A method for deploying and recovering an underwater robot

Abstract

The present disclosure relates to a method for deploying and recovering an underwater robot, including: determining an energy budget according to the battery capacity and mission duration of the underwater robot; obtaining historical ocean current and tide data to determine the influence degree of ocean currents and tides on the path planning of the underwater robot and to determine the influence degree of ocean currents and tides on the propulsion system of the underwater robot; obtaining the direction data, speed data of ocean currents and the periodic change data of tides, and analyzing the vertical profile change of flow velocity at different positions; judging the ocean current and tide conditions according to the obtained direction data, speed data of ocean currents and the periodic change data of tides, and combining with the energy budget, and making an adjustment strategy for the propulsion system according to the ocean current and tide conditions; real-time monitoring the direction data, speed data of ocean currents and the periodic change data of tides, and combining with the coordinates of the deployment task point and recovery point of the underwater robot to generate a path planning report; the present disclosure realizes the intelligent and autonomous operation of the underwater robot.

Description

Technical Field

[0001] The present disclosure relates to the technical field of underwater robot deployment and recovery, and particularly to a method for deploying and recovering an underwater robot. Background Art

[0002] When an underwater robot performs ocean exploration tasks, it needs to face complex and ever-changing ocean environments. Among them, ocean currents and tides have important impacts on the movement trajectory and energy consumption of the robot. The flow velocity and direction of ocean currents change with time and space, showing a certain periodicity. Tides also cause the rise and fall of seawater, thereby changing the vertical distribution of underwater pressure and density. These factors interfere with the movement of the robot underwater, causing it to deviate from the predetermined route and consume additional energy. At the same time, due to the limited battery capacity carried by the robot, energy budget becomes a key factor restricting the mission duration. How to adjust the buoyancy and propulsion strategy of the robot according to the real-time obtained ocean current and tide data, and complete the established tasks under the premise of meeting the energy budget is a technical problem to be solved urgently. Therefore, designing a method for deploying and recovering the robot to ensure its safe return after the mission is completed while minimizing energy consumption has become an important part of achieving this goal. This requires comprehensive consideration of multiple aspects such as sensor layout, data processing algorithms, and control strategy optimization in the design of the robot, and verifying its effectiveness and reliability through a large number of sea trial experiments. Only on the basis of mastering the characteristics of the ocean dynamic environment can the intelligent and autonomous operation of underwater robots be truly realized. Summary of the Invention

[0003] In order to solve the above problems existing in the prior art, the purpose of the present disclosure is to provide a method for deploying and recovering an underwater robot to ensure the safe return of the underwater robot after the mission is completed while minimizing energy consumption.

[0004] The method for deploying and recovering an underwater robot described in the present disclosure includes:

[0005] Determine the energy budget according to the battery capacity and mission duration of the underwater robot; obtain historical ocean current and tide data, determine the influence degree of ocean currents and tides on the path planning of the underwater robot, and determine the influence degree of ocean currents and tides on the propulsion system of the underwater robot;

[0006] Obtain the direction data, speed data of the ocean current and the periodic change data of the tide, and analyze the change of the vertical velocity profile at different positions;

[0007] According to the obtained direction data, speed data of the ocean current and the periodic change data of the tide, combined with the energy budget, judge whether the ocean current and tide conditions are favorable ocean currents or unfavorable ocean currents, and make an adjustment strategy for the propulsion system according to the conditions of the ocean current and tide;

[0008] If it is determined to be a favorable ocean current, the autoregressive moving average model is used to predict the velocity change trend in the future for a period of time, and the buoyancy of the underwater robot is adjusted in advance according to the prediction result;

[0009] If it is determined to be an unfavorable ocean current, the fuzzy control algorithm is used to adjust the buoyancy and the power output of the propulsion system of the underwater robot; through tidal data, the artificial potential field method is used to dynamically optimize the path;

[0010] The direction data, velocity data of the ocean current and the periodic change data of the tide are monitored in real time, and combined with the coordinates of the underwater robot deployment task point and the recovery point, a path planning report is generated.

[0011] Preferably, according to the battery capacity and mission duration of the underwater robot, the energy budget is determined; historical ocean current and tidal data are obtained to determine the influence degree of the ocean current and tide on the path planning of the underwater robot, and the influence degree of the ocean current and tide on the propulsion system of the underwater robot, including:

[0012] According to parameters such as the battery capacity and estimated mission duration of the underwater robot, the energy budget value during the entire mission is calculated through the energy consumption model to obtain the upper limit of the available energy of the underwater robot; historical ocean current and tidal data of the target sea area are obtained and preprocessed; the spectrum analysis of the preprocessed historical ocean current and tidal data is carried out through the fast Fourier transform algorithm to obtain the flow velocity magnitude, flow direction and periodic change characteristics of the ocean current and tide at different time scales; according to the results of the spectrum analysis, the influence degree of the ocean current and tide on the movement of the underwater robot is judged.

