An underwater glider robot anchoring process energy consumption optimization method and system based on a sub-domain gray box model

Abstract

The application discloses an underwater glider robot anchoring process energy consumption optimization method and system based on a sub-domain gray box model, and the energy consumption optimization method comprises the following steps: establishing a white box energy consumption mechanism model in the underwater glider robot anchoring process; dividing the whole anchoring energy consumption stage of the underwater glider robot into multiple sub-domains, and using a Latin hypercube design method to simulate and generate multiple sub-domain samples meeting variable range constraints by using a hardware-in-the-loop simulation platform; selecting a Kriging model as a sub-domain surrogate model, fusing the white box energy consumption mechanism model and constraint conditions of an oil tank volume and a movable mass position, forming a sub-domain gray box model, and performing segmented energy consumption fitting by using the sub-domain gray box model; taking energy consumption minimization fitted by the sub-domain gray box model as a target, iteratively optimizing the sub-domain surrogate model by using a dynamic guidance self-adjusting particle swarm optimization algorithm, and outputting optimal planning parameters. The application realizes low-energy consumption control of the whole anchoring process, and is suitable for tasks such as ocean resource exploration and hydrological monitoring.

Description

Technical Field

[0001] This invention relates to the field of underwater robot control technology, and in particular to a method and system for optimizing energy consumption during the mooring process of an underwater gliding robot based on a domain gray box model. Background Technology

[0002] With the development of marine science and engineering, underwater glider robots (UGRs) have become core equipment for marine environmental monitoring and data acquisition due to their advantages of long endurance, low energy consumption, and large-scale operation. Anchoring is a crucial function for underwater glider robots to complete fixed-point observations, and its energy consumption directly determines the observation duration. Current research on robot energy consumption optimization mainly focuses on unmanned aerial vehicles (UAVs), industrial equipment, and transportation vehicles: in the UAV field, research revolves around trajectory design, resource allocation, and cooperative control, balancing computational complexity and energy consumption through joint optimization frameworks; in the industrial equipment field, energy redundancy is reduced from the perspectives of modular control and multi-axis coupling effect optimization; in the transportation field, a "time-energy" balance is achieved through methods such as dynamic programming and mixed-integer linear programming. Research on underwater glider robots, however, focuses on path optimization and navigation energy reduction, such as full-coverage path planning based on ant colony algorithms and autonomous navigation schemes using lightweight long short-term memory networks (SN-LSTM). However, existing technologies have significant shortcomings: First, there is a lack of an energy consumption mechanism model for underwater gliders anchoring, covering the entire process of "diving-leveling-stationary holding"; second, black-box energy consumption mechanism models lack physical constraints, are prone to overfitting, and are highly dependent on sample size; third, traditional energy consumption optimization methods are mostly oriented towards continuously differentiable functions, which cannot adapt to the significant piecewise energy consumption characteristics of underwater gliders anchoring. Summary of the Invention

[0003] This invention provides a method and system for optimizing energy consumption during the anchoring process of an underwater gliding robot based on a domain-specific gray box model, in order to solve the technical problems mentioned in the background.

[0004] To achieve the above objectives, the technical solution of the present invention is implemented as follows:

[0005] This invention provides an energy consumption optimization method for the anchoring process of an underwater glider based on a domain-specific gray box model, comprising the following steps:

[0006] S1. Based on the kinematic model and gliding energy consumption model of the underwater gliding robot, establish a white-box energy consumption mechanism model and constraints on the oil bladder volume and movable mass position during the anchoring process of the underwater gliding robot.

[0007] S2. Build a hardware-in-the-loop simulation platform. Divide the entire anchoring energy consumption stage of the underwater glider into multiple subdomains according to the different ranges of the oil bladder volume. Design the variable range constraints for each subdomain. Based on the white-box energy consumption mechanism model and using the Latin hypercube design method, use the hardware-in-the-loop simulation platform to simulate and generate multiple subdomain samples that meet the variable range constraints.

[0008] S3. The Kriging model is selected as the domain proxy model. The white box energy consumption mechanism model and the constraints of oil bladder volume and movable mass position are integrated to form a domain gray box model. The domain gray box model is then used to fit the segmented energy consumption.

[0009] S4. With the goal of minimizing the energy consumption fitted by the domain gray box model, the dynamic-guided self-adjusting particle swarm algorithm is used to iteratively optimize the domain proxy model and output the optimal planning parameters.

[0010] Furthermore, step S1 specifically includes the following steps:

[0011] S11. The anchoring process of the underwater glider is divided into three stages: diving, leveling, and stationary holding.

[0012] S12. Based on the kinematic model of the underwater glider, establish the mapping relationship between the pitch angle and the fuel bladder volume and the movable mass; then construct the constraints on the fuel bladder volume and the position of the movable mass.

[0013] S13. Based on the kinematic model and gliding energy consumption model of the underwater gliding robot, establish a white-box energy consumption mechanism model of the underwater gliding robot during the anchoring process that satisfies the constraints and mapping relationships in S12.

[0014] Among them, the white-box energy consumption mechanism model is the sum of the energy consumption of the buoyancy adjustment system, pitch control system, integrated system and anchoring system on the underwater glider during the anchoring process; the integrated system is the control, measurement and communication system.

