Energy modeling and energy saving obstacle avoidance method and apparatus for natural energy capturing marine robots

Abstract

This invention discloses a method and device for energy modeling and energy-saving obstacle avoidance of marine robots that considers natural energy capture, belonging to the field of marine robot technology. The method first acquires the marine robot's mission information, navigation environment information, and its own motion state information; then, it constructs an energy modeling system including a speed response model, an energy consumption model, and a natural energy capture model; next, it introduces effective speed constraints under the influence of natural energy to construct a dynamic speed window; within the dynamic speed window, it samples candidate speed combinations and generates candidate trajectories by combining them with a kinematic model; it scores the predicted trajectories using a comprehensive evaluation function, selects the speed combination with the best score as the robot's motion control command for the next moment; and iteratively executes the above steps until the robot reaches the target point or the mission termination condition is met. This method improves the utilization of natural energy and enhances the endurance and autonomous navigation capabilities in complex time-varying environments while ensuring the marine robot's safe navigation and effective obstacle avoidance.

Description

Technical Field

[0001] This invention belongs to the field of marine robot technology, specifically relating to energy modeling and energy-saving obstacle avoidance methods and equipment for marine robots that take into account natural energy capture. Background Technology

[0002] As an autonomous marine vehicle, marine robots can perform tasks such as marine resource exploration, natural disaster early warning, marine environmental monitoring, and long-term maritime patrols. Due to the complex marine operating environment, long navigation distances, and difficulty in artificial refueling, energy supply and endurance have always been significant factors restricting the long-term autonomous operation of marine robots. For marine robots equipped with photovoltaic panels, sails, and hydrofoils / wave plates, they can capture solar energy and convert it into electricity through photovoltaic panels, utilize wind power for auxiliary propulsion through sails, and improve navigation response by utilizing waves or hydrodynamics through hydrofoils / wave plates, thereby enhancing their natural energy utilization capabilities and endurance.

[0003] However, during actual navigation, the motion state and energy expenditure of such marine robots are influenced by a combination of factors, including sea winds, waves, currents, solar radiation intensity, hull heading, attitude angles, sail angles, and the shading effect of photovoltaic panels. Their speed response is not only related to the propulsion system but also to the direction of wind, waves, currents, and hydrodynamic forces; their energy capture capability is not only related to solar radiation intensity but is also affected by the arrangement of photovoltaic panels, sail shading, and changes in hull attitude. Therefore, without modeling the coupling relationship between energy consumption, natural energy capture, and motion response, traditional obstacle avoidance methods cannot accurately evaluate the net energy consumption and energy utilization efficiency of different candidate trajectories.

[0004] The invention patent application CN117572870A, entitled "Dynamic Obstacle Avoidance Method for Unmanned Surface Vessels," achieves dynamic obstacle avoidance by adjusting the relationship between relative speed and the tangent of the obstacle circle, thereby adjusting the speed and heading of the unmanned surface vessel. However, this method does not consider the optimal utilization of ocean energy and is not suitable for obstacle avoidance by marine robots. The invention patent CN119759026A, entitled "Multi-Source Energy-Saving Obstacle Avoidance Control Method and System for Ocean Energy-Driven Robots," involves robots that utilize wind energy differently, resulting in variations in energy calculations. Furthermore, the evaluation function coefficient set involved in the dynamic window method is difficult to determine. Currently, different types of marine robots have different movement patterns, and their obstacle avoidance capabilities are limited compared to unmanned surface vessels. Indiscriminately applying traditional obstacle avoidance methods to marine robots may lead to obstacle avoidance failures. Secondly, marine robots have the advantage of utilizing ocean energy; maximizing this advantage while avoiding obstacles is the optimal obstacle avoidance strategy.

[0005] To address this, this invention proposes an energy modeling and energy-saving obstacle avoidance method for marine robots that considers natural energy capture. This method, designed for marine robots equipped with photovoltaic panels, sails, and hydrofoils / wave-force winglets, establishes an energy modeling system including a speed response model, an energy consumption model, and a natural energy capture model. This system characterizes the robot's mobility and energy expenditure under different sea states, headings, attitudes, and solar radiation conditions. The energy consumption, natural energy capture, safe obstacle avoidance, and target approach capabilities of candidate trajectories are jointly incorporated into a dynamic window method evaluation function. Simultaneously, a metaheuristic optimization algorithm is used to adaptively optimize the weight parameters of the evaluation function, generating an optimal speed control strategy that balances safety, maneuverability, and energy efficiency, thereby achieving energy-saving obstacle avoidance and navigation optimization in dynamic environments. This invention maximizes the utilization of natural energy while ensuring safe obstacle avoidance, improving the autonomous navigation capability and endurance of marine robots in complex sea areas, and providing reliable technical support for long-term autonomous operations and marine resource exploration. Summary of the Invention

[0006] The purpose of this invention is to address the challenge of establishing an energy model that reflects the coupling relationship between energy consumption, natural energy capture, and motion response when marine robots equipped with photovoltaic panels, sails, and hydrofoils / wave-driven winglets navigate in time-varying marine environments, while safely avoiding unknown obstacles, and to achieve energy-efficient obstacle avoidance based on this model. Existing obstacle avoidance methods typically do not adequately consider the impact of natural energy on the robot's mobility and energy budget. For marine robots with solar energy capture, wind-assisted propulsion, and wave / hydrodynamic-assisted propulsion capabilities, their operating modes and dynamic characteristics differ from traditional unmanned surface vessels or general surface platforms, resulting in limited adaptability of traditional methods for energy-efficient obstacle avoidance. Therefore, this invention provides an energy modeling and energy-efficient obstacle avoidance method for marine robots that considers natural energy capture. This method accurately reflects the robot's mobility and energy budget in complex time-varying environments through a velocity response model, an energy consumption model, and a natural energy capture model. It also utilizes a metaheuristic optimization algorithm to adaptively optimize the parameters of the dynamic window method evaluation function, thereby generating a safe and energy-efficient optimal velocity control strategy, achieving energy-efficient obstacle avoidance and navigation optimization in dynamic environments.