[0013] Preferably, the direction data, velocity data of the ocean current and the periodic change data of the tide are obtained, and the vertical profile change of the flow velocity at different positions is analyzed, including:

[0014] The ocean current in the target sea area is monitored in real time through a Doppler velocity profiler to obtain the three-dimensional flow field structure data of the entire water body; by measuring the velocity components in different azimuths, the velocity vectors at each position are obtained, and the ocean current distribution characteristics of the entire area are comprehensively analyzed to judge whether there are significant spatial differences; the tidal level periodic change data is obtained through a tide gauge, and the fast Fourier transform (FFT) is used to perform spectrum analysis on the tidal period to determine the main period of the tide; the tidal height rise and fall situation obtained by the tide gauge is correlated with the flow velocity change measured by the velocity profiler to obtain the interaction mechanism between the tide and the ocean current; the vertical stratification analysis of the flow velocity data obtained by the velocity profiler and the current meter is carried out, and the average flow velocity and direction of different depth layers are calculated and the vertical profile of the flow velocity changing with depth is drawn.

[0015] Preferably, based on the obtained direction data, velocity data of ocean currents, and periodic change data of tides, and combining with the energy budget, judge whether the ocean current and tide conditions are favorable or unfavorable ocean currents, and make adjustment strategies for the propulsion system according to the ocean current and tide conditions, including:

[0016] According to the measurement results of an acoustic Doppler current profiler, calculate the velocity and direction of ocean currents in the target sea area. Through statistical analysis of ocean current vectors, obtain the average velocity, main flow direction, and probability distribution characteristics of the flow velocity of ocean currents; through a probability distribution model, judge the influence degree of ocean currents on the propulsion efficiency and course-keeping ability of the underwater robot; extract the main periodic components of tides through spectral analysis methods, and establish a harmonic model of tidal level changes; through methods such as the least squares method or Fourier transform, predict the tidal ebb and flow process in a future period of time; conduct statistical analysis on the rise and fall of tidal levels, calculate the difference between the high tide level and the low tide level within one tidal cycle to determine the amplitude and uncertainty of tidal changes; comprehensively analyze multi-source monitoring data such as velocity profiles, ocean current velocities, and tidal periods, and through data fusion and machine learning algorithms, establish an energy optimization model for the underwater robot. Through the energy optimization model, predict the energy consumption change trend under different speeds, headings, and depths, and obtain the optimal combination of motion parameters with the highest energy efficiency; according to the prediction results of the energy optimization model, dynamically adjust the buoyancy of the underwater robot; the purpose of adjusting the position of the center of buoyancy is to make the underwater robot reach a balanced state in still water and reduce the coupling effect during dynamic motion. On the premise of ensuring the attitude stability of the underwater robot, minimize the energy loss during the buoyancy adjustment process. Optimize the control strategy of the propulsion system of the underwater robot for different ocean current and tide conditions;

[0017] Obtain parameters such as the battery capacity and discharge characteristics of the underwater robot, and combine with the actual mission requirements to calculate the total available energy and average power limit during the entire mission period through an energy budget model; obtain the drag characteristics and propeller characteristics of the underwater robot to obtain the drag curve and propeller efficiency curve at different speeds; according to the ocean current and tide data, combine with the drag characteristics and propeller characteristics of the underwater robot, and through computational fluid dynamics methods, obtain the propulsion efficiency curve and response surface; by fitting the propeller characteristic surface, establish the mapping relationship between thrust demand and shaft power, and through machine learning algorithms, establish the non-linear mapping relationship between ocean currents, tides, speed, and propulsion efficiency. Through training and optimization of historical data, obtain the propulsion efficiency prediction model; combine the propulsion efficiency prediction model with the objective function of energy optimization to establish a non-linear programming model, and obtain the optimal navigation strategy parameters with the highest energy utilization efficiency through the non-linear programming model; based on the optimal navigation strategy parameters, adjust the propeller speed, pitch angle, and rudder angle in real time;

[0018] Analyze the ocean current and tidal characteristics for each category, make judgments from both favorable and unfavorable perspectives, and set thresholds for parameters such as flow velocity, flow direction, and tidal level; establish a multi-objective optimization model with the energy utilization efficiency and speed retention rate as the objective functions, combined with the propeller characteristic curve, energy consumption model, and historical data of the underwater robot, to obtain the optimal speed and course angle thresholds under various environmental conditions. When the measured value exceeds the threshold range, it is considered an unfavorable ocean current, and vice versa for a favorable ocean current.

[0019] Preferably, if it is determined to be a favorable ocean current, predict the change trend of the flow velocity in the future for a period of time through an autoregressive moving average model, and adjust the buoyancy of the underwater robot in advance according to the prediction result, including:

[0020] When it is determined to be a favorable ocean current, control the drainage volume of the ballast tank of the underwater robot to balance the buoyancy and gravity of the underwater robot; at the same time, control the course angle of the robot to be consistent with the ocean current direction; combine the predicted ocean current speed, direction, and duration information, and obtain the optimal speed and course control sequence through the objective function of energy optimization, while ensuring the track tracking accuracy, minimize the energy consumption; monitor the attitude and position of the robot in real time, introduce a trajectory tracking algorithm based on model predictive control, and adjust the control quantity in advance according to the prediction of the motion state in the future for a period of time to reduce the course deviation caused by ocean current interference and maintain smooth and efficient motion.

