An improved JADE algorithm-based adaptive path planning method for underwater glider

Abstract

The application relates to an underwater glider adaptive path planning method based on an improved JADE algorithm, which comprises the following steps: firstly, a two-dimensional plane coordinate system is established, and a coordinate system conversion is performed according to the positions of a starting point and a target point; secondly, an underwater glider steady gliding model and an energy consumption model are established, a gliding speed and an energy consumption calculation method are obtained, and the influence of a sea current on the underwater glider is considered; then, the parameter updating strategies of a scaling factor and a crossover factor of the original JADE algorithm are improved, and the optimization capacity of the JADE algorithm is improved; finally, the energy consumption cost is taken as an optimization target, and the improved JADE algorithm is used to plan a series of path points and the best pitching angle and diving depth of each path section for the underwater glider. The application can enable the underwater glider to adaptively select the best movement mode according to the marine environment, and safely reach the target point with the lowest energy consumption.

Description

Technical Field

[0001] This invention belongs to the field of underwater vehicle path planning technology, and relates to an adaptive path planning method for underwater gliders based on the improved JADE algorithm. Background Technology

[0002] An Autonomous Underwater Glider (AUG) is a new type of autonomous underwater vehicle (AUV) powered by gravity-buoyancy difference. Compared to traditional propeller-powered AUVs, AUGs offer advantages such as low power consumption, long endurance, quiet operation, and stealth. Therefore, AUGs have a wide range of applications and are particularly suitable for marine environmental monitoring and resource exploration. Low-energy path planning is crucial for AUGs to successfully complete long-range, long-endurance exploration missions. Furthermore, AUGs glide at low speeds up and down in a zigzag pattern, making them highly susceptible to ocean currents during gliding; therefore, path planning must consider factors such as terrain and ocean currents.

[0003] Traditional AUG path planning primarily focuses on planning waypoints, neglecting to plan the AUG's pitch angle and diving depth. This traditional approach is advantageous due to its simplicity and ease of implementation. However, many algorithms exist to address this problem, such as A... Algorithms include fast random search tree algorithm, artificial potential field method, genetic algorithm, particle swarm algorithm and differential evolution algorithm. However, traditional path planning algorithms have the following problems: (1) They rarely take into account the AUG steady-state gliding model, AUG energy consumption model and marine environment (topography and ocean currents, etc.) for path planning; (2) The AUG motion mode is fixed, and the pitch angle and diving depth remain unchanged throughout the gliding process, and the motion parameters cannot be adaptively adjusted according to the topography and ocean currents.

[0004] To address the shortcomings of traditional path planning methods, this invention proposes an adaptive path planning method for underwater gliders based on an improved JADE algorithm, taking a practical engineering application perspective. This method combines the AUG steady-state gliding model and an energy consumption model, optimizing the motion parameters (pitch angle and diving depth) of each path segment based on terrain and ocean current information while simultaneously planning waypoints. Compared to the traditional AUG path planning method, this invention offers advantages in terms of better environmental adaptability and energy consumption optimization capabilities.

[0005] The purpose of this invention is to solve the problem of AUG adaptive path planning in complex marine environments. Considering seabed topographic constraints and ocean current interference, an adaptive path planning method with minimum energy consumption as the optimization objective is proposed. The planning includes a series of path points and the corresponding AUG pitch angle and diving depth for each path segment. Summary of the Invention

[0006] Technical problems to be solved To overcome the shortcomings of existing technologies and improve the AUG's adaptability to the marine environment and energy consumption optimization capabilities, this invention proposes an adaptive path planning method for underwater gliders based on an improved JADE algorithm. This method can rationally plan path points and the optimal pitch angle and diving depth for each path segment based on the marine environment (topography and currents). First, the method analyzes the dynamics and kinematics of the AUG, establishes a steady-state gliding model, and derives the relationship between gliding speed, net buoyancy, and gliding angle. Second, it establishes an AUG energy consumption model, analyzing energy consumption calculation methods related to and independent of gliding time. Finally, considering the marine environment, the improved JADE algorithm is used to plan an adaptive path for the AUG.