[0015] Furthermore, the specific expression of the white-box energy consumption mechanism model is as follows:

[0016] (1)

[0017] in, This represents the total energy consumption of the underwater gliding robot during the anchoring process; This indicates the energy consumption of the buoyancy adjustment system on an underwater glider during the anchoring process; This indicates the energy consumption of the pitch control system on the underwater glider during the mooring process; This indicates the energy consumption of the integrated system on the underwater glider during the mooring process; This indicates the energy consumption of the anchoring system on an underwater glider during the anchoring process;

[0018] Energy consumption , , , The expressions are as follows:

[0019] (2)

[0020] (3)

[0021] (4)

[0022] (5)

[0023] in, , These represent the current depth as follows: The power and flow rate of the pump on the buoyancy regulation system; and These represent the oil bladder volumes before and after the change, respectively. and These represent the depth ranges representing the dynamic process of oil bladder volume during descent and leveling, respectively. and These represent the depths of the underwater glider before and after the change in the oil bladder volume during the leveling phase. It refers to the mass of the anchor in the anchoring system; It is the density of water; and These represent the power and flow rate of the solenoid valve in the buoyancy adjustment system, respectively. and These represent the axial coordinates of the movable mass before it moves. axial coordinates after movement The pitch angle, with the axis being the direction of movement of the movable mass; and These represent the time before and after the movable mass position is moved, respectively. This indicates that when the pitch angle is The power of the pitch control motor on the pitch control system; It is the power of the integrated system; and These represent the start and end times of the integrated system's operation, respectively. It is the power of the winch on the anchoring system; and These represent the start and end times of anchoring, respectively. Indicates time; This indicates the depth of the underwater gliding robot.

[0024] Furthermore, the mapping relationship between the pitch angle and the oil bladder volume and movable mass in S12 is specifically expressed as follows:

[0025] (6)

[0026] in, It is movable mass; It is the radial coordinate of the movable mass before it moves; It refers to all mass on an underwater glider except for movable mass; and These are the axial and radial coordinates of the center of mass of the movable mass before it moves; It is net buoyancy; It is gravitational acceleration; and It is the hydrodynamic coefficient, and the radial direction mentioned above is the direction perpendicular to the axial direction.

[0027] Furthermore, step S2 specifically includes the following steps:

[0028] S21. Build a hardware-in-the-loop simulation platform and divide the entire underwater gliding robot anchoring energy consumption stage into four sub-domains according to the different ranges of oil bladder volume.

[0029] S22. Design the variable range constraints for each domain, as follows:

[0030] I. Constraints between depth and ground altitude;

[0031] II. Coupling constraints between the movable mass axial coordinate, oil bladder volume, and pitch angle;

[0032] III. Maximum and minimum constraints on the axial coordinates of movable mass;

[0033] S23. Based on the white-box energy consumption mechanism model and using the Latin hypercube design method, use an in-loop simulation platform to simulate and generate global samples and multiple subdomain samples that satisfy the variable range constraints of each subdomain.

[0034] Furthermore, the expression for the constraint between depth and ground altitude in S22 is as follows:

[0035] (7)

[0036] in, Indicates the first i The depth of the underwater glider in the sub-domain; Indicates the first i The ground altitude of the underwater glider in the sub-domain; This represents the maximum value of the sum of the set depth and the altitude above the ground.

[0037] The expression for the coupling constraint between the movable mass axial coordinate, oil bladder volume, and pitch angle is as follows:

[0038] (8)

[0039] in, Indicates the first i The axial coordinates of the movable mass in the domain before it moves; Indicates the first i The volume of the oil sac before the change in the domain;

[0040] The expressions for the maximum and minimum constraints of the axial coordinates of the movable mass are as follows:

[0041] (9)

[0042] (10)

[0043] in, , They represent the first time. i The maximum and minimum values ​​of the axial coordinates of the movable mass in the domain before it moves; , These represent the maximum and minimum pitch angles of the underwater glider, respectively. , They represent the first time. i The maximum and minimum values ​​of the oil bladder volume before the change in the domain; , These represent the maximum and minimum values ​​of the axial coordinates of the movable mass before it moves during the entire anchoring energy consumption phase.

[0044] Furthermore, each sub-domain sample in S3 includes basic parameters and planning parameters, wherein the basic parameters include the depth of the underwater glider before diving. Bottom height Oil sac volume before change And the axial coordinates of the movable mass before it moves. ;

[0045] Planning parameters include the changed oil bladder volume And the axial coordinates after the movable mass has moved. .

[0046] Furthermore, step S3 specifically includes the following steps:

[0047] S31. Initially select one of the response surface method, radial basis function model, and Kriging model as the domain proxy model;

[0048] S32. Based on the samples in each sub-domain, the response surface method, radial basis function model, and kriging model are validated and evaluated using median absolute deviation, maximum absolute error, root mean square error, and coefficient of determination, respectively.

[0049] S33. Based on the model validation and evaluation results, the Kriging model was selected as the domain proxy model.

[0050] S34. By integrating the white-box energy consumption mechanism model with the constraints of depth, oil bladder volume, and movable mass position, a domain-specific gray-box model is formed, and the domain-specific gray-box model is used to fit the segmented energy consumption.