[0021] Preferably, if it is determined to be a favorable ocean current, predict the change trend of the flow velocity in the future for a period of time through an autoregressive moving average model, and adjust the buoyancy of the underwater robot in advance, including:

[0022] Obtain the fuzzy control rule table of the buoyancy adjustment amount through a fuzzy control algorithm; dynamically adjust the buoyancy adjustment amount and propulsion system power in the future for a period of time through a Kalman filter combined with a model predictive control (MPC) algorithm, synchronized with the ocean current change; obtain the optimal combination of buoyancy adjustment amount and propulsion power through a nonlinear optimization algorithm; through the artificial potential field method, according to the deployment position, predetermined task path, and recovery position of the robot, combined with the real-time ocean current distribution data measured by ADCP, segmentally optimize the global path, use the gradient information of the artificial potential field function as the input of the fuzzy control and Kalman filter, and perform dynamic weight adjustment according to the state error of the underwater robot and the magnitude of environmental interference, and adjust the weight coefficients of the artificial potential field method and the feedback control law in real time.

[0023] Preferably, the direction data, speed data of the ocean current, and the periodic change data of the tide are monitored in real time, and combined with the coordinates of the underwater robot deployment task point and the recovery point, a path planning report is generated, including:

[0024] Obtain the coordinates of the deployment task points and recovery points of the underwater robot. Based on the Euclidean distance and azimuth angle between the two points, preliminarily plan a basic navigation path; use the A* algorithm for global path search to generate a collision-free and blind-free global optimal path; superimpose and analyze the global optimal path with the predicted ocean current field to evaluate the influence degree of ocean currents on the movement of the underwater robot; through numerical simulation or pool tests, obtain the optimal heading angle range under different working conditions; according to the results of the ocean current influence analysis, locally optimize the global path, and use the artificial potential field method to introduce virtual gravitational and repulsive forces to make the underwater robot stay away from the countercurrent area, approach the downstream area, and keep parallel to the ocean current direction; construct a potential field function, reasonably set the combined weights of the gravitational potential field and the repulsive potential field, and superimpose the tidal current velocity vector into the potential field gradient. The potential field function is used to make the path planning adapt to the periodic changes of tides; based on the optimized path, formulate adjustment strategies for the buoyancy and propulsion systems according to the dynamic characteristics and maneuvering performance of the underwater robot; establish a dynamic model of the underwater robot considering multiple degrees of freedom such as speed, heading, pitch, and roll; use the model predictive control algorithm to optimize the rotational speed and pitch angle of the thruster in real time, and use the adaptive neuro-fuzzy inference system to establish a non-linear mapping relationship between the buoyancy adjustment amount and the ship speed, heading, and flow velocity; establish a task risk assessment, identify potential risk factors such as equipment failures, communication interruptions, and severe sea conditions, obtain the hazard levels of each risk factor, and propose corresponding prevention and control measures and emergency plans; generate a path planning report according to the above content.

[0025] An underwater robot deployment and recovery method described in the present disclosure has the following advantages:

[0026] First, calculate the energy budget based on the battery capacity and mission duration, and use historical ocean current and tidal data to analyze their flow velocity and periodic change characteristics through Fourier transform to determine the influence on path planning and propulsion systems. Second, real-time monitor the ocean current direction and speed, tidal cycle changes, and tidal level height, and record the vertical profile changes of the flow velocity at different depths and positions. Then, combine the energy budget and real-time monitoring data to judge the favorable or unfavorable conditions of ocean currents and tides, and formulate adjustment strategies for the buoyancy and propulsion systems. For favorable ocean currents, use the autoregressive moving average model to predict the flow velocity changes and adjust the buoyancy device to utilize natural forces to assist movement. For unfavorable ocean currents, use the fuzzy control algorithm and Kalman filter algorithm to adjust the buoyancy and propulsion system power in real time and optimize the deployment and recovery path. Finally, generate a detailed path planning report, including ocean current and tidal influence analysis, buoyancy and propulsion system adjustment suggestions, energy consumption estimation, and operation risks, to ensure the efficient and stable operation of the unmanned underwater robot in a complex environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 is a flowchart of an underwater robot deployment and recovery method described in the present disclosure. Detailed implementation manners

[0028] As Figure 1 shown, a method for deploying and recovering an underwater robot according to the present disclosure includes:

[0029] S101. Determine an energy budget according to the battery capacity and mission duration of the underwater robot; obtain historical ocean current and tide data, determine the influence degree of ocean currents and tides on the path planning of the underwater robot, and determine the influence degree of ocean currents and tides on the propulsion system of the underwater robot;

[0030] Specifically, according to parameters such as the battery capacity and estimated mission duration of the underwater robot, calculate the energy budget value during the entire mission period through an energy consumption model to obtain the upper limit of the available energy of the underwater robot; obtain historical ocean current and tide data of the target sea area and perform preprocessing on the data; perform spectral analysis on the preprocessed historical ocean current and tide data through a fast Fourier transform algorithm to obtain the flow velocity magnitude, flow direction, and periodic change characteristics of ocean currents and tides at different time scales; according to the results of the spectral analysis, judge the influence degree of ocean currents and tides on the movement of the underwater robot.

