments, the inefficiency of traditional algorithms may lead to the loss of the opportunity to select the best path. In addition, traditional algorithms usually require frequent reinitialization or parameter adjustment in dynamic environments, which is not only time-consuming, but may also interrupt the continuity of path planning, thereby reducing the effectiveness and flexibility of the AUV in dynamic environments. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides an AUV multi-objective global path planning method based on a sparrow search algorithm, which simulates the foraging behavior of sparrows to find a path when the AUV performs path search, randomly generates an initial path in three-dimensional space, and gradually improves the path solution through iterative updates to find an optimal path from the starting point to the end point, thereby effectively avoiding obstacles and reaching the target point.

[0006] To achieve the above object, the present invention adopts the following technical scheme: a multi-objective global path planning method for AUV based on sparrow search algorithm, comprising the following steps:

[0007] Step 1) Construct a multi-objective optimization model with path length and path safety as the core. Its goal is to minimize the path length and path safety cost at the same time, that is, to minimize the total length of the path while ensuring that the path avoids obstacles, so as to find a path that is both short and safe;

[0008] Step 2) establishing the constraints for the AUV to perform path search in a complex static environment with multiple obstacles;

[0009] Step 3) Solving the multi-objective optimization model established in step 1 using the sparrow search strategy;

[0010] Step 4) When the algorithm converges, the optimal solution is output, and a 3D schematic diagram and a Pareto frontier diagram of obstacle avoidance in the AUV path planning are drawn based on the optimal solution.

[0011] As a further improvement of the present invention, step 1) is specifically:

[0012] The path length objective function f1 is established as the first objective function to measure the total length of the path. The calculation formula is as follows:

[0013]

[0014] Where N is the number of points in the path; p i =(x i ,y i ,z i ), represents the i-th point in the path; ||p i+1 -p i || means from point p i To point p i+1 The Euclidean distance of

[0015] The safety objective function f2 is established as the second objective function, which is used to measure the distance between the path and the obstacle. It is determined by calculating the combination of the minimum distance and the average distance from each point on the path to the obstacle. The calculation formula is as follows:

[0016]

[0017] Among them, minDist is the minimum distance between all points on the path and obstacles; avgDist is the average distance between all points on the path and obstacles; α is an adjustment parameter used to control the weight of path length and path safety; δ is a constant used to avoid division by zero errors.

[0018] As a further improvement of the present invention, step 2) is specifically:

[0019] Establish constraints on collision detection, boundary conditions, personal and global optimal choices, and path optimization strategies as follows:

[0020] Step 2.1) Establish obstacle constraints and detect whether the path or path point collides with obstacles;

[0021] Step 2.2) Establish spatial boundary constraints, by setting the boundaries of the three-dimensional space to ensure that the points in the path and the positions of the sparrows are within the given three-dimensional space range;

[0022] Step 2.3) Establish individual and global optimal solution constraints to ensure that the path of each sparrow moves towards a better solution, so as to achieve a balance between global search and local optimization, and gradually approach the Pareto optimal frontier of multi-objective optimization;

[0023] Step 2.4) Path smoothing and optimization: The path is further optimized through path smoothing and greedy algorithm to improve the feasibility and safety of the path.

[0024] As a further improvement of the present invention, step 2.1) is specifically as follows:

[0025] Spherical obstacles and cylindrical obstacles are randomly generated as follows:

[0026] Step 2.1.1) Establish the constraints of the spherical obstacle:

[0027] p i Represents any point on the path. If the path point meets the following conditions, it means that the collision condition is met, and the path point needs to be regenerated. The formula is as follows:

[0028]

[0029] ||p i -c j ||<R j

[0030] Among them, p i =(x i ,y i ,z i );The center of the jth spherical obstacle The radius is R j ;||p i -c j || represents the path point p i To the center of the ball c j The Euclidean distance of

[0031] Step 2.1.2) Establish the constraints of the cylindrical obstacle:

[0032] If the path point p i If the following conditions are met, it means that the collision conditions are met, and the path points need to be regenerated. The formula is as follows:

[0033]

[0034] d axis <R k

[0035]

[0036] Among them, the bottom center position of the kth cylindrical obstacle is The radius is R k , height H k ;d axis Indicates the horizontal distance from the path point to the cylinder axis.