[0051] Furthermore, the specific details of the domain-specific gray box model in S3 are as follows:

[0052] (11)

[0053] in, Indicates that the underwater gliding robot is in the i The minimum value of the total energy consumption in the domain; The underwater gliding robot is in the first i The minimum value of the changed oil bladder volume in the domain; , They represent the first time. i The minimum and maximum values ​​of the axial coordinates of the movable mass after it has been moved within the domain; Indicates the first i Kriging models in a domain; The underwater gliding robot is in the first i The changed oil bladder volume in the domain; Indicates the first i The axial coordinates of the movable mass after it has been moved within the domain; Indicates that it is subject to; This indicates the depth range of the underwater gliding robot during the entire anchoring energy consumption phase; This indicates the ground altitude range of the underwater glider during the entire anchoring energy consumption phase.

[0054] In another aspect, the present invention provides an energy consumption optimization system for the mooring process of an underwater glider, comprising an underwater glider, wherein the underwater glider is configured to or executes the above-described energy consumption optimization method for the mooring process of an underwater glider.

[0055] The beneficial effects of this invention are:

[0056] 1. This invention discloses an energy consumption optimization method for the mooring process of an underwater glider based on a domain gray box model. It internally designs a white box energy consumption mechanism model that covers the entire mooring process of "diving-leveling-fixed-point holding". The energy consumption calculated by the white box energy consumption mechanism model is more realistic.

[0057] 2. This invention combines a white-box energy consumption mechanism model with data-driven approaches, balancing physical interpretability with adaptability to complex characteristics, computational efficiency with modeling accuracy. It addresses the shortcomings of single modeling methods and the high dependence of black-box energy consumption mechanism models on sample size. Existing black-box energy consumption mechanism models suffer from excessively high costs in acquiring sea trial energy consumption data, and sea trial motion parameters cannot cover all gliding states. In contrast, the white-box energy consumption mechanism model established in this invention calculates energy consumption based on the ideal power of components, without considering the impact of load on component power.

[0058] 3. The present invention internally designs and uses a domain-specific gray box model, which performs segmented energy consumption fitting on each domain, thereby achieving independent and accurate fitting and joint optimization of the domain-specific gray box model.

[0059] Furthermore, the energy consumption obtained by fitting the data in this invention is closer to the actual energy consumption, and further optimization based on this data yields even better results. Attached Figure Description

[0060] Figure 1 A schematic diagram of the anchoring process of an underwater gliding robot;

[0061] Figure 2 The diagram shows the expected and response data of the oil bladder volume in the embodiment of the present invention, the traditional global gray box model, and the no-optimization scheme.

[0062] Figure 3 The diagram shows the expected and response data of the axial coordinates of the movable mass in the embodiments of the present invention, the traditional global gray box model, and the no-optimization scheme.

[0063] Figure 4 The diagram shows the depth and pitch angle response data of the domain gray box model, the traditional global gray box model, and the no-optimization scheme in the embodiments of the present invention.

[0064] Figure 5 This is a data graph showing the cumulative energy consumption of the domain gray box model, the traditional global gray box model, and the no-optimization scheme in the embodiments of the present invention;

[0065] Figure 6 This is a schematic diagram of the on-site marine anchoring experiment of the present invention;

[0066] Figure 7 This is a data graph showing the volume of the oil bladder and the location of its movable mass during the offshore anchoring experiment of this invention.

[0067] Figure 8 This is a data graph showing the depth, height, and pitch angle during the marine anchoring experiment of this invention. Detailed Implementation

[0068] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Preferred embodiments of the invention are shown in the drawings. However, the invention can be implemented in many other different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a thorough and complete understanding of the disclosure of the invention.

[0069] Reference Figure 1 This application provides an energy consumption optimization method for the anchoring process of an underwater glider based on a domain-specific gray box model, comprising the following steps:

[0070] S1. Based on the kinematic model and gliding energy consumption model of the underwater gliding robot, a white-box energy consumption mechanism model and constraints on the oil sac volume and movable mass position during the anchoring process of the underwater gliding robot are established. The kinematic model and gliding energy consumption model of the underwater gliding robot can be found in the paper (A. Jing, J. Gao, G. Pan, B. Song and C. Yang, "A Real-Time Updatable Gray-Box Energy Consumption Model for Underwater Gliders Considering Steady-State Dynamics Constraints," in IEEE Transactions on Industrial Electronics, vol. 72, no. 11, pp. 11704-11713, Nov. 2025, doi: 10.1109 / TIE.2025.3558024).

[0071] S2. Build a hardware-in-the-loop simulation platform. Divide the entire anchoring energy consumption stage of the underwater glider into multiple subdomains according to the different ranges of the oil bladder volume. Design the variable range constraints for each subdomain. Based on the white-box energy consumption mechanism model and using the Latin hypercube design method, use the hardware-in-the-loop simulation platform to simulate and generate multiple subdomain samples that meet the variable range constraints.

[0072] S3. The Kriging model is selected as the domain proxy model. The white box energy consumption mechanism model and the constraints of oil bladder volume and movable mass position are integrated to form a domain gray box model. The domain gray box model is then used to fit the segmented energy consumption.