[0031] More specifically, calculate the energy budget according to the battery capacity and estimated mission duration of the underwater robot. By obtaining historical ocean current and tide data and using Fourier transform to analyze its flow velocity and periodic change characteristics, determine the influence degree of ocean currents and tides on the path planning and propulsion system of the underwater robot;

[0032] According to parameters such as the battery capacity and estimated mission duration of the underwater robot, calculate the energy budget value during the entire mission period through an energy consumption model to obtain the upper limit of the available energy of the underwater robot. Among them, the energy consumption model adopts a prediction model based on historical data and machine learning algorithms. By analyzing the energy consumption situation in historical missions, a non-linear relationship between energy consumption and mission duration, speed, load, and other factors is established, thereby realizing a relatively accurate energy budget estimation. After obtaining the historical ocean current and tide data of the target sea area, use data preprocessing technology to clean, verify, and standardize the data, eliminate noise interference in the data, and improve data quality. Use an improved fast Fourier transform (FFT) algorithm to perform spectral analysis on the preprocessed historical ocean current and tide data. By selecting appropriate time windows and frequency resolutions, obtain the flow velocity magnitude, flow direction, and periodic change characteristics of ocean currents and tides at different time scales. According to the results of the spectral analysis, judge the influence degree of ocean currents and tides on the movement of the underwater robot, determine the ocean current and tide factors to be considered, and incorporate them into the path planning model as constraint conditions.

[0033] For example, in the energy budget estimation, an energy consumption prediction model based on long short-term memory (LSTM) neural network is adopted. By extracting and standardizing the historical mission data, taking the mission duration, average speed, load power and other parameters as input, and taking the actual energy consumption value as output, a three-layer LSTM network is constructed. During the training process, the Adam optimizer is used, the learning rate is set to 0.001, the batch size is 32, and the early stopping and dropout regularization techniques are used, and finally the prediction accuracy of more than 90% is achieved on the test set. For the spectrum analysis of ocean current and tidal data, the improved Cooley-Tukey FFT algorithm is used. By segmenting the data, 128 sampling points are selected as the time window, and the frequency resolution is set to 0.01 Hz. Before the FFT transformation, the data is weighted by the Hanning window to reduce the influence of spectrum leakage. By analyzing the amplitude and phase information in the frequency domain results, key parameters such as the dominant frequency, average flow velocity, and flow direction change period are extracted, which provides an important decision basis for subsequent path planning. In the multi-objective path planning model, three objective functions, namely, range, energy consumption, and sailing time, were considered, and the weight coefficients were set to 0.4, 0.4, and 0.2, respectively. The ocean current and tidal factors were converted into speed and direction constraints, which were combined with the robot's kinematic and dynamic equations to form a mixed integer nonlinear programming problem. The improved non-dominated sorting genetic algorithm (NSGA-II) was used to solve the problem, with the population size set to 100, the crossover probability to 0.8, and the mutation probability to 0.1. The elite retention strategy and the crowding distance operator were introduced, and the global optimal solution was converged after 50 iterations. In the adaptive control of the propulsion system, a fuzzy PID-based controller was designed. With the heading deviation and angular velocity deviation as input, the Mamdani reasoning and center of gravity defuzzification were used to output the speed and rudder angle control of the propeller. In the fuzzy rule base, 25 IF-THEN rules were defined, and the triangular membership function was used. By adaptively adjusting the parameters of the PID controller, the robustness and adaptability under different sea conditions were achieved. After testing, the control strategy can control the heading control error within 2°, and the response time is less than 1 second. The threshold for triggering path replanning is set to flow velocity deviation exceeding 0.5m / s or direction deviation exceeding 30°, which is an empirical value obtained based on statistical analysis of a large amount of historical data. During replanning, a local path optimization algorithm based on artificial potential field is adopted. By constructing virtual gravitational field and repulsive field, the local path is smoothed and optimized while keeping the overall structure of the original path unchanged, so that it can better adapt to the dynamically changing ocean environment.

[0034] S102, obtaining the direction data, speed data and tidal periodic variation data of the ocean current, and analyzing the variation of the vertical profile of the flow velocity at different locations.

[0035] Specifically, a Doppler current profiler is used to monitor the ocean currents in the target sea area in real time to obtain three-dimensional flow field structure data of the entire water body; by measuring the velocity components in different directions, the velocity vectors at each position are obtained, and the distribution characteristics of the ocean currents in the entire area are comprehensively analyzed to determine whether there are significant spatial differences; tide gauge is used to obtain the data of the periodic change of the tide level, and the fast Fourier transform (FFT) is used to perform spectral analysis on the tidal period to determine the main period of the tide; the relationship between the tide level rise and fall obtained by the tide gauge and the velocity change measured by the current profiler is analyzed to obtain the interaction mechanism between the tide and the ocean current; the vertical stratification analysis of the velocity data obtained by the current profiler and the current meter is carried out, and the average velocity and direction of different depth layers are calculated and the vertical profile of the velocity change with depth is drawn.

[0036] More specifically, a Doppler current profiler and an acoustic Doppler current meter are used to monitor the direction and velocity of the ocean current in real time, the tide gauge is used to obtain the data of the periodic change of the tide, the rise and fall of the tide level is recorded, and the vertical profile change of the velocity at different depths and positions is analyzed;