[0037] As a further improvement of the present invention, the formula of the space boundary constraint condition in step 2.2) is:

[0038]

[0039] Among them, x i ,y i 、z i They are path points p i The formula ensures that the waypoints do not exceed the predefined working space.

[0040] As a further improvement of the present invention, step 2.3) is specifically as follows:

[0041] Step 2.3.1) Leader Behavior

[0042] The leader behavior refers to the updating of individual sparrows according to the global optimal solution. The specific calculation formula is as follows:

[0043]

[0044] in, is the new position of the i-th sparrow in the t+1-th iteration; g is the position of the global optimal solution; ε is a scaling factor; r1 is a random number vector with the same dimension as the position, which obeys the standard normal distribution and is used to add a certain random disturbance;

[0045] Step 2.3.2) Follower Behavior

[0046] The follower behavior refers to the updating of individual sparrows according to their own optimal solutions or through heuristic random exploration. The specific calculation formula is as follows:

[0047]

[0048] Among them, p i is the optimal position of the individual solution of the i-th sparrow; r2 is a random number vector that follows a standard normal distribution;

[0049] Updated through heuristic random exploration, the specific calculation formula is as follows:

[0050]

[0051] Among them, r3 is a random number vector that follows a uniform distribution [0, 1]; λ is an exploration factor.

[0052] As a further improvement of the present invention, step 2.4) is specifically as follows:

[0053] Step 2.4.1) Path smoothing formula

[0054] Use cubic spline interpolation to smooth the path and increase the continuity and smoothness of the path. The specific formula is as follows:

[0055] p(t)=(x(t),y(t),z(t)),t∈[1,N]

[0056] Where N is the number of points in the path; t is the interpolation parameter; the original point set in the path is x(t), y(t), and z(t) are the x, y, and z coordinate values ​​corresponding to t obtained by cubic spline interpolation;

[0057] Step 2.4.2) Greedy algorithm optimizes the path

[0058] Use the greedy algorithm to optimize the path. By trying multiple locations in a small range, select the path point with the minimum cost to increase the safety of the path. The specific process is as follows:

[0059] Let the current path point be p i =(x i ,y i ,z i), use the step size Δ to make a small adjustment to obtain the candidate path point p i ', the specific formula is as follows:

[0060] p i '=(x i +Δx,y i +Δy,z i +Δz)

[0061] Cost function cost(p i ') is defined as:

[0062]

[0063] Among them, d min is the new position p i 'The distance to the nearest obstacle, η is a constant used to prevent division by zero errors;

[0064] Then, from all the attempted positions, the position with the lowest cost is selected. The specific formula is as follows:

[0065] p i =argmincost(p i ')

[0066] Through this greedy local search strategy, a local optimal solution is selected at each step.

[0067] As a further improvement of the present invention, step 3) is specifically:

[0068] The sparrow search strategy is used to find an optimal path from the starting point to the end point in three-dimensional space, so that the AUV can effectively avoid obstacles and move in a safe environment. The sparrow search algorithm simulates the foraging behavior of sparrows, performs global search and local optimization to find a balance between path length and safety, including the following stages:

[0069] 3.1) Encoding stage: The path planning problem is expressed in the form of sparrow positions, and the path is represented by a series of three-dimensional coordinates;

[0070] 3.2) Population initialization stage: randomly generate an initial sparrow population, each individual represents a potential movement strategy or action plan from the starting point to the end point;

[0071] 3.3) Fitness evaluation phase: Calculate the target value of each path, comprehensively consider the path length and safety, and evaluate the path quality of each sparrow;

[0072] 3.4) Selection, update and diversity maintenance phase: Update the sparrow positions based on the leader and follower behaviors, while maintaining the diversity of solutions and saving non-dominated solutions through external archives;

[0073] 3.5) Iteration and termination conditions: Continue iterating until the maximum number of iterations is reached, and output the optimal path as the final solution.