[0007] This invention provides a method for energy modeling and energy-saving obstacle avoidance of marine robots that considers natural energy capture, comprising the following steps:

[0008] Acquire information about the marine robot's mission, navigation environment, and its own motion status;

[0009] An energy modeling system is constructed, which includes a velocity response model, an energy consumption model, and a natural energy capture model; the natural energy capture model is used to calculate the solar power capture after considering the correction for wind and sail obstruction.

[0010] An effective speed constraint under the influence of natural energy is introduced to construct a dynamic speed window; candidate speed combinations are sampled within the dynamic speed window, and candidate trajectories are generated by combining them with a kinematic model.

[0011] The trajectory is segmented based on the maximum collision avoidance distance from the obstacle. The candidate trajectory's speed, heading, distance from the target point, distance from the obstacle, and net energy consumption are used as evaluation indicators to construct a comprehensive evaluation function.

[0012] After acquiring multiple sets of predicted trajectories, the predicted trajectories are scored using a comprehensive evaluation function. The velocity combination with the best score is selected as the motion control command for the robot at the next moment. This process is repeated until the robot reaches the target point or the task termination condition is met.

[0013] Furthermore, the velocity response model is specifically as follows:

[0014] Step 1.1: The robot adjusts the sail angle in real time according to the current wind direction angle to obtain the maximum auxiliary thrust under the current working conditions, and obtains the corresponding optimal speed through computational fluid dynamics (CFD) numerical simulation.

[0015]

[0016] in, Indicates sea state L, the CFD simulation of ship speed under different wind angles; This represents the CFD numerical simulation solution function; Indicates the first Group wind direction angle; Indicates the first The encounter angle is the optimal sail angle that allows the robot to obtain the maximum auxiliary thrust.

[0017] Step 1.2: Obtain the continuous speed response curve using a piecewise cubic conformal interpolation method. Furthermore, the truncated Fourier cosine series method is used for fitting, generating a function that can be called in real time using the dynamic window method. ;

[0018] Step 1.3: Perform linear interpolation of the speed between adjacent sea state classes to calculate the wind speed introduced from the sea surface. Post-continuous speed prediction function ;

[0019]

[0020] in, and These represent the wind speed boundaries corresponding to two adjacent sea state levels; and These represent the velocity response functions under two adjacent sea states, respectively.

[0021] Step 1.4: Superimpose the ocean current effect onto the wind and wave speed response model;

[0022]

[0023]

[0024] in, and These represent the speed vector and scalar respectively under the combined influence of wind, waves, and current during surface navigation. and These represent the ocean current velocity vector and scalar, respectively. Indicates the relative direction angle of ocean currents. Indicates the absolute direction angle of ocean currents. Indicates the heading angle.

[0025] Furthermore, the energy consumption model The output power of environmental perception, communication, intelligent control, and actuator is calculated jointly.

[0026] Furthermore, the natural energy capture model for:

[0027]

[0028]

[0029] in, Indicates the intensity of solar radiation; This represents the total area of ​​actual effective illumination; Indicates the light energy conversion coefficient; This indicates the effective light-receiving area of ​​the photovoltaic panels on the sail. This represents the effective light-receiving area of ​​the photovoltaic panels, taking into account the shading effect of the sails on the deck photovoltaic panels. This indicates the solar altitude angle.

[0030] Furthermore, the effective light-irradiated area of ​​the photovoltaic panel considering the shading effect of the sail on the deck photovoltaic panel... Calculate using the following method:

[0031]

[0032] in, and This indicates the width and length of the photovoltaic panels on the ship's hull. This represents the angle of incidence of sunlight, that is, the angle between sunlight and the x-axis of the ship's coordinate system. Indicates the heading angle. Indicates the solar azimuth angle; This represents the angle of sunlight incidence when the upper boundary of the shadow cast by the sail on the deck coincides with the diagonal of the photovoltaic rectangle on the deck.

[0033] The effective light-emitting area of ​​the photovoltaic panels on the sail for:

[0034]

[0035] in, and This indicates the width and length of the photovoltaic panels on the sail surface; Indicates the sail's angle.

[0036] Furthermore, the speed dynamic window includes speed boundary constraints, acceleration constraints, obstacle constraints, and effective speed constraints under the influence of natural energy; the speed constraints are:

[0037]

[0038] in, and These represent the basic minimum and maximum linear velocities of the marine robot, determined by its own motion capabilities, respectively. and These represent the minimum and maximum angular velocities of the marine robot, respectively.

[0039] The acceleration constraint is:

[0040]

[0041] in, and These represent the robot's current linear velocity and angular velocity, respectively. and These represent the robot's maximum linear acceleration and angular acceleration, respectively.

[0042] The obstacle distance constraint is:

[0043]

[0044] in, This represents the closest distance between the simulated trajectory corresponding to the combination of the robot's linear velocity and angular velocity and the obstacle.

[0045] The effective speed constraint under the influence of natural energy is:

[0046]

[0047]

[0048] in, This represents the minimum linear velocity within the dynamic window of speed after effective speed correction by natural energy.