[0037] The Doppler current profiler is used to monitor the ocean currents in the target sea area in real time. By transmitting and receiving high-frequency acoustic signals, the Doppler frequency shift effect is utilized to calculate the flow velocity and direction of different depth layers, and the three-dimensional flow field structure data of the entire water body is obtained. The Doppler frequency shift effect refers to the change in frequency caused by the movement of the medium when the acoustic wave propagates in a moving medium. By measuring the change in frequency, the movement speed and direction of the medium can be deduced. The Doppler current profiler usually operates in the high-frequency range of 300 - 600 kHz, with a measurement range of up to several hundred meters and a vertical resolution of 0.5 - 1 m. Acoustic Doppler current meters are deployed at multiple locations. By measuring the flow velocity components in different directions, the flow velocity vectors at each location are obtained, and the distribution characteristics of the ocean currents in the entire area are comprehensively analyzed to determine whether there are significant spatial differences. The deployment of the acoustic Doppler current meter needs to consider the representativeness and coverage of the measurement points, and usually the equilateral triangle or five-point method is used for deployment to ensure the accuracy and reliability of the measurement. According to the tidal level periodic change data recorded by the tide gauge, the fast Fourier transform (FFT) is used to perform spectral analysis on the tidal period, and the dominant frequency components are extracted through power spectral density estimation to determine the main period of the tide. At the same time, the least squares method is used to perform harmonic analysis on the tidal level data, fitting the harmonic constants of each partial tide to obtain the amplitudes and phases of different partial tides, revealing the non-linear characteristics of the tide. The relationship between the tidal height fluctuations obtained by the tide gauge and the flow velocity changes measured by the current profiler is analyzed through methods such as cross-spectrum and coherence analysis to reveal the interaction mechanism between the tide and the ocean current. The cross-spectrum can estimate the phase difference and time delay between two signals, revealing the causal relationship between them; the coherence analysis can measure the correlation degree between two signals to determine whether they are affected by common influencing factors. The vertical stratification analysis of the flow velocity data obtained by the current profiler and the current meter is carried out. According to parameters such as temperature, salinity, and density, the water body is divided into the surface layer, middle layer, and bottom layer. The average flow velocity and direction of different depth layers are calculated, and a vertical profile of the flow velocity changing with depth is drawn to visually display the stratification structure and vertical gradient characteristics of the flow field.

[0038] Exemplarily, when conducting real-time ocean current monitoring in the target sea area, a Doppler current profiler with a working frequency of 600 kHz, a measurement range of 300 m, and a vertical resolution of 0.5 m is used to estimate the flow velocity and direction by measuring the Doppler frequency shift. Assuming that the acoustic wave frequency changes by 100 Hz, according to the Doppler frequency shift formula Δf = 2vf / c (where v is the flow velocity, f is the acoustic wave frequency, and c is the speed of sound), the corresponding flow velocity can be deduced to be 0.75 m / s. When deploying the acoustic Doppler current meter, the six-point method is adopted, that is, 6 representative measuring points are selected in the target area, and the distance between each measuring point is not less than 500 m to ensure the spatial representativeness and statistical reliability of the data. For the analysis of tidal level data, the FFT algorithm is used for spectral analysis. Assuming that the sampling interval of the tidal level data is 1 hour and a total of 1024 data points are collected, the spectral resolution is 1 / 1024 h ≈ 0.001 h, which can effectively identify the main tidal components such as the semidiurnal tide and the diurnal tide. By calculating the power spectral density, it is found that the frequency of the semidiurnal tide is 0.08 h and the frequency of the diurnal tide is 0.04 h, corresponding to periods of 12.5 h and 25 h respectively. In the harmonic analysis, 8 main tidal components (M2, S2, N2, K2, K1, O1, P1, Q1) are used as basis functions, and the least squares method is used to fit the tidal level data to obtain the harmonic constants of each tidal component. For example, the amplitude of the M2 tidal component is 1.5 m and the phase is 60°. In the correlation analysis of tides and ocean currents, the cross-spectrum and coherence analysis methods are adopted. The calculation formula of the cross-spectrum is Sxy(f) = X(f)Y*(f), where X(f) and Y(f) are the Fourier transforms of the tidal level and flow velocity data respectively, and Y*(f) is the conjugate complex number of Y(f). By analyzing the amplitude and phase of Sxy(f), it is found that there is a significant correlation between the tidal level and the flow velocity at the semidiurnal tide frequency, and the phase difference is about 30°, indicating that the change in the flow velocity lags behind the tidal level by about 1.25 h. The calculation formula of the coherence analysis is Cxy(f) = |Sxy(f)|^2 / [Sx(f)Sy(f)], where Sx(f) and Sy(f) are the auto-power spectral densities of the tidal level and the flow velocity respectively. By calculating Cxy(f), it is found that at the semidiurnal tide and diurnal tide frequencies, the coherence coefficients are both greater than 0.8, indicating that tides are the main factors affecting the change of ocean currents. In the vertical stratification analysis, according to the temperature, salinity, and density data measured by the CTD, the density of seawater is calculated using the UNESCO seawater state equation, and combined with the flow velocity profile data, the water body is divided into the mixed layer, the thermocline layer, and the deep layer. The density gradient of the mixed layer is less than 0.05 kg / m^4, the density gradient of the thermocline layer is greater than 0.2 kg / m^4, and the density gradient of the deep layer is between the two. The analysis shows that the average flow velocity of the mixed layer is 0.6 m / s and the direction is northeast; the average flow velocity of the thermocline layer is 0.3 m / s and the direction is southeast; the average flow velocity of the deep layer is less than 0.1 m / s and the direction change is not obvious.

[0039] The Doppler current profiler measures the vertical profile velocity of water bodies through the Doppler effect of sound waves propagating in water, while the acoustic Doppler current meter can measure the local velocity.