[0074] Compared with the prior art, the present invention has the following beneficial effects:

[0075] (1) The present invention adopts the sparrow search algorithm (SSA) through multiple iterations and fitness evaluations, and finally forms a Pareto frontier set containing multiple non-dominated solutions in the external archive. By selecting the optimal solution in the archive and further smoothing and optimizing it, the code effectively finds the path planning solution of the AUV in a complex three-dimensional environment. The output optimal solution has undergone multiple smoothing and greedy optimization to ensure the smoothness and feasibility of the path.

[0076] (2) Compared with traditional optimization algorithms such as particle swarm optimization, the latter is prone to fall into local optimal solutions when dealing with complex non-convex spaces, especially in multi-objective optimization problems. This problem is more common. The sparrow search algorithm used in the present invention combines the behavioral strategies of leaders and followers to enable the population to have a stronger ability to jump out of local optimal solutions, thereby improving search efficiency.

[0077] (3) The traditional particle swarm optimization algorithm may lose population diversity in the later stage, resulting in insufficient search ability. In contrast, the sparrow search algorithm used in the present invention can maintain a high diversity of the population during the initialization and iteration process, and can better cope with the dynamic changes of the environment, thereby finding multiple high-quality solutions.

[0078] (4) In a dynamic or unpredictable environment, traditional optimization algorithms often need to reinitialize or adjust parameters, while the sparrow search algorithm used in the present invention can quickly adapt to changes in the environment through flexible individual behavior adjustment, thereby maintaining the continuity and effectiveness of the optimization process. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0080] Figure 1 The figure is a flow chart of the algorithm of the present invention.

[0081] Figure 2 This is a flow chart of the multi-objective optimization method for AUV path planning of the present invention.

[0082] Figure 33D schematic diagram of obstacle avoidance in AUV path planning of the present invention.

[0083] Figure 4 The present invention draws a Pareto frontier graph based on the optimal solution. DETAILED DESCRIPTION

[0084] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0085] like Figure 1-2 The multi-objective global path planning method for AUV based on the sparrow search algorithm shown includes the following steps:

[0086] Step 1) Build a multi-objective optimization model with path length and path safety as the core:

[0087] The goal of the constructed multi-objective optimization model is to minimize the path length and path safety cost at the same time, that is, to minimize the total length of the path while ensuring that the path avoids obstacles, so as to find a path that is both short and safe. Specifically:

[0088] The path length objective function f1 is established as the first objective function to measure the total length of the path. The calculation formula is as follows:

[0089]

[0090] Where N is the number of points in the path (intermediate points plus the start and end points); p i =(x i ,y i ,z i ), represents the i-th point in the path; ||p i+1 -p i || means from point p i To point p i+1 The Euclidean distance of .

[0091] The safety objective function f2 is established as the second objective function to measure the distance between the path and the obstacle. The purpose is to minimize the cost of the path approaching the obstacle. It is determined by calculating the combination of the minimum distance and the average distance from each point on the path to the obstacle. The calculation formula is as follows:

[0092]

[0093] Where minDist is the minimum distance between all points on the path and obstacles; avgDist is the average distance between all points on the path and obstacles; α is an adjustment parameter that controls the weight of path length and path safety. In the present invention, α=0.6, which means that it is more inclined to minimize the shortest distance between the path and obstacles; δ is a very small constant used to avoid zero division errors. In the present invention, δ=1×10 -3 .

[0094] Step 2) Establish the constraints for the AUV to perform path search in a complex static environment with multiple obstacles:

[0095] 2.1) Obstacle Constraints

[0096] The constraint condition of the obstacle is used to detect whether the path or the path point collides with the obstacle. If a collision is detected, the path point should be regenerated. In the present invention, in order to simulate the obstacles on the seabed, spherical obstacles and cylindrical obstacles are randomly generated. The following is a specific formula for the constraint conditions of the spherical obstacles and the cylindrical obstacles.