[0049] Furthermore, the candidate velocity combination Candidate trajectories are generated by combining the kinematic model, which is as follows:

[0050]

[0051] in, and They represent the first Candidate trajectories in The predicted location coordinates at that moment; Indicates the prediction step time interval; and They represent the first The position coordinates of each candidate trajectory at the next prediction time step; Indicates the first Total output power of the robot for each candidate trajectory; Indicates the first The cumulative net energy consumption of the candidate trajectories at time t; Indicates the first The cumulative net energy consumption of the candidate trajectories; Indicates the first Candidate linear velocities corresponding to each candidate trajectory; Indicates the first The linear velocity of each candidate trajectory at the next prediction time step; Indicates the first The headings corresponding to the candidate trajectories; Indicates the first The heading of each candidate trajectory at the next prediction time step; Indicates the first The effective speed of each candidate trajectory under the influence of natural energy in the next prediction time step; Indicates the first The candidate angular velocities corresponding to the candidate trajectories; Indicates the first The angular velocity of each candidate trajectory at the next prediction time step; Indicates the relative direction of ocean currents; Relative wind direction angle; This indicates the current wind speed at sea.

[0052] Furthermore, the comprehensive evaluation function for:

[0053]

[0054]

[0055]

[0056]

[0057]

[0058]

[0059] in, Indicates the maximum collision avoidance distance of the obstacle; , , , , , , , , and These are the weight coefficients for the evaluation items; This indicates the total number of candidate trajectories generated within the current speed dynamic window; This represents the traversal index during normalized summation, used to traverse all candidate trajectories. ; Indicates the first The heading indicator of the trajectory; This indicates the angular deviation between the trajectory's final heading and the direction of the line connecting the predicted point and the target point; This represents the heading evaluation value of the corresponding trajectory after normalization. Indicates the first Distance metrics for each trajectory; No. The predicted position coordinates of the candidate trajectories at the end of the prediction time window; Indicates the coordinates of the target point; This represents the target distance evaluation value of the corresponding trajectory after normalization. Indicates the first The speed index corresponding to each trajectory; This represents the speed evaluation value of the corresponding trajectory after normalization. Indicates the first Energy indicators for each trajectory; This represents the energy evaluation value of the corresponding trajectory after normalization. Indicates the first Distance indicators between candidate trajectories and obstacles; Indicates the first The first candidate trajectory The coordinates of each predicted sampling point; Indicates the coordinates of the obstacle; This represents the obstacle distance evaluation value after normalization.

[0060] Furthermore, the comprehensive evaluation function adopts a metaheuristic optimization algorithm for adaptive optimization, using the total net energy consumption of the candidate trajectory within the prediction time window and the minimum distance between it and the obstacle as the fitness function, and obtaining the evaluation function weights for optimal energy utilization through algorithm iteration.

[0061] Furthermore, the fitness function is designed in segments based on the maximum collision avoidance distance of the obstacle: when the minimum distance between the candidate trajectory and the obstacle is less than or equal to the maximum collision avoidance distance, the fitness function introduces a dimensionless maximum collision avoidance distance correction term based on the reciprocal of the net energy consumption of the candidate trajectory; when the minimum distance between the candidate trajectory and the obstacle is greater than the maximum collision avoidance distance, the fitness function uses the reciprocal of the energy consumption of the candidate trajectory within the current prediction time window.

[0062] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the above-described method for energy modeling and energy-saving obstacle avoidance of marine robots that takes into account natural energy capture.

[0063] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method for energy modeling and energy-saving obstacle avoidance of marine robots considering natural energy capture.

[0064] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method for energy modeling and energy-saving obstacle avoidance of marine robots considering natural energy capture.

[0065] The beneficial effects of this invention are as follows:

[0066] This invention provides an energy modeling and energy-saving obstacle avoidance method for marine robots that considers natural energy capture, applicable to marine robots equipped with photovoltaic panels, sails, and hydrofoils. While ensuring the robot can safely avoid unknown obstacles, this invention establishes a velocity response model, an energy consumption model, and a natural energy capture model to accurately characterize the robot's motion response, energy consumption, and natural energy utilization under different sea states, headings, attitudes, and solar radiation conditions. This method uses energy consumption, natural energy capture, target approach capability, and obstacle collision avoidance safety during navigation as comprehensive evaluation indicators in a dynamic window method. A metaheuristic optimization algorithm is used to adaptively optimize the evaluation function parameters to obtain the optimal combination of control parameters, thereby determining the optimal velocity control quantity for the robot and achieving energy-saving obstacle avoidance and navigation optimization in dynamic environments. This method not only ensures the robot's safe navigation in complex and time-varying sea areas but also improves the comprehensive utilization efficiency of natural energy sources such as solar, wind, and wave / hydrodynamic forces, reduces net energy consumption during obstacle avoidance, and enhances the endurance, reliability, and continuous operation capability of marine robots in long-term autonomous navigation missions. Attached Figure Description

[0067] Figure 1 This is a flowchart illustrating the principle of the marine robot energy modeling and energy-saving obstacle avoidance method that considers natural energy capture, as described in this invention.

[0068] Figure 2 This is a flowchart illustrating the construction of the velocity response model for the marine robot energy modeling and energy-saving obstacle avoidance method that considers natural energy capture, as described in this invention.

[0069] Figure 3 This is a schematic diagram of sail obstruction in the marine robot energy modeling and energy-saving obstacle avoidance method that considers natural energy capture in this invention.

[0070] Figure 4 This is a schematic diagram of the sail obstruction angle analysis of the marine robot energy modeling and energy-saving obstacle avoidance method that considers natural energy capture in this invention.

[0071] Figure 5 This is a flowchart of the metaheuristic optimization algorithm (taking genetic algorithm as an example) for the energy modeling and energy-saving obstacle avoidance method of marine robots that considers natural energy capture in this invention. Detailed Implementation

[0072] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0073] The marine robot in this invention refers to a robot that can store electrical energy through solar panels and use sails and wave-assisted propulsion.