[0040] According to the working principle of the Doppler current profiler, by using the Doppler effect of sound waves propagating in water, the velocity data at different depths on the vertical profile of the water body are obtained. Through the analysis and processing of the velocity data, the velocity distribution characteristics of the vertical profile of the water body are obtained. According to the working principle of the acoustic Doppler current meter, by using the Doppler frequency shift effect of sound waves, the three-dimensional velocity component data of the local water body are obtained. Through the synthesis calculation of the three-dimensional velocity component data, the velocity magnitude and direction of the local water body are obtained. By comparing and analyzing the vertical profile velocity distribution data measured by the Doppler current profiler and the local velocity data measured by the acoustic Doppler current meter, the relationship between the local velocity and the vertical profile velocity is judged, and the position and variation characteristics of the local velocity in the vertical profile velocity distribution are obtained. According to the relationship between the local velocity and the vertical profile velocity, mathematical methods such as interpolation or fitting are used to estimate the extension of the local velocity in the vertical direction, and the continuous distribution data of the local velocity on the vertical profile are obtained, and the variation trend of the local velocity within the entire water depth range is determined. By comprehensively analyzing the measurement data of the Doppler current profiler and the acoustic Doppler current meter, the spatio-temporal distribution characteristics of the water body velocity are obtained, and the variation laws of the velocity in the vertical and horizontal directions are judged, providing data support for further studying the hydrodynamic characteristics of the water body.

[0041] Exemplarily, the Doppler current profiler uses sound waves with a frequency of 600 kHz to propagate in water, and obtains the velocity distribution of the vertical profile of the water body by measuring the Doppler frequency shift at different depths. The formula for calculating the frequency shift is Δf=(2v / c)f, where v is the velocity, c is the sound speed, and f is the sound wave frequency. The echo intensity is proportional to the concentration of suspended particles. By analyzing the change of the echo intensity, the distribution of suspended substances in the water body can be inferred. The acoustic Doppler current meter emits sound waves with a frequency of 2 MHz at an angle of 25° at a depth of 5 m underwater, and calculates the instantaneous velocity vector at this point according to the Doppler frequency shift of the echo and the sound wave propagation time. The measurement results of the two instruments are compared. The velocities measured by the Doppler current profiler at depths of 5 m, 1 m, and 5 m are 8 m / s, 2 m / s, and 5 m / s respectively, and the velocity measured by the acoustic Doppler current meter at 5 m is 48 m / s, with a difference of 3% between the two, indicating that the measurement results have high consistency. Kriging interpolation is performed on the continuously collected velocity data to obtain the continuous distribution image of the three-dimensional flow field of the water body. Through the statistical analysis of the velocity data within 1 hour, the average velocity is obtained as 35 m / s, the velocity change range is 9 - 8 m / s, and the turbulence intensity is 15%, providing an important reference for water environment monitoring and water conservancy project design.

[0042] S103. According to the obtained direction data, velocity data of ocean currents, and periodic change data of tides, combined with the energy budget, determine whether the ocean current and tide conditions are favorable or unfavorable ocean currents, and make adjustment strategies for the propulsion system according to the conditions of ocean currents and tides;

[0043] Specifically, according to the measurement results of an acoustic Doppler current profiler, calculate the velocity and direction of ocean currents in the target sea area. Through the statistical analysis of ocean current vectors, obtain the average velocity, main direction of ocean currents, and probability distribution characteristics of velocities; through a probability distribution model, judge the influence degree of ocean currents on the propulsion efficiency and course-keeping ability of the underwater robot; extract the main periodic components of tides through spectral analysis methods, and establish a harmonic model of tidal level changes; through methods such as the least squares method or Fourier transform, predict the tidal ebb and flow process in a future period of time; conduct statistical analysis on the ebb and flow conditions of tidal levels, calculate the difference between the high tide level and the low tide level within a tidal cycle to determine the amplitude and uncertainty of tidal changes; comprehensively analyze multi-source monitoring data such as velocity profiles, ocean current velocities, and tidal periods, and through data fusion and machine learning algorithms, establish an energy optimization model for the underwater robot. Through the energy optimization model, predict the energy consumption change trend under different speeds, headings, and depths, and obtain the combination of motion parameters with the optimal energy efficiency; according to the prediction results of the energy optimization model, dynamically adjust the buoyancy of the underwater robot; the purpose of adjusting the position of the center of buoyancy is to make the underwater robot reach a balanced state in still water and reduce the coupling effect during dynamic motion. On the premise of ensuring the attitude stability of the underwater robot, minimize the energy loss during the buoyancy adjustment process. Optimize the control strategy of the propulsion system of the underwater robot for different ocean current and tide conditions;

[0044] Obtain parameters such as the battery capacity and discharge characteristics of the underwater robot, and combined with the actual mission requirements, calculate the total available energy and average power limit during the entire mission period through an energy budget model; obtain the resistance characteristics and propeller characteristics of the underwater robot to obtain the resistance curve and propeller efficiency curve at different speeds; according to the ocean current and tide data, combined with the resistance characteristics and propeller characteristics of the underwater robot, through computational fluid dynamics methods, obtain the propulsion efficiency curve and response surface; establish a mapping relationship between thrust demand and shaft power by fitting the propeller characteristic surface, and establish a non-linear mapping relationship between ocean currents, tides, speed, and propulsion efficiency through machine learning algorithms. Through the training and optimization of historical data, obtain a propulsion efficiency prediction model; combine the propulsion efficiency prediction model with the objective function of energy optimization to establish a non-linear programming model, and obtain the optimal navigation strategy parameters for energy utilization efficiency through the non-linear programming model; based on the optimal navigation strategy parameters, adjust the propeller speed, pitch angle, and rudder angle in real time;