[0073] S4. With the goal of minimizing the energy consumption fitted by the Domain-Specific Gray Box Model (DDGBM), the dynamic-guided self-regulating particle swarm algorithm is used to iteratively optimize the domain proxy model and output the optimal planning parameters.

[0074] In some embodiments, S1 specifically includes the following steps:

[0075] S11. The anchoring process of the underwater glider is divided into three stages: diving, leveling, and stationary holding.

[0076] S12. Based on the kinematic model of the underwater glider, establish the mapping relationship between the pitch angle and the fuel bladder volume and the movable mass; then construct the constraints on the fuel bladder volume and the movable mass position; where the constraints on the fuel bladder volume and the movable mass position refer to the constraints in formula (2). , , , ;

[0077] S13. Based on the kinematic model and gliding energy consumption model of the underwater gliding robot, establish a white-box energy consumption mechanism model of the underwater gliding robot during the anchoring process that satisfies the constraints and mapping relationships in S12.

[0078] Among them, the white-box energy consumption mechanism model is the sum of the energy consumption of the buoyancy adjustment system, pitch control system, integrated system and anchoring system on the underwater glider during the anchoring process; the integrated system is the control, measurement and communication system.

[0079] In some embodiments, the expression of the white-box energy consumption mechanism model is as follows:

[0080] (1)

[0081] in, This represents the total energy consumption of the underwater gliding robot during the anchoring process; This indicates the energy consumption of the buoyancy adjustment system on an underwater glider during the anchoring process; This indicates the energy consumption of the pitch control system on the underwater glider during the mooring process; This indicates the energy consumption of the integrated system on the underwater glider during the mooring process; This indicates the energy consumption of the anchoring system on an underwater glider during the anchoring process;

[0082] Energy consumption , , , The expressions are as follows:

[0083] (2)

[0084] (3)

[0085] (4)

[0086] (5)

[0087] in, , These represent the current depth as follows: The power and flow rate of the pump on the buoyancy regulation system; and These represent the oil bladder volumes before and after the change, respectively. and These represent the depth ranges representing the dynamic process of oil bladder volume during descent and leveling, respectively. and These represent the depths of the underwater glider before and after the change in the oil bladder volume during the leveling phase. It refers to the mass of the anchor in the anchoring system; It is the density of water; and These represent the power and flow rate of the solenoid valve in the buoyancy adjustment system, respectively. and These represent the axial coordinates of the movable mass before it moves. axial coordinates after movement The pitch angle, with the axis being the direction of movement of the movable mass; and These represent the time before and after the movable mass position is moved, respectively. This indicates that when the pitch angle is The power of the pitch control motor on the pitch control system; It is the power of the integrated system; and These represent the start and end times of the integrated system's operation, respectively. It is the power of the winch on the anchoring system; and These represent the start and end times of anchoring, respectively. Indicates time; This indicates the depth of the underwater gliding robot.

[0088] In some embodiments, the mapping relationship between the pitch angle and the oil bladder volume and movable mass in S12 is specifically expressed as follows:

[0089] (6)

[0090] in, It is movable mass; It is the radial coordinate of the movable mass before it moves; It refers to all mass on an underwater glider except for movable mass; and These are the axial and radial coordinates of the center of mass of the movable mass before it moves; It is net buoyancy; It is gravitational acceleration; and It is the hydrodynamic coefficient, and the radial direction mentioned above is the direction perpendicular to the axial direction.

[0091] In some embodiments, S2 specifically includes the following steps:

[0092] S21. Build a hardware-in-the-loop simulation platform and divide the entire underwater gliding robot anchoring energy consumption stage into four sub-domains according to the different ranges of oil bladder volume.

[0093] S22. Design the variable range constraints for each domain, as follows:

[0094] I. Constraints between depth and ground altitude;

[0095] II. Coupling constraints between the movable mass axial coordinate, oil bladder volume, and pitch angle;

[0096] III. Maximum and minimum constraints on the axial coordinates of movable mass;

[0097] S23. Based on the white-box energy consumption mechanism model and using the Latin hypercube design method, a hardware-in-the-loop simulation platform is used to simulate and generate global samples and multiple subdomain samples that satisfy the variable range constraints of each subdomain.

[0098] In some embodiments, the hardware-in-the-loop simulation platform includes a hardware-in-the-loop simulation platform centralized control system, a hardware-in-the-loop simulation platform main control system, and a hardware-in-the-loop simulation system that are interconnected.

[0099] The hardware-in-the-loop simulation platform's centralized control system includes a wireless communication module, a satellite communication module, a shore-based control system, a signal receiving module, and a host computer. The shore-based control system includes a display and control unit, an operation panel processing unit, and a first serial port communication module. The wireless module, satellite communication module, and signal receiving module are connected to the host computer via the first serial port communication module. The host computer downloads navigation tasks through the operation panel. The display and control unit is electrically connected to the host computer.

[0100] The hardware-in-the-loop simulation platform's main control system consists of a main control board, peripherals, a second serial communication module, a CAN bus communication module, and multiple sensors. Peripherals include an oil bladder, movable mass, a dwell system, and servos. The multiple sensors include a depth gauge, altimeter, rangefinder, CTD (Conductivity, Temperature, Depth) meter, and TCM (Attitude, Measurement, and Communication) sensor. The CAN bus communication module is electrically connected to the main control board via the second serial communication module. The dwell system simulates the anchoring system on an underwater glider. The multiple sensors and servos are all electrically connected to the main control board.