[0074] This invention discloses an energy-saving obstacle avoidance method for marine robots that considers energy capture, such as... Figure 1 As shown, the specific steps are as follows:

[0075] Step 1: Determine the marine robot's mission information, its own position, attitude, speed, and obstacle information within the sea area. Utilize the environmental perception system to acquire the marine environment information of the marine robot at the current moment (including sea breeze, waves, ocean currents, and solar radiation intensity).

[0076] Step Two: Establish an energy modeling system based on environmental information and robot structural characteristics. This system includes a velocity response model, an energy consumption model, and a natural energy capture model. The velocity response model characterizes the impact of sails and hydrofoils / wave panels on the robot's effective speed under wind, waves, and currents. The energy consumption model calculates the energy consumption of environmental perception, communication, intelligent control, and actuators. The natural energy capture model calculates the effective power generation of the photovoltaic panels under solar radiation, hull attitude, heading, and sail shading conditions, and further obtains the net energy consumption under candidate navigation states.

[0077] Step 3: Combining the wind, wave, and current information detected by the environmental perception system, and based on robot kinematic constraints, acceleration constraints, obstacle distance constraints, and effective speed constraints under the influence of natural energy, determine the robot's dynamic velocity window. Discretely sample the linear velocity and angular velocity to obtain multiple candidate velocity combinations. Substitute each candidate velocity combination into the robot's kinematic model to generate candidate trajectories corresponding to each candidate velocity combination within a preset prediction time window.

[0078] Step 4: Expand the safe range of obstacles based on their locations, and construct a dynamic window method comprehensive evaluation function using heading, speed, distance from the target point, distance from the obstacle, and net energy consumption as evaluation indicators.

[0079] Step 5: Using the evaluation function weight coefficients as optimization variables, and taking the net energy consumption and obstacle avoidance safety of the candidate trajectory within the current prediction time window as optimization criteria, a metaheuristic optimization algorithm is used to adaptively optimize the evaluation function weights to obtain the optimal evaluation function coefficient set.

[0080] Step 6: Substitute the coefficient set of the optimal evaluation function into the dynamic window method to comprehensively evaluate the candidate trajectories, select the speed combination with the best score, and issue the speed expectation value by the control system so that the marine robot can reduce the net energy consumption of navigation while avoiding obstacles.

[0081] Steps one through six are executed cyclically during the robot's navigation until the robot reaches the target point or the mission termination condition is met.

[0082] The energy modeling system for the marine robot described in step two includes a velocity response model, an energy consumption model, and a natural energy capture model.

[0083] The speed response model takes sea state level, relative wind and wave angle, ocean current speed, and ocean current direction as input, and outputs the robot's effective ground speed in the current heading, such as... Figure 2 As shown, the specific construction steps are as follows:

[0084] Step 21: Obtain CFD speed response samples under typical sea states. The computational fluid dynamics (CFD) numerical simulation method is used to calculate the robot's speed response characteristics under pure wind and wave action. Several typical sea state levels are selected, and considering the physical characteristic that the wind direction and wave direction are consistent under wind-driven wave conditions, multiple sets of typical wave encounter angles are set for simulation. The wave encounter angle refers to the angle between the robot's heading and the direction of the incoming wave.

[0085] The robot adjusts the sail angle in real time based on the current encounter angle to obtain the maximum auxiliary thrust under the current operating conditions, and obtains the corresponding optimal speed sample through CFD simulation:

[0086]

[0087] in, Indicates sea state L, the CFD simulation of ship speed under different wind angles; This represents the CFD numerical simulation solution function; Indicates the first Group wind direction angle; Indicates the first The encounter angle is the optimal sail angle that allows the robot to obtain the maximum auxiliary thrust.

[0088] Step 22: Generate continuous velocity data based on PCHIP interpolation. Since CFD simulations only yield corresponding speeds at a few discrete wind direction angles, the speed samples are sparse. Directly using the dynamic window method would result in discontinuous candidate heading evaluations. Therefore, piecewise cubic conformal interpolation (PCHIP) is performed on the discrete speed samples to obtain angular resolution data. Continuous speed data. For each sea state, a known discrete sample set is:

[0089]

[0090] In this embodiment, Q is set to 7.

[0091] The continuous speed response curve was obtained using PCHIP interpolation:

[0092]

[0093] Steps two and three: Fit the continuous velocity mathematical function using a cosine series. To improve the online calculation efficiency of the dynamic window method, the obtained continuous velocity polar curves are... Further fitting using truncated Fourier cosine series yields a function form that is convenient for real-time calling by the DWA algorithm:

[0094]

[0095] in, This indicates the speed of the ship under different sea conditions and the influence of wind and waves. The relative wind direction angle, ; Sea state rating ; To truncate the order, this embodiment takes coefficient vector Estimated using the least squares method.

[0096] Step 24: Perform continuous interpolation between adjacent sea states based on wind speed. Since actual sea states are not always strictly equal to sea states 1, 2, 3, or 4, directly looking up the most recent sea state level would cause abrupt changes in speed prediction. Therefore, this invention uses sea surface wind speed as an environmental intensity indicator to perform linear interpolation of speed between adjacent sea state levels.

[0097] When the current sea surface wind speed is at the [number]th Sea state level corresponding wind speed With the When the wind speed corresponds to sea state level, the continuous speed prediction function is:

[0098]

[0099] in, Indicates the current sea surface wind speed; Indicates the current wind speed The robot's speed is affected by the wind and waves at a relative angle θ. and These represent the wind speed boundaries corresponding to two adjacent sea state levels; and These represent the velocity response functions under two adjacent sea states, respectively.