[0007] Technical solution An adaptive path planning method for underwater gliders based on an improved JADE algorithm, characterized by the following steps: Step 1: Establish the AUG steady-state gliding model, that is, the horizontal and vertical velocities during steady-state gliding are: (3); (4); In the formula, and They are horizontal velocity and vertical velocity, respectively. For gliding speed, The pitch angle for gliding; Step 2: Establish the AUG velocity synthesis model under the influence of ocean currents as follows: (8); In the formula, The desired direction of the synthesis rate; path point To the waypoint The speed of the ocean current; It is the matrix transpose; Step 3: Establish a time-energy consumption model for a profile of the AUG gliding, including the energy consumption of devices related to gliding time. Energy consumption of a profile buoyancy adjustment system Energy consumption of a profile attitude adjustment system ; Step 4: Determine the decision variables and encoding method for AUG adaptive path planning: The AUG adaptive path planning establishes a planar coordinate system. Determine the starting coordinates of the path planning task. and endpoint coordinates Starting point As the origin, pointing to direction as The axis is then rotated according to the right-hand rule. Sure The axis, then the new coordinate system and The points in the array have the following correspondence: (13); In the formula, coordinate axes with coordinate axes The included angle, coordinate system The point in the middle; The decision variables and encoding method for the AUG adaptive path planning are as follows: (14); In the formula, Path segments New coordinate system Projected length on the axis; In the new coordinate system, the intermediate path points The vertical coordinate on; The corresponding pitch angle of the AUG in each path segment; The corresponding diving depth of the AUG in each path segment; the intermediate path points are path points excluding the start and end points; Step 5: Based on the decision variables in Step 4, establish the optimization objective function, i.e., the total energy consumption of AUG adaptive path planning: (18); In the formula, The average power of devices related to gliding time. For the first The length of each path segment For the first Gliding time of the path segment No. The horizontal glide distance of the AUG path segment; Step 6: Improve the crossover factor in the original JADE algorithm and scaling factor for: (19); (20); In the formula, The mean of the scaling factors, and Follows a normal distribution; Let be the mean of the cross-factor, and It follows a Cauchy distribution; The improved update strategy is as follows: Sort all individuals in the population in descending order of their fitness value (i.e., the optimization target value); then, prioritize individuals with higher fitness values. For individuals in the population, the scaling factor and crossover factor are updated in the same way, according to the crossover factor. and scaling factor Update; and after the fitness value Individuals adopt new update strategies: The new update strategy: (twenty one); (twenty two); (twenty three); (twenty four); in, and for Random numbers between This is the minimum value of the scaling factor. The maximum value of the scaling factor. It is a fixed constant; Step 7: Optimize the objective function in Step 5 using the improved JADE algorithm, and output the optimal decision variables.

[0008] The process of establishing the AUG steady-state gliding model in step 1 is as follows: when the AUG is steadily descending or ascending, the gliding speed... It is a gliding angle and net buoyancy Functions: (1); In the formula, , , and The hydrodynamic coefficients are constant; the glide angle during steady-state gliding using the AUG. With pitch angle and attack angle The relationship between them is: (2); Since the angle of attack of the AUG during steady-state gliding is negligible, the glide angle is approximately equal to the pitch angle. Based on the characteristics of the AUG's V-shaped glide motion, the horizontal and vertical velocities during steady-state gliding are obtained: (3); (4); In the formula, and These are the horizontal velocity and the vertical velocity, respectively.

[0009] The process of establishing the AUG velocity synthesis model under the influence of ocean currents in step 2 is as follows: Let the horizontal velocity of the AUG in still water be... Path points To the waypoint The ocean current speed is ,but arrive The synthesis rate of ocean currents is considered in this stage: (5); (6); (7); (8); In the formula, The desired direction of synthesis speed.

[0010] The energy consumption of the gliding time-related devices in step 3 Energy consumption of a profile buoyancy control system Energy consumption of a profile attitude control system ;in: For the time it takes for the AUG to glide across a profile, For diving depth, The average power of devices related to gliding time. Indicates the density of seawater. Represents gravitational acceleration. Indicates the hydraulic pump flow rate. Indicates the power factor. For the power of the moving motor, The slider's movement speed, For the quality of AUG, For the mass of the slider, To maintain a stable mindset.