[0101] The hardware-in-the-loop (HIL) simulation platform includes a simulation system (simulation computer) running the simulation, a kinematic model of the underwater glider, a communication module on the buoyancy adjustment system under normal conditions, a simulation interface server, a signal transmission module, and a signal conversion unit. The main control system of the HIL simulation platform inputs data into the simulation computer via the signal transmission module and transmits simulation data back to the main control system for calculation via the transmission module. The communication module, simulation interface server, signal transmission module, and signal conversion unit are all electrically connected to the simulation computer, which contains a built-in kinematic model of the underwater glider.

[0102] In some embodiments, the expression for the constraint between depth and ground altitude in S22 is as follows:

[0103] (7)

[0104] in, Indicates the first i The depth of the underwater glider in the sub-domain; Indicates the first i The ground altitude of the underwater glider in the sub-domain; This represents the maximum value of the sum of the set depth and the altitude above the ground. In this embodiment, The value is 100.

[0105] The expression for the coupling constraint between the movable mass axial coordinate, oil bladder volume, and pitch angle is as follows:

[0106] (8)

[0107] in, Indicates the first i The axial coordinates of the movable mass in the domain before it moves; Indicates the first i The volume of the oil sac before the change in the domain;

[0108] The expressions for the maximum and minimum constraints of the axial coordinates of the movable mass are as follows:

[0109] (9)

[0110] (10)

[0111] in, , They represent the first time. i The maximum and minimum values ​​of the axial coordinates of the movable mass in the domain before it moves; , These represent the maximum and minimum pitch angles of the underwater glider, respectively. , They represent the first time. i The maximum and minimum values ​​of the oil bladder volume before the change in the domain; , These represent the maximum and minimum values ​​of the axial coordinates of the movable mass before it moves during the entire anchoring energy consumption phase.

[0112] In some embodiments, each domain sample in S3 includes basic parameters and planning parameters, wherein the basic parameters include the depth of the underwater glider before diving. Bottom height Oil sac volume before change And the axial coordinates of the movable mass before it moves. ;

[0113] Planning parameters include the changed oil bladder volume And the axial coordinates after the movable mass has moved. .

[0114] The first domain sample in S2 is shown in Table 1.

[0115] Table 1: First domain sample data:

[0116]

[0117] In some embodiments, S3 specifically includes the following steps:

[0118] S31. Initially select one of the following methods as the domain proxy model: Response Surface Method (RSM), Radial Basis Function Model (RBF), or Kriging Model;

[0119] S32. Based on the samples in each subdomain, the response surface method, radial basis function model, and Kriging model are evaluated and validated using median absolute deviation (MAD), maximum absolute error (MAE), root mean square error (RMSE), and coefficient of determination (R-square).

[0120] S33. Based on the model validation and evaluation results, the Kriging model was selected as the domain proxy model.

[0121] The specific results of model validation and evaluation are shown in Table 2.

[0122] Table 2: Model Validation Evaluation Results;

[0123]

[0124] Based on the model validation and evaluation results in Table 2, the Kriging model demonstrates a more significant advantage across multiple datasets. Its mean absolute deviation, maximum absolute error, and root mean square error are mostly at low levels, and its coefficient of determination approaches 1. It exhibits strong explanatory power for data variation, optimal stability and fitting accuracy, and is suitable for complex, high-dimensional scenarios with varying fluctuation characteristics. Overall, the Kriging model demonstrates the best accuracy and stability across multiple scenarios; therefore, it was selected as the domain-specific surrogate model.

[0125] S34. By integrating the white-box energy consumption mechanism model with the constraints of depth, oil bladder volume, and movable mass position, a domain-specific gray-box model is formed, and the domain-specific gray-box model is used to fit the segmented energy consumption.

[0126] In some embodiments, the domain-based gray box model in S3 is specifically as follows:

[0127] (11)

[0128] in, The underwater gliding robot is in the first i The minimum value of the total energy consumption in the domain; The underwater gliding robot is in the first i The minimum value of the changed oil bladder volume in the domain; , They represent the first time. i The minimum and maximum values ​​of the axial coordinates of the movable mass after it has been moved within the domain; Indicates the first i Kriging models in a domain; The underwater gliding robot is in the first i The changed oil bladder volume in the domain; Indicates the first i The axial coordinates of the movable mass after it has been moved within the domain; Indicates that it is subject to; This indicates the depth range of the underwater gliding robot during the entire anchoring energy consumption phase; This indicates the ground altitude range of the underwater glider during the entire anchoring energy consumption phase.

[0129] In some embodiments, S4 specifically includes the following steps:

[0130] S41. Based on the STM32F429 microcontroller with ARM Cortex-M4 32-bit RISC core, three typical heuristic optimization algorithms are developed: Real Number Encoded Genetic Algorithm (RCGA), Dynamically Guided Self-Regulating Particle Swarm Optimization (DMeSR-PSO), and Curiosity Simulated Annealing (CSA).

[0131] S42. Three typical heuristic optimization algorithms were used to optimize and compare several domains respectively. The optimization results are shown in Table 3.