[0100] Step 25: Superimpose the ocean current effect onto the wind and wave speed response model. After obtaining the predicted speed under pure wind and wave influence, further consider the impact of ocean currents on the robot's ground speed. Express the pure wind and wave speed and ocean current speed in vector form:

[0101]

[0102]

[0103]

[0104]

[0105]

[0106] in, and These are the speed vector and scalar of the surface navigation state under the combined influence of wind, waves and current, respectively; the latter is the output of the speed response model. and Representing the ocean current velocity vector and scalar, respectively. This represents the ship's speed vector under the influence of wind and waves in a calm state. For heading angle, , These represent the absolute and relative direction angles of the ocean current, respectively. This indicates the effective speed that a marine robot can achieve under the influence of natural forces such as sea breeze, waves, and currents.

[0107] Furthermore, the total power output of the marine robot involved in step two is represented as follows:

[0108]

[0109] in, This is an energy consumption model that characterizes the energy consumption power of a marine robot, specifically including the output power of the environmental perception subsystem, communication subsystem, intelligent control subsystem, and actuator subsystem. It can be represented as:

[0110]

[0111] The The natural energy capture model characterizes the solar energy capture power of the marine robot under the current environment and attitude. Since the marine robot of this invention mainly obtains electricity through photovoltaic modules, and the deck photovoltaic panels are affected by the wind and sails, the energy capture power is expressed as the solar energy capture power after considering the wind and sail shading correction.

[0112] The shadow cast by the sails on the ship's hull on the horizontal plane of the deck under sunlight is the shadow cast by the sails on the deck's photovoltaic panels. As the sun's position and the ship's course change relative to these changes, the shadow on the deck transitions from a triangle to a trapezoid; therefore, the area of ​​the shadow needs to be calculated piecewise to avoid errors caused by a single formula. For example... Figure 3 and 4As shown, the angle between sunlight and the x-axis of the ship's coordinate system is defined as... The critical angle is defined as the situation where the upper boundary of the shadow cast on the deck coincides exactly with the diagonal of the rectangular photovoltaic panel on the deck. ,when When the shadow area is triangular, it is triangular; otherwise, it is trapezoidal. From this, we can gradually derive the formula for the effective illumination area:

[0113]

[0114]

[0115]

[0116]

[0117]

[0118] In the formula, This indicates the effective light-receiving area of ​​the photovoltaic panels on the sail. This represents the effective light-receiving area of ​​the photovoltaic panels, taking into account the impact of sails shading the solar panels on the deck. This represents the total area of ​​actual effective illumination. and This indicates the width and length of the photovoltaic panels on the ship's hull. and This indicates the width and length of the photovoltaic panels on the sail surface. Indicates the azimuth angle of the sun. Indicates the solar altitude angle. Indicates the sail's angle.

[0119]

[0120] in, G represents the light energy conversion coefficient, which can be obtained through photovoltaic panel experiments, and G is the solar radiation intensity.

[0121] In step 3, during the construction of the speed dynamic window, the robot's maximum and minimum speeds and acceleration constraints are first determined. Then, the maximum collision avoidance distance is limited by the obstacle position, and the effective speed provided by natural energy is incorporated into the constraints, thus forming the final speed dynamic window. This ensures that all candidate velocities sampled meet safety and energy utilization requirements. Robot speed dynamic window size optimization:

[0122] Speed ​​boundary limits:

[0123] Acceleration limit:

[0124] Obstacle restrictions:

[0125] Minimum speed limit optimization:

[0126] in, and These represent the basic minimum and maximum linear velocities of the marine robot, determined by its own motion capabilities, respectively. and These represent the minimum and maximum angular velocities of the marine robot, respectively. and These represent the robot's current linear velocity and angular velocity, respectively. and These represent the robot's maximum linear acceleration and angular acceleration, respectively. This represents the closest distance between the simulated trajectory corresponding to the combination of the robot's linear velocity and angular velocity and the obstacle. This represents the maximum speed that ocean energy can provide. It is related to information such as waves, sea winds, and ocean currents, and can be predicted using a speed response model. This represents the minimum linear velocity within the dynamic speed window after natural energy-based effective speed correction; where, , For the correction coefficient, satisfying This allows control over the correction range of the minimum speed limit relative to the effective speed of natural energy. By adjusting... The size of the value can control the minimum velocity increase in the candidate velocity space to adapt to different natural energy conditions and obstacle avoidance requirements.

[0127] This leads to the final set of velocities for the dynamic window:

[0128] This invention introduces a natural energy-efficient speed to correct the minimum linear velocity boundary, enabling the candidate velocity space to dynamically adjust with changes in wind, wave, and current conditions. This process avoids the algorithm from evaluating a large number of low-speed and inefficient trajectories even when natural energy availability is good, thereby improving the matching degree between the velocity sampling space and the natural energy-driven characteristics. This enhances natural energy utilization efficiency and energy-saving obstacle avoidance effects while ensuring safe obstacle avoidance.