[0045] Analyze the ocean current and tidal characteristics for each category, make judgments from both favorable and unfavorable perspectives, and set thresholds for parameters such as flow velocity, flow direction, and tide level; take the energy utilization efficiency and the speed retention rate as the objective functions, and establish a multi-objective optimization model in combination with the propeller characteristic curve, energy consumption model, and historical data of the underwater robot to obtain the optimal speed and course angle thresholds under various environmental conditions. When the measured value exceeds the threshold range, it is considered an unfavorable ocean current, and vice versa for a favorable ocean current.

[0046] More specifically, based on the real-time monitoring data, including the vertical profile changes of the flow velocity at different depths and positions, the direction and speed of the ocean current, the tidal cycle changes, and the rise and fall of the tide level, combined with the energy budget, judge the favorable or unfavorable conditions of the ocean current and tide, and make preliminary adjustment strategies for the buoyancy and propulsion systems.

[0047] Tidal signals are extracted from the original flow velocity data by methods such as low-pass filtering or harmonic analysis to obtain the time series of tidal flow velocity, providing a basis for subsequent tidal parameter calculations. The tidal flow velocity sequence is processed using spectral analysis methods. By extracting the tidal frequency components, the periodic flow velocity changes caused by tides are obtained, and the tidal ellipse parameters are calculated to determine the main axis direction and ellipticity of the tidal current, and to judge the impact degree of tides on navigation. The tidal ellipse refers to the ellipse formed by the trajectory of the endpoints of the tidal current vectors within one tidal cycle. Its major axis represents the main direction of the tidal current, the minor axis represents the amplitude of the tidal current, and the ratio of the major axis to the minor axis is the ellipticity. The tidal ellipse parameters can be estimated by performing principal component analysis or least squares fitting on the tidal current vectors. The real-time change data of the tidal level height is obtained using a tide gauge, and the harmonic constants of each tidal component are extracted through harmonic analysis to establish a tidal level prediction model. Commonly used tidal level prediction models include harmonic analysis method, response method, and numerical model method, etc. Based on the measured tidal level data, methods such as least squares fitting or Kalman filtering can be used to estimate the harmonic constants of each tidal component and obtain the harmonic prediction model of the tidal level. Combining astronomical tides and meteorological factors, the tidal level change trend in the future period is predicted to provide a basis for the formulation of the navigation plan. The available energy of the unmanned underwater vehicle is estimated according to the energy budget model. Considering factors such as battery capacity and the power consumption curve at different speeds, the maximum range and endurance time at different speeds are calculated. By analyzing historical data, an empirical relationship between speed, payload power, range, and endurance time is established, and considering the influence of environmental factors (such as temperature, salinity, etc.), methods such as multiple regression or neural network are used to construct the energy budget model. Combining the prediction results of the tidal cycle and ocean current velocity, the navigation path and time plan are optimized. Thresholds for the course angle and ocean current velocity are set. When the actual values exceed the threshold range, it is determined as an unfavorable navigation condition. The setting of the threshold needs to comprehensively consider factors such as the maneuverability of the underwater vehicle, energy consumption budget, and mission requirements. If the ocean current direction is consistent with the navigation direction and the ocean current velocity is greater than the set threshold, it is judged as a favorable navigation condition, and the speed can be appropriately increased to utilize the driving effect of the ocean current and save the energy consumption of the propulsion system. If the ocean current direction is opposite to the navigation direction and the ocean current velocity is greater than the set threshold, it is judged as an unfavorable navigation condition, and the speed needs to be reduced and the course angle adjusted to minimize the resistance effect of the ocean current. If the vertical velocity profile caused by the tidal cycle change shows an obvious layered structure, at the transition moment between the flood tide period and the ebb tide period, the flow velocities of the upper and lower layers are in opposite directions, forming a strong shear effect, which may have an adverse impact on the attitude stability of the underwater vehicle. By changing the liquid filling amount of the ballast tank or the density of the buoyancy material, the height of the center of buoyancy of the underwater vehicle can be controlled to keep it away from the shear layer and reduce the risk of attitude disturbance. Comprehensively analyzing the real-time monitoring data and the energy budget evaluation results, a fuzzy decision algorithm is used to dynamically adjust the buoyancy and propulsion system parameters of the underwater vehicle.Determine decision variables such as ship speed, course, and height of the center of buoyancy and their universes of discourse, construct membership functions such as triangular and trapezoidal functions, establish a fuzzy rule base in the form of IF-THEN, and through Mamdani or Sugeno inference and defuzzification by the centroid method or the maximum membership degree method, obtain control instructions. On the premise of meeting the requirements of navigation safety and target tracking accuracy, minimize energy consumption, extend the continuous working time of the underwater robot, and improve the success rate of long-range observation tasks.