[0011] The process of establishing the optimization objective function in step 5 is as follows: taking AUG as the starting point... The target point is Since the number of path segments is Therefore, the number of intermediate waypoints (excluding the waypoints between the origin and the destination) is And use Indicates intermediate waypoints; make The set of path segments is: (15); Based on the AUG velocity synthesis model from step 2, the path segment is obtained. The corresponding synthesis speeds Therefore, AUG in the The gliding time for the path segment is: (16); In the formula, , ; No. The horizontal gliding distance of the AUG path segment is: (17); Based on the energy consumption of devices related to gliding time Energy consumption of a profile buoyancy adjustment system Energy consumption of a profile attitude adjustment system for: (18).

[0012] The optimization process of the objective function in step 7 using the improved JADE algorithm is as follows: Step 71: Initialize the parameters of the improved JADE algorithm, including the average scaling factor. Cross factor mean Maximum number of iterations Population size ,in Even numbers, fixed constant values The initial population is determined according to formula (14) in step 3. (25); Step 72: Calculate the population individuals using formula (18). (in, First iteration Each iteration The objective function value (26); The original population was sorted in ascending order of fitness value (i.e., the value of formula (26)). Sort the population and obtain the new population after sorting. ; Step 73, targeting Individual populations are updated using formulas (19) and (20). and The update results are as follows: and ; and Individual populations are updated using formulas (23) and (24). and The update results are as follows: and ; Step 74: Perform the mutation operation. The mutation method is as follows: (27); In the formula, This indicates the current iteration number, during the first iteration. ; Indicates a fixed constant; Indicates from An individual randomly selected from the population; Indicates from An individual randomly selected from the population; Indicates from A randomly selected individual Offspring with lower internal storage fitness values ​​are prioritized in the first iteration. It is an empty set; Step 75: Perform the crossover operation. First, randomly generate an integer. (28); In the formula Decision variables The number of; and then the specific crossover method is: (29); In the formula, ; ; Step 76: Perform the selection operation: (30); In the formula, ; This represents the value calculated by formula (18); Good storage value; Better storage value; Step 77: Randomly delete Some individuals in the population made ; Step 78: Update the mean of the scaling factor and cross factor Mean; (31); (32); (33); (34); In the formula, It is a fixed constant and ; for Number of individuals in the in-population; Step 79: Determine if the algorithm termination condition is met. If the termination condition is not met, repeat formulas (26) to (34); if the termination condition is met, output the optimal decision variable.

[0013] (35).

[0014] Beneficial effects This invention proposes an adaptive path planning method for underwater gliders based on an improved JADE algorithm. First, a two-dimensional coordinate system is established, and coordinate transformation is performed based on the positions of the starting point and the target point. Second, a steady-state gliding model and an energy consumption model for the underwater glider are established, deriving methods for calculating gliding speed and energy consumption, while considering the influence of ocean currents on the underwater glider. Then, the parameter update strategies for scaling and cross factors in the original JADE algorithm are improved, enhancing the algorithm's optimization capability. Finally, using energy consumption cost as the optimization objective, the improved JADE algorithm is used to plan a series of waypoints and the optimal pitch angle and diving depth for each waypoint segment for the underwater glider. This invention enables the underwater glider to adaptively select the optimal motion mode according to the marine environment, safely reaching the target point with minimal energy consumption.

[0015] The present invention proposes an adaptive path planning method for underwater gliders based on an improved JADE algorithm, which brings the following beneficial effects: 1. The path planning takes into account the AUG steady-state gliding model and energy consumption model, and the calculation method of gliding speed and energy consumption is derived. 2. The adaptive parameter (cross factor and scaling factor) update strategy of the original JADE algorithm was improved, which improved the convergence performance of the algorithm; 3. Based on practical engineering applications of AUG, this invention focuses on adaptive path planning for AUG. The research content is novel, with the planning content consisting of a series of path points and the pitch angle and diving depth of each path segment; 4. Adaptive path planning methods can effectively reduce the energy consumption of AUG and improve its adaptability to the marine environment (topography, ocean currents). Attached Figure Description

[0016] Figure 1 This is a force analysis diagram of the steady-state gliding process of the method of the present invention.