[0129] i candidate velocity combinations obtained by discrete sampling under the constraint of the final velocity set in the velocity dynamic window. The parameters are substituted into the robot's kinematics model as control inputs to predict the trajectory position, heading, velocity, angular velocity, and energy changes within the current time window, thereby generating each candidate trajectory for the evaluation function to calculate the overall performance. The specific kinematic model is as follows:

[0130]

[0131] In the formula, and Let represent the predicted position coordinates of the i-th candidate trajectory at time t; Indicates the prediction step time interval; and They represent the first The position coordinates of each candidate trajectory at the next prediction time step; This represents the candidate linear velocity corresponding to the i-th candidate trajectory, which is the ground speed obtained by sampling from the velocity dynamic window; This represents the linear velocity of the i-th candidate trajectory at the next prediction time step. It remains consistent with the sampled candidate velocity within the prediction time window and is used to predict the position of the trajectory in the next step. Indicates the first Total output power of the robot for each candidate trajectory; This indicates the robot's total output power; Indicates the heading corresponding to the i-th candidate trajectory; Indicates the first The heading of each candidate trajectory at the next prediction time step; This represents the cumulative net energy consumption of the i-th candidate trajectory at time t; This represents the cumulative net energy consumption of the i-th candidate trajectory; The effective speed of the i-th candidate trajectory under the influence of natural energy in the next prediction time step is represented by the velocity response model. Indicates the relative direction of ocean currents; Relative wind direction angle; Indicates the current sea surface wind speed; This represents the candidate angular velocity corresponding to the i-th candidate trajectory, which is the rate of change of heading obtained by sampling from the velocity dynamic window; This represents the angular velocity of the i-th candidate trajectory at the next prediction time step. It remains consistent with the sampled candidate angular velocities within the prediction time window and is used to predict the trajectory's heading in the next step. Both velocity and angular velocity are sampled candidate values, already satisfying the constraints.

[0132] In step 4, the comprehensive evaluation function is:

[0133]

[0134]

[0135]

[0136]

[0137]

[0138]

[0139] In the formula, Indicates the maximum collision avoidance distance from the obstacle; when the distance between the obstacle and the robot is greater than... When using the evaluation item weight coefficient Multiply by the scores of each evaluation item; otherwise, use the weight coefficient of the evaluation item. Multiply by the scores of each evaluation item; This represents the total number of candidate trajectories generated within the current speed dynamic window; i represents the i-th candidate trajectory currently being evaluated. This represents the traversal index during normalized summation, used to traverse all candidate trajectories. ; Indicates the first The heading indicator of the trajectory, This represents the angular deviation between the trajectory's final heading and the direction of the line connecting the predicted point and the target point. This represents the heading evaluation value of the corresponding trajectory after normalization. Indicates the first The distance index of the trajectory is the reciprocal of the Euclidean distance from the end of the trajectory to the target point. The predicted position coordinates of the i-th candidate trajectory at the end of the prediction time window. Indicates the coordinates of the target point. This represents the target distance evaluation value of the corresponding trajectory after normalization. Indicates the first The speed index corresponding to each trajectory, that is, the absolute magnitude of the speed. The represented raw speed, including directional information. This represents the speed evaluation value of the corresponding trajectory after normalization. Indicates the first The energy metric for this trajectory is the reciprocal of the net energy consumption value within the predicted time window. This indicates the predicted travel time for the corresponding trajectory. This represents the energy evaluation value of the corresponding trajectory after normalization. Indicates the first The obstacle distance index corresponding to each trajectory is the shortest distance between the i-th candidate trajectory and the obstacle within the prediction time window. This represents the coordinates of the m-th predicted sampling point on the i-th candidate trajectory. Represents the coordinates of the obstacle. This represents the obstacle distance evaluation value after normalization.

[0140] In step five, the comprehensive evaluation function is optimized using a metaheuristic optimization algorithm: the weight coefficients of the evaluation function in the dynamic window method are used as optimization variables, and the total net energy consumption of the candidate trajectory within the prediction time window and the minimum distance between the candidate trajectory and the obstacle are used as the objective function of the metaheuristic optimization algorithm. The evaluation function weights for optimal energy utilization are obtained through algorithm iteration.

[0141] Step 51: Initialize the population size: The individuals in the population are a vector consisting of ten real numbers, each of which is in the range of [0,1].

[0142] Step 52: Calculate Fitness: For each individual in the population, substitute the candidate weight vector into the evaluation function to weight and score all candidate trajectories. Select the trajectory with the highest score under this weight vector as the representative trajectory. Then, calculate the fitness value of this individual based on the energy consumption and minimum distance to obstacles within the current prediction time window. The specific fitness function is constructed as follows:

[0143]

[0144] Where time represents the travel time of the i-th candidate trajectory within the current prediction time window. obs(i) represents the closest distance between the i-th candidate trajectory and the obstacle. This represents the maximum collision avoidance distance from the obstacle. The fitness function described above is based on obs(i) and... The relationship is constructed in segments: when the robot is in a relatively safe state, the fitness function uses the reciprocal of the energy consumption of the candidate trajectory within the current prediction time window, so that the optimization process prioritizes the candidate trajectory with lower energy consumption; when the robot is in an obstacle avoidance priority state, a dimensionless maximum obstacle collision avoidance danger distance correction term is introduced on the basis of the reciprocal of the predicted energy consumption of the candidate trajectory, so that the optimization process can improve obstacle avoidance safety while taking into account the energy saving goal, and avoid the problem of low speed lingering near obstacles or insufficient obstacle avoidance caused by simply pursuing low energy consumption.

[0145] Step 53: Search and Update: Update the weight vector using a metaheuristic optimization algorithm (such as particle swarm optimization, whale optimization, ant colony optimization, or any continuous optimization method). Individuals move towards the fitness-optimal solution to ensure that random perturbations are introduced into the search process, avoiding getting trapped in local optima. In the iteration, the energy consumption of the candidate trajectory corresponding to each weight update is calculated using the latest solar energy capture model, ensuring that the search process is closely integrated with the robot's energy constraints.

[0146] Step 54: Termination Condition: Stop optimization when the maximum number of iterations is reached, or when the fitness of all individuals converges. Output the weight vector with the highest fitness in the current solution set.

[0147] Step 55: Application of weight vector: Based on the iteration of the optimization algorithm, the optimal weight coefficient group is obtained and substituted into the dynamic window method to calculate the comprehensive evaluation value of the trajectory cluster. The speed combination with the best evaluation value is selected, and the control system issues the desired speed. The actuator completes the operation to avoid obstacles.