[0097] 2.1.1) Constraints of spherical obstacles: This method calculates the distance from each point in the path to the center of each spherical obstacle and determines whether it is less than the radius of the sphere. If it is less than, it indicates a collision. The specific calculation formula is as follows:

[0098]

[0099] ||p i -c j ||<R j

[0100] Among them, p i =(x i ,y i ,z i );The center of the jth spherical obstacle The radius is R j ,||p i -c j || represents the path point p i To the center of the ball c j Euclidean distance, in order to ensure that the path point p i There is no collision with obstacle j, and the above distance must be greater than the radius R of the obstacle j .

[0101] 2.1.2) Constraints of cylindrical obstacles: This method calculates the distance from each point in the path to the cylinder axis and checks whether the height of the point is within the range of the cylinder. If the height of the point is within the range of the cylinder, it indicates a collision and the position of the path point needs to be regenerated to avoid the collision.

[0102] The condition for the path point to collide with the cylinder is that the horizontal distance from the path point to the cylinder axis is less than the radius of the cylinder and the height z of the path point is i Within the height range of the cylinder, the specific formula is as follows:

[0103]

[0104] d axis <R k

[0105]

[0106] Among them, the bottom center position of the jth cylindrical obstacle is The radius is R c , height is H c ;p i represents any point on the path; d axis Indicates the horizontal distance from the path point to the cylinder axis. For a cylindrical obstacle, if a collision is detected, it means that each point p on the path i =(x i ,y i ,z i ) all satisfy the above constraints.

[0107] 2.2) Spatial boundary constraints

[0108] In this method, the boundary of the three-dimensional space is set. The constraint of the space boundary is used to ensure that the points and positions of the sparrows in the path are within the given three-dimensional space range, which means that all sparrows and path points must be within this range. The specific formula is as follows:

[0109]

[0110] Among them, x i ,y i 、z i They are path points p i The entire path planning process is restricted to the cubic space of [0,100]×[0,100]×[0,100] to ensure that the path points do not exceed the predefined working space, thereby ensuring the rationality of the planned path.

[0111] 2.3) Constraints on individual and global optimal solutions

[0112] The constraints of personal optimal solution and global optimal solution are used to ensure that the path of each sparrow moves towards a better solution. This ensures that the update of the sparrow is not only constrained by the personal historical optimal path, but also guided by the global optimal path, so as to gradually optimize the effect of path planning, thereby achieving a balance between global search and local optimization, and gradually approaching the Pareto optimal frontier of multi-objective optimization.

[0113] 2.3.1) Leader Behavior (Based on Global Optimal Solution)

[0114] Leader behavior refers to the updating of individual sparrows according to the global optimal solution to guide the solution to converge in a better direction. The specific calculation formula is as follows:

[0115]

[0116] in, is the new position of the i-th sparrow in the t+1th iteration; g is the position of the global optimal solution; ε is a proportional factor, which is used in this method to control the amplitude of random noise, set to ε = 0.03; r1 is a random number vector with the same dimension as the position, which obeys the standard normal distribution and is used to add a certain random disturbance. This formula indicates that the i-th sparrow moves closer to the global optimal solution with a small amount of random disturbance ε to maintain the diversity of solutions.

[0117] 2.3.2) Follower behavior (based on personal optimal solution or global exploration)

[0118] Follower behavior refers to the updating of individual sparrows according to their personal optimal solution or through heuristic random exploration. The specific calculation formula for the update of individual sparrows according to their personal optimal solution is as follows:

[0119]

[0120] Among them, p i is the personal optimal solution position of the i-th sparrow; ε is also the amplitude control of random noise, set to ε = 0.03; r2 is a random number vector that obeys the standard normal distribution.