[0017] Figure 2 This is a schematic diagram of coordinate system transformation and two-dimensional plane adaptive path planning in the method of the present invention.

[0018] Figure 3 This is a schematic diagram of the three-dimensional adaptive path planning method of the present invention.

[0019] Figure 4 The flowchart shows the improved JADE algorithm of the present invention.

[0020] Figure 5 This is a diagram showing the two-dimensional planar path planning result of the method of the present invention.

[0021] Figure 6 This is a diagram showing the planning results of pitch angle, diving depth, and number of gliding profiles for each path segment using the method of this invention.

[0022] Figure 7 This is a diagram showing the three-dimensional spatial path planning result of the method of the present invention.

[0023] Figure 8 This research focuses on adaptive path planning methods for underwater gliders. Detailed Implementation

[0024] The present invention will now be further described in conjunction with the embodiments and accompanying drawings: The technical solution of this invention is as follows: The aforementioned adaptive path planning method for underwater gliders includes the following steps: Step 1: Establish an AUG steady-state gliding model; Step 2: Establish an AUG velocity synthesis model under the influence of ocean currents; Step 3: Establish the AUG energy consumption model; Step 4: Determine the decision variables and encoding method for AUG adaptive path planning; Step 5: Based on the decision variables in Step 4, establish the optimization objective function; Step 6: Improve the update strategy for the crossover factor and scaling factor in the original JADE algorithm; Step 7: Optimize the objective function in Step 5 using the improved JADE algorithm, and output the optimal decision variables.

[0025] To fully utilize favorable terrain and ocean currents to reduce AUG energy consumption, this invention proposes an adaptive path planning method for underwater gliders based on an improved JADE algorithm. The implementation of the algorithm requires establishing a steady-state gliding model and an energy consumption model for the AUG, while also considering the influencing factors such as terrain and ocean currents. By combining the model with the marine environment, a series of waypoints and the pitch angle and diving depth for each path segment are planned for the AUG, enabling it to safely reach the target point with minimal energy consumption.

[0026] In step 1, when the AUG is in steady-state descent or ascent, the gliding speed... It is a gliding angle and net buoyancy Functions: (1); In the formula, , , and The hydrodynamic coefficients are constant; the glide angle during steady-state gliding using the AUG. With pitch angle and attack angle The relationship between them is: (2); Since the angle of attack of the AUG during steady-state gliding is negligible, the glide angle is approximately equal to the pitch angle. Based on the characteristics of the AUG's V-shaped glide motion, the horizontal and vertical velocities during steady-state gliding are obtained: (3); (4); In the formula, and These are the horizontal velocity and the vertical velocity, respectively.

[0027] In step 2, assume the horizontal velocity of the AUG in still water is... Path points To the waypoint The ocean current speed is ,but arrive The method for calculating the composite velocity of ocean currents considering each stage is as follows: (5); (6); (7); (8); In the formula, The desired direction of synthesis speed.

[0028] In step 3, the AUG's energy consumption can be divided into energy consumption related to gliding time and energy consumption unrelated to gliding time. Energy consumption related to gliding time is mainly generated by the control and navigation system, measurement sensors, and positioning and communication system; energy consumption unrelated to gliding time is mainly generated by the buoyancy adjustment system and attitude adjustment system.

[0029] Step 3 yields the following specific process: 1. Calculate the energy consumption related to gliding time. The time for gliding one profile in AUG is: (9); In the formula, Let be the diving depth. Assume the average power of the devices related to the gliding time is... Therefore, the energy consumption of this part is (10).

[0030] 2. Calculate energy consumption independent of gliding time. The energy consumption of the AUG gliding profile buoyancy control system is... (11); In the formula, Indicates the density of seawater. Represents gravitational acceleration. Indicates the hydraulic pump flow rate. This represents the power factor.

[0031] 3. The energy consumption of the AUG gliding attitude control system in one profile is... (12); In the formula, For the power of the moving motor, The slider's movement speed, For the quality of AUG, For the mass of the slider, To maintain a stable mindset.