[0148] Example 2

[0149] In this embodiment, a genetic algorithm is used to optimize the weight coefficients of the comprehensive evaluation function. Figure 5 As shown; the evaluation function coefficients are used as the dependent variable, and the total energy consumption during the entire voyage and the minimum distance between the voyage trajectory and obstacles are used as the objective function of the genetic algorithm. The optimal energy evaluation function coefficients are obtained through iterative genetic algorithm. The specific operation is as follows:

[0150] Step 1: Initialize the population size: The individuals in the population are a vector consisting of ten real numbers, each in the range [0,1]. Considering that the individuals in the population are real numbers, the chromosome encoding uses real number encoding.

[0151] Step 2: Evaluate each individual in the population and calculate its fitness value;

[0152]

[0153] Step 3: Select parent individuals based on their fitness values. The higher the fitness value, the greater the probability of being selected. Use roulette wheel selection. Generate new individuals (offspring) through crossover operations. Use two-point crossover, that is, randomly set two crossover points in the individual's encoding string, and then perform partial gene exchange.

[0154] Step 4: Randomly modify the genes of individuals through mutation operations to introduce diversity and prevent the population from getting trapped in local optima. Use an adaptive feasible mutation strategy, that is, adjust the mutation rate according to the fitness value of the current solution to increase or decrease the number of mutation operations, thereby optimizing the search performance.

[0155] Step 5: Replace all parent individuals with the newly generated offspring to form a new generation of population; check if the termination condition (maximum number of iterations) is met. If it is met, end the algorithm; otherwise, return to step 2 to continue iterating.

[0156] Step 6: Output the individual with the best fitness in the population as the optimal solution.

[0157] Furthermore, the optimal evaluation function coefficient set is obtained through iterative genetic algorithm and substituted into the dynamic window method to calculate the comprehensive evaluation value of the trajectory cluster. The speed combination with the optimal evaluation value is selected, and the control system issues the desired speed. The actuator completes the operation to avoid obstacles.

[0158] In particular, in some preferred embodiments of the present invention, a computer device is also provided, including a memory and a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the marine robot energy modeling and energy-saving obstacle avoidance method considering natural energy capture described in any of the above embodiments.

[0159] In some other preferred embodiments of the present invention, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the marine robot energy modeling and energy-saving obstacle avoidance method considering natural energy capture described in any of the above embodiments.

[0160] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above embodiments of the marine robot energy modeling and energy-saving obstacle avoidance method considering natural energy capture, which will not be repeated here.

[0161] While the invention has been described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the invention. Therefore, it should be understood that many modifications can be made to the exemplary embodiments, and other arrangements can be designed without departing from the spirit and scope of the invention as defined by the appended claims. It should be understood that different dependent claims and features described herein can be combined in ways different from those described in the original claims. It is also understood that features described in conjunction with individual embodiments can be used in other described embodiments.

Claims

1. A method for energy modeling and energy-saving obstacle avoidance of marine robots considering natural energy capture, characterized in that, Includes the following steps: Acquire information about the marine robot's mission, navigation environment, and its own motion status; Construct an energy modeling system that includes a velocity response model, an energy consumption model, and a natural energy capture model; The natural energy capture model is calculated using solar power capture power after considering sail obstruction correction. An effective speed constraint under the influence of natural energy is introduced to construct a dynamic speed window; candidate speed combinations are sampled within the dynamic speed window, and candidate trajectories are generated by combining them with a kinematic model. The dynamic speed window includes speed boundary constraints, acceleration constraints, obstacle constraints, and effective speed constraints under the influence of natural energy; the effective speed constraints under the influence of natural energy are as follows: in, This represents the minimum linear velocity within the dynamic speed window after effective speed correction based on natural energy. and These represent the basic minimum and maximum linear velocities of the marine robot, determined by its own motion capabilities, respectively. and These represent the minimum and maximum angular velocities of the marine robot, respectively. Indicates the correction factor; This indicates the effective speed that a marine robot can achieve under the influence of natural energy. The trajectory is segmented based on the maximum collision avoidance distance from the obstacle. The candidate trajectory's speed, heading, distance from the target point, distance from the obstacle, and net energy consumption are used as evaluation indicators to construct a comprehensive evaluation function. After acquiring multiple sets of predicted trajectories, the predicted trajectories are scored using a comprehensive evaluation function, and the velocity combination with the best score is selected as the motion control command for the robot at the next moment. Repeat the above steps until the robot reaches the target point or the task termination condition is met.

2. The method for energy modeling and energy-saving obstacle avoidance of marine robots considering natural energy capture according to claim 1, characterized in that, The velocity response model is specifically as follows: Step 1.1: The robot adjusts the sail angle in real time according to the current wind direction angle to obtain the maximum auxiliary thrust under the current working conditions, and obtains the corresponding optimal speed through CFD simulation; in, Indicates sea state L, the CFD simulation of ship speed under different wind angles; This represents the CFD numerical simulation solution function; Indicates the first Group wind direction angle; Indicates the first The encounter angle is the optimal sail angle that allows the robot to obtain the maximum auxiliary thrust. Step 1.2: Obtain the continuous speed response curve using a piecewise cubic conformal interpolation method. Furthermore, the truncated Fourier cosine series method is used for fitting, generating a function that can be called in real time using the dynamic window method. ; Step 1.3: Perform linear interpolation of the speed between adjacent sea state classes to calculate the wind speed introduced from the sea surface. Post-continuous speed prediction function ; in, and These represent the wind speed boundaries corresponding to two adjacent sea state levels; and These represent the velocity response functions under two adjacent sea states, respectively. Step 1.4: Superimpose the ocean current effect onto the wind and wave speed response model; in, and These represent the speed vector and scalar respectively under the combined influence of wind, waves, and current during surface navigation. and These represent the ocean current velocity vector and scalar, respectively. Indicates the relative direction angle of ocean currents. Indicates the absolute direction angle of ocean currents. Indicates the heading angle.