[0132] Table 3: Optimization results of three typical heuristic optimization algorithms;

[0133]

[0134] Table 3 shows that all three typical heuristic optimization algorithms achieved significant energy savings. After optimization of the third domain sample data, the energy consumption was reduced to 23.42 kW·s, representing energy savings of 35.5% and 12.9% respectively compared to no optimization and full-domain optimization, demonstrating the adaptability advantage of the domain gray-box model to segmented energy consumption characteristics. In terms of operational efficiency, the dynamic-guided self-regulating particle swarm optimization algorithm and the curiosity simulated annealing algorithm had a running time of less than 0.08 s, which is more than 46.7% shorter than the real-number encoded genetic algorithm, making them more suitable for the real-time requirements of the embedded control system of the underwater glider. Therefore, the dynamic-guided self-regulating particle swarm optimization algorithm was selected as the optimization algorithm for subsequent simulation verification and sea trial phases.

[0135] S43. With the goal of minimizing the energy consumption fitted by the domain gray box model, the dynamic-guided self-adjusting particle swarm algorithm is used to iteratively optimize the domain proxy model and output the optimal planning parameters.

[0136] This invention also conducted comparative experiments on the proposed Domain-Specific Gray Box Model (DDGBM), the traditional Global Gray Box Model (FDGBM), and a no-optimization scheme. The comparative data can be found in [link to relevant documentation]. Figures 2 to 5 As shown.

[0137] Comparative experiment Set the initial base parameters. The anchor chain is 2 m long and 50 m deep; the mooring process simulation is run on a hardware-in-the-loop (HIL) simulation platform. Figures 2 to 5The dynamic response characteristics of core state variables during mooring were quantitatively compared between the Domain-Specific Gray Box Model (DDGBM), the Global-Specific Gray Box Model (FDGBM), and the no-optimization scheme, covering oil bladder volume, movable mass position, and depth-pitch angle. The scheme using the Domain-Specific Gray Box Model (DDGBM) maintained the initial oil bladder volume until a depth of -35.4 m, then increased it to -1.9 L. The axial coordinate of the movable mass was increased from its initial value to 70 mm and then held. The pitch angle stabilized at -26.4° after initial oscillations until a depth of -35.4 m was reached, at which point the oil bladder volume was adjusted, and the pitch angle decreased to -28.4°, taking 183.6 s and consuming 138.93 kW·s.

[0138] In the Full-Domain Gray Box Model (FDGBM) scheme, at depths greater than -19.8 m, the oil bladder volume increases to -2.04 L, the axial coordinate of the movable mass increases from its initial value to 70 mm and then remains constant, and the pitch angle, after initial oscillations, stabilizes at -27.3° and then slowly decreases to 28.1°. During the descent phase at depths less than -19.8 m, the volume briefly decreases from -2.04 L to -2.06 L before continuously increasing to -1.60 L, the movable mass position decreases to 26.3 mm, and the pitch angle increases to -17.3°. The total time taken is 224.8 s, and the total energy consumption is 174.03 kW·s.

[0139] Without optimization, the initial parameters were maintained during the descent to the leveling height, and then the oil bladder volume was adjusted to -1 L. The axial coordinate of the movable mass was adjusted to 23.05 mm. The pitch angle during the descent was maintained at -22.3°. The descent time was 243.1 s, and the energy consumption was 174.01 kW·s.

[0140] The difference in planning parameter responses between the global gray box model (FDGBM) and the local gray box model (DDGBM) is essentially due to the different adaptability of the two models to the "segmented characteristics of energy consumption models". Compared with the global gray box model (FDGBM) and the no-optimization scheme, the local gray box model (DDGBM) reduces the anchoring process time by 18.3% and 24.5% respectively, and reduces energy consumption by 20.2% in both cases, showing a significant advantage.

[0141] In addition, this invention also underwent offshore anchoring tests, and the experimental data are available in […]. Figures 6 to 8 As shown. Specifically, the offshore anchoring trials were conducted in the South China Sea, with the deployment site as follows: Figure 6 As shown. The initial basic parameters are... ,set up The anchor chain is 2 m long. See Figure 7 and Figure 8As shown, the initial oil bladder volume was maintained until the leveling height, then adjusted to -1 L. The movable mass position was increased from the initial value to 70 mm and then maintained. When the pitch angle reached -21.84°, the underwater glider began to descend. During the descent, the pitch angle oscillated initially before stabilizing at -24.64°. At 503 seconds, the leveling phase began, and the pitch angle oscillated and increased to -19.74°. At 714 seconds, the anchor touched the bottom, the pitch angle increased to -13.16°, then slowly decreased and stabilized at -17.78°. The anchoring process ended at this point, with an anchoring depth of -87.9 m, an anchoring height of 1.68 m, and an energy consumption of 214.82 kW·s. Based on the depth and height data, the longitudinal profile of the seabed can be calculated, demonstrating that the proposed method remains effective for seabeds with varying elevations.

[0142] In another aspect, the present invention provides an energy consumption optimization system for the mooring process of an underwater glider, comprising an underwater glider, wherein the underwater glider is configured to or executes the above-described energy consumption optimization method for the mooring process of an underwater glider; the underwater glider includes:

[0143] The buoyancy adjustment system is installed on the underwater glider. The buoyancy adjustment system mainly includes an oil bladder, which is used to adjust the volume of the oil bladder through a solenoid valve, thereby adjusting the buoyancy of the underwater glider.