[0032] In step 4, a planar coordinate system is established. Determine the starting coordinates of the path planning task. and endpoint coordinates Starting from the beginning As the origin, pointing to direction as The axis is then rotated according to the right-hand rule. Sure The axis, then the new coordinate system and The points in the array have the following correspondence: (13); In the formula, coordinate axes with coordinate axes The included angle, coordinate system The point in the middle.

[0033] Assume the number of path segments from the start point to the end point in AUG is . The length of each path segment can be determined by the coordinates of the two path points. In addition, a path segment can consist of multiple gliding profiles. Within the same path segment, the motion parameters (pitch angle and dive depth) of each gliding profile of the AUG are the same.

[0034] To solve the adaptive path planning problem, the decision variables and encoding method are designed as follows: (14); In the formula, Path segments New coordinate system Projected length on the axis; In the new coordinate system, the intermediate path points The vertical coordinate on; The corresponding pitch angle of the AUG in each path segment; The corresponding diving depth of the AUG in each path segment; the intermediate path points are path points excluding the starting point and the ending point.

[0035] In step 5, the starting point of AUG is... The target point is Since the number of path segments is Therefore, the number of intermediate waypoints (excluding the waypoints between the origin and the destination) is And use Indicates intermediate waypoints.

[0036] make The set of path segments is: (15); According to formula (8), the path segment can be obtained. The corresponding synthesis speeds Therefore, AUG in the The gliding time of the path segment is (16); In the formula, , .

[0037] No. The horizontal gliding distance of the AUG on the path segment is (17); Based on formulas (10), (11), and (12) in step 3, the optimization objective function (i.e., total energy consumption) of AUG adaptive path planning is expressed as follows: (18).

[0038] In step 6, to improve the search capability of the original JADE algorithm, this invention improves the parameter update strategy for the scaling factor and cross factor of the original JADE algorithm. The scaling factor of the original JADE algorithm... and cross factor The update methods are respectively (19); (20); In the formula, The mean of the scaling factors, and Follows a normal distribution; Let be the mean of the cross-factor, and It follows a Cauchy distribution.

[0039] Improved and The parameter update strategy is as follows: First, sort all individuals in the population in descending order of their fitness value (i.e., the optimization target value); then, sort the individuals with the highest fitness values... For individuals in the population, the scaling factor and crossover factor are updated in the same way as in formulas (19) and (20); while the fitness value is updated after... Individuals adopt new update strategies: (twenty one); (twenty two); (twenty three); (twenty four); in, and for Random numbers between This is the minimum value of the scaling factor. The maximum value of the scaling factor. It is a fixed constant.

[0040] In step 7, the objective function from step 5 is optimized using the improved JADE algorithm. The specific process is as follows: First, initialize the parameters of the improved JADE algorithm, including the mean scaling factor. Cross factor mean Maximum number of iterations Population size ( (even number), fixed constant value The initial population is determined according to formula (14) in step 3. (25).

[0041] Then, the population individuals are calculated using formula (18). (in, First iteration Each iteration The objective function value (26).

[0042] The original population was sorted in ascending order of fitness value (i.e., the value of formula (26)). Sort the population and obtain the new population after sorting. .

[0043] against Individual populations are updated using formulas (19) and (20). and The update results are as follows: and ; and Individual populations are updated using formulas (23) and (24). and The update results are as follows: and .

[0044] Perform a mutation operation, the mutation method is as follows: (27); In the formula, This indicates the current iteration number, during the first iteration. ; Indicates a fixed constant; Indicates from An individual randomly selected from the population; Indicates from An individual randomly selected from the population; Indicates from A randomly selected individual Offspring with lower internal storage fitness values ​​are prioritized in the first iteration. It is an empty set.

[0045] To perform the crossover operation, first generate a random integer: (28); In the formula Decision variables The number of; and then the specific crossover method is: (29); In the formula, ; .

[0046] Perform the selection operation: (30); In the formula, ; This represents the value calculated by formula (18); Good storage value; Better storage value.

[0047] Random deletion Some individuals in the population made .

[0048] Update the mean of the scaling factor and cross factor Mean.

[0049] (31); (32); (33); (34); In the formula, It is a fixed constant and ; for The number of individuals in the in-population.