[0121] The specific calculation formula for updating through heuristic random exploration is as follows:

[0122]

[0123] Among them, g is the position of the global optimal solution; r3 is a random number vector that follows a uniform distribution [0,1]; λ is an exploration factor. This formula indicates that the individual sparrows conduct random exploration based on the global optimal position to maintain the global search capability of the algorithm.

[0124] 2.4) Path smoothing and optimization

[0125] In this method, the path is further optimized through path smoothing and greedy algorithm to improve the feasibility and safety of the path. The specific calculation formula is as follows:

[0126] 2.4.1) Path smoothing formula

[0127] This method uses cubic spline interpolation to smooth the path, increase the continuity and smoothness of the path, and make the path more in line with the requirements of practical applications. The cubic spline interpolation formula is as follows:

[0128] p(t)=(x(t),y(t),z(t)),t∈[1,N]

[0129] Where N is the number of points in the path; t is the interpolation parameter; the original point set in the path is x(t), y(t), and z(t) are the x, y, and z coordinate values ​​of t obtained by cubic spline interpolation.

[0130] 2.4.2) Greedy algorithm optimization path

[0131] This method uses a greedy algorithm to optimize the path. By trying multiple positions in a small range, the path point with the minimum cost, that is, the path point with the largest distance from the obstacle, is selected to increase the safety of the path. The specific process of greedy optimization is as follows:

[0132] Let the current path point be p i =(x i ,y i ,z i ), use the step size Δ to make a small adjustment to obtain the candidate path point p i ', the specific formula is as follows:

[0133] p i '=(x i +Δx,y i +Δy,z i +Δz)

[0134] In this method, Δ=0.3 is set, and the current point is adjusted in each dimension to generate a new path point.

[0135] Cost function cost(p i ') is defined as:

[0136]

[0137] Among them, d min is the new position p i 'The distance to the nearest obstacle; η = 1×10 -3is a small constant used to prevent division by zero errors. This formula defines the cost function as the inverse of the distance to the obstacle.

[0138] Then, from all the attempted positions, the position with the lowest cost is selected. The specific formula is as follows:

[0139] p i =argmincost(p i ')

[0140] Through this greedy local search strategy, the local optimal solution can be selected at each step, thereby improving the overall safety of the path.

[0141] The above optimization steps and formulas ensure that the path maintains the shortest path length and maximizes the distance to obstacles during the path planning process, thereby achieving better path planning results.

[0142] Step 3) Use the sparrow search strategy to solve the multi-objective optimization model established in step 1):

[0143] The Sparrow Search Algorithm (SSA) is used for path planning, which is to find an optimal path from the starting point to the end point in three-dimensional space, so that the AUV can effectively avoid obstacles and move in a safe environment. SSA simulates the foraging behavior of sparrows, performs global search and local optimization, and thus finds a balance between multiple objectives (path length and safety), including the following stages:

[0144] 3.1) Encoding stage: The path planning problem is expressed in the form of sparrow positions, and the path is represented by a series of three-dimensional coordinates;

[0145] 3.2) Population initialization stage: randomly generate an initial sparrow population, each individual represents a potential movement strategy or action plan from the starting point to the end point;

[0146] 3.3) Fitness evaluation phase: Calculate the target value of each path, comprehensively consider the path length and safety, and evaluate the path quality of each sparrow;

[0147] 3.4) Selection, update and diversity maintenance phase: Update the sparrow positions based on the leader and follower behaviors, while maintaining the diversity of solutions and saving non-dominated solutions through external archives;

[0148] Step 4) When the algorithm converges, output the optimal solution:

[0149] When the algorithm converges, the optimal solution is output, and a 3D schematic diagram of obstacle avoidance and a Pareto frontier diagram in AUV path planning are drawn based on the optimal solution, as shown in Figure 3 and Figure 4 shown.

[0150] In the present invention, the sparrow search algorithm (SSA) forms a Pareto front set containing multiple non-dominated solutions in the external archive through multiple iterations and fitness evaluations. By selecting the optimal solution in the archive and further smoothing and optimizing it, the code effectively finds the path planning solution for the AUV in a complex three-dimensional environment. The output optimal solution has undergone multiple smoothing and greedy optimization to ensure the smoothness and feasibility of the path.