[0144] The pitch adjustment system is located at the head of the underwater glider. The pitch adjustment system mainly includes a movable mass, which is used to adjust the position of the movable mass, thereby realizing the adjustment of the pitch angle of the underwater glider.

[0145] The integrated system includes an integrated circuit board, which has a built-in control module, measurement module, and communication module.

[0146] An anchoring system, installed on an underwater glider, includes a winch, an anchor chain wound around the winch, and an anchor attached to the anchor chain.

[0147] The anchoring process is as follows:

[0148] Before diving, the underwater glider floated on the surface of the water, with an oil sac volume of [missing information]. The axial coordinate of the movable mass is in The underwater glider is controlled to ensure its stern antenna, connected to the integrated system, surfaces to enable communication. During the descent phase, the glider activates its buoyancy adjustment system. A solenoid valve on this system opens, allowing hydraulic oil to flow from the external oil bladder into the internal cavity, reducing the bladder's volume and decreasing the glider's buoyancy. The pitch adjustment system adjusts the axial coordinates of the movable mass to control the glider's descent angle (pitch angle). This continues until the glider reaches the set leveling height. (Generally judged based on the height from the bottom) when entering the leveling stage.

[0149] Once the leveling phase begins, the oil bladder volume is adjusted to within the mooring volume range if necessary (depending on the anchor mass and water density), and the axial coordinate of the movable mass is controlled to ensure the pitch angle meets mooring requirements. Simultaneously, the anchoring system begins releasing the anchor chain according to the set mooring height. Once the anchor touches the bottom and the anchor chain has been fully released, the mooring phase begins. At this point, the mooring process is complete.

[0150] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Furthermore, the technical solutions of the various embodiments of the present invention can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for optimizing energy consumption during the anchoring process of an underwater glider based on a domain-specific gray box model, characterized in that, Includes the following steps: S1. Based on the kinematic model and gliding energy consumption model of the underwater gliding robot, establish a white-box energy consumption mechanism model and constraints on the oil bladder volume and movable mass position during the anchoring process of the underwater gliding robot. S2. Build a hardware-in-the-loop simulation platform. Divide the entire anchoring energy consumption stage of the underwater glider into multiple subdomains according to the different ranges of the oil bladder volume. Design the variable range constraints for each subdomain. Based on the white-box energy consumption mechanism model and using the Latin hypercube design method, use the hardware-in-the-loop simulation platform to simulate and generate multiple subdomain samples that meet the variable range constraints. S3. The Kriging model is selected as the domain proxy model. The white box energy consumption mechanism model and the constraints of oil bladder volume and movable mass position are integrated to form a domain gray box model. The domain gray box model is then used to fit the segmented energy consumption. S4. With the goal of minimizing the energy consumption fitted by the domain gray box model, the dynamic-guided self-regulating particle swarm algorithm is used to iteratively optimize the domain proxy model and output the optimal planning parameters. S3 specifically includes the following steps: S31. Initially select one of the response surface method, radial basis function model, and Kriging model as the domain proxy model; S32. Based on the samples in each sub-domain, the response surface method, radial basis function model, and kriging model are validated and evaluated using median absolute deviation, maximum absolute error, root mean square error, and coefficient of determination, respectively. S33. Based on the model validation and evaluation results, the Kriging model was selected as the domain proxy model. S34. By integrating the white-box energy consumption mechanism model with the constraints of depth, oil bladder volume, and movable mass position, a domain-specific gray-box model is formed, and the domain-specific gray-box model is used to fit the segmented energy consumption.

2. The energy consumption optimization method for the anchoring process of an underwater gliding robot based on a domain-specific gray box model according to claim 1, characterized in that, S1 specifically includes the following steps: S11. The anchoring process of the underwater glider is divided into three stages: diving, leveling, and stationary holding. S12. Based on the kinematic model of the underwater glider, establish the mapping relationship between the pitch angle and the fuel bladder volume and the movable mass; then construct the constraints on the fuel bladder volume and the position of the movable mass. S13. Based on the kinematic model and gliding energy consumption model of the underwater gliding robot, establish a white-box energy consumption mechanism model of the underwater gliding robot during the anchoring process that satisfies the constraints and mapping relationships in S12. Among them, the white-box energy consumption mechanism model is the sum of the energy consumption of the buoyancy adjustment system, pitch control system, integrated system and anchoring system on the underwater glider during the anchoring process; the integrated system is the control, measurement and communication system.