[0050] Finally, determine whether the algorithm termination condition is met. If the termination condition is not met, repeat equations (26) to (34); if the termination condition is met, output the optimal decision variable.

[0051] (35). Specific implementation examples: This invention uses a path planning task for an underwater glider in a certain sea area from the starting point to the target point as an example to verify the adaptive path planning method: 1. Analysis of AUG steady-state gliding model.

[0053] The force analysis process for steady-state diving and steady-state surfacing in AUG is similar. This invention takes steady-state diving as an example (e.g.) Figure 1 ), obtained from torque balance ; ; In the formula, This indicates that the AUG is experiencing resistance. This indicates the lift force experienced by the AUG. This represents the net buoyancy of the AUG. The lift can be calculated using the above formula. and resistance Respectively with net buoyancy and gliding angle The relationship between them: ; ; The hydrodynamics of AUG can also be approximated as: ; ; Based on torque balance and hydrodynamic formulas, the steady-state gliding speed can be obtained. With angle of attack and gliding angle The relationship between them and the angle of attack with gliding angle Relationship: ; ; Once the angle of attack is determined, the steady-state gliding speed depends only on the net buoyancy and the gliding angle. Therefore, the steady-state gliding speed is: ; Based on the above steady-state gliding model analysis, the parameters are taken as follows: , , , and It can obtain the relationship between AUG gliding speed and net buoyancy and gliding angle, and the relationship between horizontal gliding speed and vertical gliding speed and gliding attitude.

[0054] 2. AUG energy consumption model analysis.

[0055] Get parameters , , , , , , , , and The energy consumption of an AUG gliding through a profile can be calculated using the formula described in step 3.

[0056] ; ; ; In the formula, Energy consumption related to gliding time; Energy consumption related to diving depth; Energy consumption related to attitude adjustment.

[0057] 3. Obtain seabed topography and ocean current data; The adaptive path planning method requires seabed topography and ocean current data, which are obtained in the following way: For seafloor topography data, this method utilizes publicly available seafloor topography data obtained through Global Multi-Resolution Topography (GMRT). GMRT offers various file formats; this method downloads ArcASCII files with a grid resolution of 61 m / node. For ocean current data, this method uses current data provided by CopernicusMarine.

[0058] 4. Determine the decision variables and coding method for AUG adaptive path planning. like Figure 2 and Figure 3 As shown, The proposed adaptive path planning method for underwater gliders can be described as adaptively planning waypoints based on the marine environment (topography and ocean currents). and each path segment pitch angle between and diving depth This adaptive path planning method has better environmental adaptability than traditional path planning methods with fixed pitch angles and diving depths, and can effectively reduce the energy consumption of underwater gliders. Based on the characteristics of the adaptive path planning method, the decision variables and encoding methods designed in this method are as follows: ; In the formula, Path segments New coordinate system Projected length on the axis; In the new coordinate system, the intermediate path points The vertical coordinate on; The corresponding pitch angle of the AUG in each path segment; This corresponds to the AUG's diving depth in each path segment. In this embodiment, the number of intermediate waypoints in the path planning task is set to 9, resulting in a total of 10 path segments. The specific decision variables and encoding methods are as follows: .

[0059] 5. Solve the adaptive path planning problem for underwater gliders using an improved JADE algorithm. First, initialize the parameters of the improved JADE algorithm, including the mean scaling factor. Cross factor mean Maximum number of iterations Population size Fixed constant value The initial population is determined according to formula (14) in step 3: .

[0060] Then, the population individuals are calculated using formula (18). (in, First iteration Each iteration The objective function value ; The original population was sorted in ascending order of fitness value (i.e., the value of formula (26)). Sort the population and obtain the new population after sorting. .

[0061] against Individual populations are updated using formulas (19) and (20). and ; The update results are as follows: and ; and Individual populations are updated using formulas (23) and (24). and The update results are as follows: and .

[0062] Furthermore, perform a mutation operation, the mutation method being: ; In the formula, This indicates the current iteration number, during the first iteration. ; Indicates a fixed constant; Indicates from An individual randomly selected from the population; Indicates from An individual randomly selected from the population; Indicates from A randomly selected individual Offspring with lower internal storage fitness values ​​are prioritized in the first iteration. It is an empty set.