[0151] Overall, the performance of the sparrow search algorithm in path planning fully demonstrates its ability to handle nonlinear and multi-objective optimization problems. By combining global and local optimal solutions, and maintaining diversity and exploration, the output optimal solution achieves the expected performance goals. This provides a reliable path planning solution for the application of AUVs in complex underwater environments, which can not only ensure the effective completion of the task, but also significantly reduce energy consumption and potential risks.

[0152] The invention is not limited to the above-mentioned embodiments. On the basis of the technical solution disclosed in the present invention, technicians in this field can make some substitutions and deformations to some technical features therein according to the disclosed technical content without creative labor, and these substitutions and deformations are all within the protection scope of the present invention.

Claims

1. A multi-objective global path planning method for AUV based on sparrow search algorithm, characterized in that: The following steps are involved: Step 1) Construct a multi-objective optimization model with path length and path safety as the core. Its goal is to minimize the path length and path safety cost at the same time, that is, to minimize the total length of the path while ensuring that the path avoids obstacles, so as to find a path that is both short and safe; Step 2) establishing the constraints for the AUV to perform path search in a complex static environment with multiple obstacles; Step 3) Solving the multi-objective optimization model established in step 1 using the sparrow search strategy; Step 4) When the algorithm converges, the optimal solution is output, and a 3D schematic diagram and a Pareto frontier diagram of obstacle avoidance in the AUV path planning are drawn based on the optimal solution.

2. The AUV multi-objective global path planning method based on the sparrow search algorithm according to claim 1 is characterized in that: Step 1) is specifically as follows: The path length objective function f1 is established as the first objective function to measure the total length of the path. The calculation formula is as follows: Where N is the number of points in the path; p i =(x i ,y i ,z i ), represents the i-th point in the path; ||p i+1 -p i || means from point p i To point p i+1 The Euclidean distance of The safety objective function f2 is established as the second objective function, which is used to measure the distance between the path and the obstacle. It is determined by calculating the combination of the minimum distance and the average distance from each point on the path to the obstacle. The calculation formula is as follows: Among them, minDist is the minimum distance between all points on the path and obstacles; avgDist is the average distance between all points on the path and obstacles; α is an adjustment parameter used to control the weight of path length and path safety; δ is a constant used to avoid division by zero errors.

3. The AUV multi-objective global path planning method based on the sparrow search algorithm according to claim 1 is characterized in that: Step 2) is specifically as follows: Establish constraints on collision detection, boundary conditions, personal and global optimal choices, and path optimization strategies as follows: Step 2.1) Establish obstacle constraints and detect whether the path or path point collides with obstacles; Step 2.2) Establish spatial boundary constraints, by setting the boundaries of the three-dimensional space to ensure that the points in the path and the positions of the sparrows are within the given three-dimensional space range; Step 2.3) Establish individual and global optimal solution constraints to ensure that the path of each sparrow moves towards a better solution, so as to achieve a balance between global search and local optimization, and gradually approach the Pareto optimal frontier of multi-objective optimization; Step 2.4) Path smoothing and optimization: The path is further optimized through path smoothing and greedy algorithm to improve the feasibility and safety of the path.

4. The AUV multi-objective global path planning method based on the sparrow search algorithm according to claim 3 is characterized in that: Step 2.1) is specifically: Spherical obstacles and cylindrical obstacles are randomly generated as follows: Step 2.1.1) Establish the constraints of the spherical obstacle: p i Represents any point on the path. If the path point meets the following conditions, it means that the collision condition is met, and the path point needs to be regenerated. The formula is as follows: ||p i -c j ||<R j Among them, p i =(x i ,y i ,z i );The center of the jth spherical obstacle The radius is R j ;||p i -c j || represents the path point p i To the center of the ball c j The Euclidean distance of Step 2.1.2) Establish the constraints of the cylindrical obstacle: If the path point p i If the following conditions are met, it means that the collision conditions are met, and the path points need to be regenerated. The formula is as follows: d axis <R k Among them, the bottom center position of the kth cylindrical obstacle is The radius is R k , height H k ;d axis Indicates the horizontal distance from the path point to the cylinder axis.