3. The energy consumption optimization method for the anchoring process of an underwater gliding robot based on a domain-specific gray box model according to claim 2, characterized in that, The specific expression of the white-box energy consumption mechanism model is as follows: （1） in, This represents the total energy consumption of the underwater gliding robot during the anchoring process; This indicates the energy consumption of the buoyancy adjustment system on an underwater glider during the anchoring process; This indicates the energy consumption of the pitch control system on the underwater glider during the mooring process; This indicates the energy consumption of the integrated system on the underwater glider during the mooring process; This indicates the energy consumption of the anchoring system on an underwater glider during the anchoring process; Energy consumption , , , The expressions are as follows: （2） （3） （4） （5） in, , These represent the current depth as follows: The power and flow rate of the pump on the buoyancy regulation system; and These represent the oil bladder volumes before and after the change, respectively. and These represent the depth ranges representing the dynamic process of oil bladder volume during descent and leveling, respectively. and These represent the depths of the underwater glider before and after the change in the oil bladder volume during the leveling phase. It refers to the mass of the anchor in the anchoring system; It is the density of water; and These represent the power and flow rate of the solenoid valve in the buoyancy adjustment system, respectively. and These represent the axial coordinates of the movable mass before it moves. axial coordinates after movement The pitch angle, with the axis being the direction of movement of the movable mass; and These represent the time before and after the movable mass position is moved, respectively. This indicates that when the pitch angle is The power of the pitch control motor on the pitch control system; It is the power of the integrated system; and These represent the start and end times of the integrated system's operation, respectively. It is the power of the winch on the anchoring system; and These represent the start and end times of anchoring, respectively. Indicates time; This indicates the depth of the underwater gliding robot.

4. The energy consumption optimization method for the anchoring process of an underwater gliding robot based on a domain-specific gray box model according to claim 3, characterized in that, The mapping relationship between the pitch angle and the oil bladder volume and movable mass in S12 is expressed as follows: （6） in, It is movable mass; It is the radial coordinate of the movable mass before it moves; It refers to all mass on an underwater glider except for movable mass; and These are the axial and radial coordinates of the center of mass of the movable mass before it moves; It is net buoyancy; It is gravitational acceleration; and It is the hydrodynamic coefficient, and the radial direction mentioned above is the direction perpendicular to the axial direction.

5. The energy consumption optimization method for the anchoring process of an underwater gliding robot based on a domain-specific gray box model according to claim 4, characterized in that, S2 specifically includes the following steps: S21. Build a hardware-in-the-loop simulation platform and divide the entire underwater gliding robot anchoring energy consumption stage into four sub-domains according to the different ranges of oil bladder volume. S22. Design the variable range constraints for each domain, as follows: I. Constraints between depth and ground altitude; II. Coupling constraints between the movable mass axial coordinate, oil bladder volume, and pitch angle; III. Maximum and minimum constraints on the axial coordinates of movable mass; S23. Based on the white-box energy consumption mechanism model and using the Latin hypercube design method, a hardware-in-the-loop simulation platform is used to simulate and generate global samples and multiple subdomain samples that satisfy the variable range constraints of each subdomain.

6. The energy consumption optimization method for the anchoring process of an underwater gliding robot based on a domain-specific gray box model according to claim 5, characterized in that, The expression for the constraint between depth and ground altitude in S22 is as follows: （7） in, Indicates the first i The depth of the underwater glider in the sub-domain; Indicates the first i The ground altitude of the underwater glider in the sub-domain; This represents the maximum value of the sum of the set depth and the altitude above the ground. The expression for the coupling constraint between the movable mass axial coordinate, oil bladder volume, and pitch angle is as follows: （8） in, Indicates the first i The axial coordinates of the movable mass in the domain before it moves; Indicates the first i The volume of the oil sac before the change in the domain; The expressions for the maximum and minimum constraints of the axial coordinates of the movable mass are as follows: （9） （10） in, , They represent the first time. i The maximum and minimum values ​​of the axial coordinates of the movable mass in the domain before it moves; , These represent the maximum and minimum pitch angles of the underwater glider, respectively. , They represent the first time. i The maximum and minimum values ​​of the oil bladder volume before the change in the domain; , These represent the maximum and minimum values ​​of the axial coordinates of the movable mass before it moves during the entire mooring energy consumption phase.

7. The energy consumption optimization method for the anchoring process of an underwater glider based on a domain-specific gray box model according to claim 6, characterized in that, Each sub-sample in S3 contains basic parameters and planning parameters, where the basic parameters include the depth of the underwater glider before diving. Bottom height Oil sac volume before change And the axial coordinates of the movable mass before it moves. ; Planning parameters include the changed oil bladder volume And the axial coordinates after the movable mass has moved. .

8. The energy consumption optimization method for the anchoring process of an underwater glider based on a domain-specific gray box model according to claim 7, characterized in that, The specific domain gray box model in S3 is as follows: （11） in, The underwater gliding robot is in the first i The minimum value of the total energy consumption in the domain; Indicates that the underwater gliding robot is in the i The minimum value of the changed oil bladder volume in the domain; , They represent the first time. i The minimum and maximum values ​​of the axial coordinates of the movable mass after it has been moved within the domain; Indicates the first i Kriging models in a domain; The underwater gliding robot is in the first i The changed oil bladder volume in the domain; Indicates the first i The axial coordinates of the movable mass after it has been moved within the domain; Indicates that it is subject to; This indicates the depth range of the underwater gliding robot during the entire anchoring energy consumption phase; This indicates the ground altitude range of the underwater glider during the entire anchoring energy consumption phase.

9. An energy consumption optimization system for the anchoring process of an underwater glider robot, characterized in that, Includes an underwater glider, which is configured to perform the energy consumption optimization method for the anchoring process of the underwater glider as described in any one of claims 1 to 8.

Images