[0063] Furthermore, to perform the crossover operation, first, a random integer is generated: ; In the formula The specific crossover method is as follows: ; In the formula, ; .

[0064] Further, perform the selection operation: ; In the formula, ; This represents the value calculated by formula (18); Good storage value; Better storage value.

[0065] Furthermore, random deletion Some individuals in the population made .

[0066] Furthermore, update the mean of the scaling factor. and cross factor Mean.

[0067] ; ; ; ; In the formula, ; for The number of individuals in the in-population.

[0068] Finally, determine whether the algorithm termination condition is met. If the termination condition is not met, repeat equations (26) to (34); if the termination condition is met, output the optimal decision variable. : ; The AUG two-dimensional plane path can be obtained based on the optimal decision variables. Figure 5 ), AUG pitch angle, diving depth and number of gliding profiles in different path segments ( Figure 6 ) and the gliding path of the AUG in three-dimensional space ( Figure 7 The planning results show that the AUG can adaptively select its pitch angle and diving depth according to the marine environment, and reach the target point with the current.

Claims

1. An adaptive path planning method for underwater gliders based on an improved JADE algorithm, characterized in that... The steps are as follows: Step 1: Establish the AUG steady-state gliding model, that is, the horizontal and vertical velocities during steady-state gliding are: （3）； （4）； In the formula, and They are horizontal velocity and vertical velocity, respectively. For gliding speed, The pitch angle for gliding; Step 2: Establish the AUG velocity synthesis model under the influence of ocean currents as follows: （8）； In the formula, The desired direction of the synthesis rate; path point To the waypoint The speed of the ocean current; It is the matrix transpose; Step 3: Establish a time-energy consumption model for a profile of the AUG gliding, including the energy consumption of devices related to gliding time. Energy consumption of a profile buoyancy adjustment system Energy consumption of a profile attitude adjustment system ; Step 4: Determine the decision variables and encoding method for AUG adaptive path planning: The AUG adaptive path planning establishes a planar coordinate system. Determine the starting coordinates of the path planning task. and endpoint coordinates Starting point As the origin, pointing to direction as The axis is then rotated according to the right-hand rule. Sure The axis, then the new coordinate system and The points in the array have the following correspondence: （13）； In the formula, coordinate axes with coordinate axes The included angle, coordinate system The point in the middle; The decision variables and encoding method for the AUG adaptive path planning are as follows: （14）； In the formula, Path segments New coordinate system Projected length on the axis; In the new coordinate system, the intermediate path points The vertical coordinate on; The corresponding pitch angle of the AUG in each path segment; The corresponding diving depth of the AUG in each path segment; the intermediate path points are path points excluding the start and end points; Step 5: Based on the decision variables in Step 4, establish the optimization objective function, i.e., the total energy consumption of AUG adaptive path planning: （18）； In the formula, The average power of devices related to gliding time. For the first The length of each path segment For the first Gliding time of the path segment No. The horizontal glide distance of the AUG path segment; Step 6: Improve the crossover factor in the original JADE algorithm and scaling factor for: （19）； （20）； In the formula, The mean of the scaling factors, and Follows a normal distribution; Let be the mean of the cross-factor, and It follows a Cauchy distribution; The improved update strategy is as follows: Sort all individuals in the population in descending order of their fitness value (i.e., the optimization target value); then, prioritize individuals with higher fitness values. For individuals in the population, the scaling factor and crossover factor are updated in the same way, according to the crossover factor. and scaling factor Update; and after the fitness value Individuals adopt new update strategies: The new update strategy: （21）； （22）； （23）； （24）； in, and for Random numbers between This is the minimum value of the scaling factor. The maximum value of the scaling factor. It is a fixed constant; Step 7: Optimize the objective function in Step 5 using the improved JADE algorithm, and output the optimal decision variables.