5. The AUV multi-objective global path planning method based on the sparrow search algorithm according to claim 3 is characterized in that: The formula for the spatial boundary constraint in step 2.2) is: Among them, x i ,y i 、z i They are path points p i The formula ensures that the waypoints do not exceed the predefined working space.

6. The AUV multi-objective global path planning method based on the sparrow search algorithm according to claim 3 is characterized in that: Step 2.3) is as follows: Step 2.3.1) Leader Behavior The leader behavior refers to the updating of individual sparrows according to the global optimal solution. The specific calculation formula is as follows: in, is the new position of the i-th sparrow in the t+1-th iteration; g is the position of the global optimal solution; ε is a scaling factor; r1 is a random number vector with the same dimension as the position, which obeys the standard normal distribution and is used to add a certain random disturbance; Step 2.3.2) Follower Behavior The follower behavior refers to the updating of individual sparrows according to their own optimal solutions or through heuristic random exploration. The specific calculation formula is as follows: Among them, p i is the optimal position of the individual solution of the i-th sparrow; r2 is a random number vector that follows a standard normal distribution; Updated through heuristic random exploration, the specific calculation formula is as follows: Among them, r3 is a random number vector that follows a uniform distribution [0, 1]; λ is an exploration factor.

7. The AUV multi-objective global path planning method based on the sparrow search algorithm according to claim 3 is characterized in that: Step 2.4) is specifically: Step 2.4.1) Path smoothing formula Use cubic spline interpolation to smooth the path and increase the continuity and smoothness of the path. The specific formula is as follows: p(t)=(x(t),y(t),z(t)),t∈[1,N] Where N is the number of points in the path; t is the interpolation parameter; the original point set in the path is x(t), y(t), and z(t) are the x, y, and z coordinate values ​​corresponding to t obtained by cubic spline interpolation; Step 2.4.2) Greedy algorithm optimizes the path Use the greedy algorithm to optimize the path. By trying multiple locations in a small range, select the path point with the minimum cost to increase the safety of the path. The specific process is as follows: Let the current path point be p i =(x i ,y i ,z i ), use the step size Δ to make a small adjustment to obtain the candidate path point p i ', the specific formula is as follows: p i '=(x i +Δx,y i +Δy,z i +Δz) Cost function cost(p i ') is defined as: Among them, d min is the new position p i 'The distance to the nearest obstacle, η is a constant used to prevent division by zero errors; Then, from all the attempted positions, the position with the lowest cost is selected. The specific formula is as follows: p i =argmincost(p i ') Through this greedy local search strategy, a local optimal solution is selected at each step.

8. The AUV multi-objective global path planning method based on the sparrow search algorithm according to claim 1 is characterized in that: Step 3) is specifically: The sparrow search strategy is used to find an optimal path from the starting point to the end point in three-dimensional space, so that the AUV can effectively avoid obstacles and move in a safe environment. The sparrow search algorithm simulates the foraging behavior of sparrows and performs global search and local optimization to find a balance between path length and safety. The following phases are included: 3.1) Encoding stage: The path planning problem is expressed in the form of sparrow positions, and the path is represented by a series of three-dimensional coordinates; 3.2) Population initialization stage: randomly generate an initial sparrow population, each individual represents a potential movement strategy or action plan from the starting point to the end point; 3.3) Fitness evaluation phase: Calculate the target value of each path, comprehensively consider the path length and safety, and evaluate the path quality of each sparrow; 3.4) Selection, update and diversity maintenance phase: Update the sparrow positions based on the leader and follower behaviors, while maintaining the diversity of solutions and saving non-dominated solutions through external archives; 3.5) Iteration and termination conditions: Continue iterating until the maximum number of iterations is reached, and output the optimal path as the final solution.

Images