3. The method for energy modeling and energy-saving obstacle avoidance of marine robots considering natural energy capture according to claim 1, characterized in that, The energy consumption model The output power of environmental perception, communication, intelligent control, and actuator is calculated jointly.

4. The method for energy modeling and energy-saving obstacle avoidance of marine robots considering natural energy capture according to claim 1, characterized in that, The natural energy capture model for: in, Indicates the intensity of solar radiation; This represents the total area of ​​actual effective illumination; Indicates the light energy conversion coefficient; This indicates the effective light-receiving area of ​​the photovoltaic panels on the sail. This represents the effective light-receiving area of ​​the photovoltaic panels, taking into account the shading effect of the sails on the deck photovoltaic panels. This indicates the solar altitude angle.

5. The method for energy modeling and energy-saving obstacle avoidance of marine robots considering natural energy capture according to claim 4, characterized in that, The effective solar irradiance area of ​​the photovoltaic panels considering the shading effect of sails on the deck photovoltaic panels. Calculate using the following method: in, and This indicates the width and length of the photovoltaic panels on the ship's hull. This represents the angle of incidence of sunlight, that is, the angle between sunlight and the x-axis of the ship's coordinate system. Indicates the heading angle. Indicates the solar azimuth angle; This represents the angle of sunlight incidence when the upper boundary of the shadow cast by the sail on the deck coincides with the diagonal of the photovoltaic rectangle on the deck. The effective light-emitting area of ​​the photovoltaic panels on the sail for: in, and This indicates the width and length of the photovoltaic panels on the sail surface; Indicates the sail's angle.

6. The method for energy modeling and energy-saving obstacle avoidance of marine robots considering natural energy capture according to claim 1, characterized in that, The candidate velocity combination Candidate trajectories are generated by combining the kinematic model, which is as follows: in, and They represent the first Candidate trajectories in The predicted location coordinates at that moment; Indicates the prediction step time interval; and They represent the first The position coordinates of each candidate trajectory at the next prediction time step; Indicates the first Total output power of the robot for each candidate trajectory; Indicates the first The cumulative net energy consumption of the candidate trajectories at time t; Indicates the first The cumulative net energy consumption of the candidate trajectories; Indicates the first Candidate linear velocities corresponding to each candidate trajectory; Indicates the first The linear velocity of each candidate trajectory at the next prediction time step; Indicates the heading corresponding to the i-th candidate trajectory; Indicates the first The heading of each candidate trajectory at the next prediction time step; Indicates the first The effective speed of each candidate trajectory under the influence of natural energy in the next prediction time step; Indicates the first The candidate angular velocities corresponding to the candidate trajectories; Indicates the first The angular velocity of each candidate trajectory at the next prediction time step; Indicates the relative direction of ocean currents; Relative wind direction angle; This indicates the current wind speed at sea.

7. The method for energy modeling and energy-saving obstacle avoidance of marine robots considering natural energy capture according to claim 1, characterized in that, The comprehensive evaluation function for: in, Indicates the maximum collision avoidance distance from the obstacle; , , , , , , , , and These are the weight coefficients for the evaluation items; This indicates the total number of candidate trajectories generated within the current speed dynamic window; This represents the traversal index during normalized summation, used to traverse all candidate trajectories. ; Indicates the first The heading indicator of the trajectory; This indicates the angular deviation between the trajectory's final heading and the direction of the line connecting the predicted point and the target point; This represents the heading evaluation value of the corresponding trajectory after normalization. Indicates the first Distance metrics for each trajectory; No. The predicted position coordinates of the candidate trajectories at the end of the prediction time window; Indicates the coordinates of the target point; This represents the target distance evaluation value of the corresponding trajectory after normalization. Indicates the first The speed index corresponding to each trajectory; This represents the speed evaluation value of the corresponding trajectory after normalization. Indicates the first Energy indicators for each trajectory; This represents the energy evaluation value of the corresponding trajectory after normalization. Indicates the first Distance indicators between candidate trajectories and obstacles; Indicates the first The first candidate trajectory The coordinates of each predicted sampling point; Indicates the coordinates of the obstacle; This represents the obstacle distance evaluation value after normalization.

8. The method for energy modeling and energy-saving obstacle avoidance of marine robots considering natural energy capture according to claim 1, characterized in that, The comprehensive evaluation function adopts a metaheuristic optimization algorithm for adaptive optimization. The total net energy consumption of the candidate trajectory within the prediction time window and the minimum distance between the candidate trajectory and the obstacle are used as the fitness function. The evaluation function weights for optimal energy utilization are obtained through algorithm iteration.

9. The method for energy modeling and energy-saving obstacle avoidance of marine robots considering natural energy capture according to claim 8, characterized in that, The fitness function is designed in segments based on the maximum collision avoidance distance of the obstacle: when the minimum distance between the candidate trajectory and the obstacle is less than or equal to the maximum collision avoidance distance, the fitness function introduces a dimensionless maximum collision avoidance distance correction term based on the reciprocal of the net energy consumption of the candidate trajectory; when the minimum distance between the candidate trajectory and the obstacle is greater than the maximum collision avoidance distance, the fitness function uses the reciprocal of the energy consumption of the candidate trajectory within the current prediction time window.

10. A computer device, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method as described in any one of claims 1 to 9.

Images