2. The underwater glider adaptive path planning method based on the improved JADE algorithm according to claim 1, characterized in that: The process of establishing the AUG steady-state gliding model in step 1 is as follows: when the AUG is steadily descending or ascending, the gliding speed... It is a gliding angle and net buoyancy Functions: （1）； In the formula, , , and The hydrodynamic coefficients are constant; the glide angle during steady-state gliding using the AUG. With pitch angle and attack angle The relationship between them is: （2）； Since the angle of attack of the AUG during steady-state gliding is negligible, the glide angle is approximately equal to the pitch angle. Based on the characteristics of the AUG's V-shaped glide motion, the horizontal and vertical velocities during steady-state gliding are obtained: （3）； （4）； In the formula, and These are the horizontal velocity and the vertical velocity, respectively.

3. The underwater glider adaptive path planning method based on the improved JADE algorithm according to claim 1, characterized in that: The process of establishing the AUG velocity synthesis model under the influence of ocean currents in step 2 is as follows: Let the horizontal velocity of the AUG in still water be... Path points To the waypoint The ocean current speed is ,but arrive The synthesis rate of ocean currents is considered in this stage: （5）； （6）； （7）； （8）； In the formula, The desired direction of synthesis speed.

4. The underwater glider adaptive path planning method based on the improved JADE algorithm according to claim 1, characterized in that: The energy consumption of the gliding time-related devices in step 3 ; Energy consumption of a profile buoyancy control system ; Energy consumption of a profile attitude control system ;in: For the time it takes for the AUG to glide across a profile, For diving depth, The average power of devices related to gliding time. Indicates the density of seawater. Represents gravitational acceleration. Indicates the hydraulic pump flow rate. Indicates the power factor. For the power of the moving motor, The slider's movement speed, For the quality of AUG, For the mass of the slider, To maintain a stable mindset.

5. The underwater glider adaptive path planning method based on the improved JADE algorithm according to claim 1, characterized in that: The process of establishing the optimization objective function in step 5 is as follows: taking AUG as the starting point... The target point is Since the number of path segments is Therefore, the number of intermediate waypoints is And use This indicates an intermediate waypoint, which is a waypoint other than the origin and destination. make The set of path segments is: （15）； Based on the AUG velocity synthesis model from step 2, the path segment is obtained. The corresponding synthesis speeds Therefore, AUG in the The gliding time for the path segment is: （16）； In the formula, , ; No. The horizontal gliding distance of the AUG path segment is: （17）； Based on the energy consumption of devices related to gliding time Energy consumption of a profile buoyancy adjustment system Energy consumption of a profile attitude adjustment system The calculation yielded: （18）。 6. The underwater glider adaptive path planning method based on the improved JADE algorithm according to claim 1, characterized in that: The optimization process of the objective function in step 7 using the improved JADE algorithm is as follows: Step 71: Initialize the parameters of the improved JADE algorithm, including the average scaling factor. Cross factor mean Maximum number of iterations Population size ,in Even numbers, fixed constant values The initial population is determined according to formula (14) in step 4. （25）； Step 72: Calculate the population individuals using formula (18). objective function value （26）； in, First iteration Each iteration According to the fitness value, i.e., the value of formula (26), the original population is processed in ascending order. Sort the population and obtain the new population after sorting. ; Step 73, targeting Individual populations are updated using formulas (19) and (20). and The update results are as follows: and ; and Individual populations are updated using formulas (23) and (24). and The update results are as follows: and ; Step 74: Perform the mutation operation. The mutation method is as follows: （27）； In the formula, This indicates the current iteration number, during the first iteration. ; Indicates a fixed constant; Indicates from An individual randomly selected from the population; Indicates from An individual randomly selected from the population; Indicates from A randomly selected individual Offspring with lower internal storage fitness values ​​are prioritized in the first iteration. It is an empty set; Step 75: Perform the crossover operation. First, randomly generate an integer. （28）； In the formula Decision variables The number of; and then the specific crossover method is: （29）； In the formula, ; ； Step 76: Perform the selection operation: （30）； In the formula, ; This represents the value calculated by formula (18); Good storage value; Better storage value; Step 77: Randomly delete Some individuals in the population made ; Step 78: Update the mean of the scaling factor and cross factor Mean; （31）； （32）； （33）； （34）； In the formula, It is a fixed constant and ; for Number of individuals in the in-population; Step 79: Determine if the algorithm termination condition is met. If the termination condition is not met, repeat formulas (26) to (34); if the termination condition is met, output the optimal decision variable. （35）。

Images