[0048] Exemplarily, in practical applications, a Doppler velocity profiler with a working frequency of 600 kHz and an acoustic Doppler velocimeter with a frequency of 1 MHz are used to obtain three-dimensional structure data of the flow field with vertical resolutions of 1 m and 0.5 m respectively. The original velocity data is preprocessed using a 5th-order Butterworth low-pass filter with a cut-off frequency set to 1 / 30 Hz to remove high-frequency noise and spike interference. Then, the least squares method is used to fit the tidal current ellipse, and it is found that the main axis direction of the tidal current is northeast-southwest, and the ellipticity is 0.2, indicating that the tidal current exhibits obvious reciprocating flow characteristics. Harmonic analysis is performed on the tide level data to obtain the amplitudes and phase lags of the four main tidal components, namely M2, S2, K1, and O1, and a tide level harmonic prediction model is constructed to predict the hourly tide level change process for the next 30 days, with an average prediction error of less than 10 cm. Based on the hourly tide level and tidal current prediction results, the A* search algorithm is used to plan the optimal navigation path, and the global optimal solution is solved under the conditions of restricted voyage and energy consumption, saving an average of 8% of the voyage and 12% of the energy consumption. In terms of the navigation control strategy, 10° and 0.5 m / s are used as the thresholds for the course angle and ocean current speed respectively. When the angle between the ocean current and the course is less than 10° and the ocean current speed is greater than 0.5 m / s, it is determined as a favorable condition, and the cruising speed is increased by 20%; when the angle between the ocean current and the course is greater than 170° and the ocean current speed is greater than 0.5 m / s, it is determined as an unfavorable condition, and the speed is reduced by 50%, and the course angle is adjusted by 15°. Fuzzy PID control is comprehensively used, with depth, ship speed, and pitch angle as inputs, and through 9 fuzzy control rules, the liquid level of the ballast tank and the angle of the rudder surface are adjusted to suppress vertical and yaw motions and ensure the motion stability of the underwater robot in complex sea conditions.

[0049] Analyze the vertical profile changes of flow velocities at different depths and positions, the direction and speed of ocean currents, the tidal cycle changes, and the rise and fall of tide levels in the real-time monitoring data to determine the adjustment strategy for optimizing the buoyancy and propulsion systems.

[0050] The Doppler current profiler is used to obtain real-time vertical profile data of flow velocities at different depths, revealing the three-dimensional structural characteristics of the flow field and determining the average flow velocity and the main flow direction of each layer. At the same time, an acoustic Doppler velocimeter is used to measure the fixed-point velocity vector to obtain the flow velocity distribution in the horizontal plane. The complementary use of the two instruments can obtain more comprehensive three-dimensional flow field information, providing a basis for the subsequent optimization of the buoyancy and propulsion systems. According to the measurement results of the acoustic Doppler velocimeter, the ocean current velocity and direction in the target sea area are calculated. Through the statistical analysis of the ocean current vectors, the average flow velocity, the main flow direction, and the probability distribution characteristics of the flow velocity of the ocean current are obtained. Common probability distribution models such as normal distribution, Weibull distribution, and Rayleigh distribution are adopted, and methods such as maximum likelihood estimation or moment estimation are used to fit the model parameters to judge the influence degree of the ocean current on the propulsion efficiency and course-keeping ability of the underwater robot. Using the continuous observation data of the tide gauge, the main periodic components of the tide are extracted by spectral analysis methods, and a harmonic model of the tide level change is established. This model represents the tide level time series as the superposition of sine functions of several tidal constituents (such as M2, S2, K1, O1, etc.), and each tidal constituent has its specific frequency, amplitude, and phase. Methods such as the least squares method or Fourier transform are used to fit the harmonic constants to predict the tide ebb and flow process in a future period, providing a reference for the optimization of the power system and energy management of the underwater robot. The statistical analysis of the rise and fall of the tide level is carried out, and the difference between the high tide level and the low tide level within one tidal cycle, that is, the tidal range, is calculated to determine the amplitude and uncertainty of the tidal change. If the tidal range is large and the change is drastic, a larger margin needs to be reserved in the buoyancy adjustment strategy to cope with the pressure change caused by the tide. The multi-source monitoring data such as the flow velocity profile, ocean current velocity, and tidal cycle are comprehensively analyzed, and through data fusion and machine learning algorithms, an energy optimization model of the underwater robot is established. The objective function of the model is defined as minimizing energy consumption or maximizing endurance time, and the constraint conditions are performance indicators such as speed, voyage, and load, as well as environmental condition limitations. Optimization algorithms such as nonlinear programming and dynamic programming are used to solve the model to predict the energy consumption change trend under different speed, course, and depth conditions, and to seek the combination of motion parameters with the optimal energy efficiency. According to the prediction results of the energy optimization model, the ballast state of the buoyancy material of the underwater robot is dynamically adjusted. By controlling the filling and drainage volume of the ballast tank, the vertical distance between the center of buoyancy and the center of gravity and the total buoyancy are changed to adapt to the hydrostatic pressure and flow velocity distribution at different depths. The purpose of adjusting the position of the center of buoyancy is to make the underwater robot reach an equilibrium state in still water and reduce the coupling effect during dynamic motion. On the premise of ensuring the attitude stability of the underwater robot, the energy loss during the buoyancy adjustment process is minimized. The control strategy of the propulsion system of the underwater robot is optimized for different ocean current and tidal conditions.

[0051] Exemplarily, by arranging multiple current meters at different depths and positions, the vertical profile data of the flow velocity is monitored and obtained in real time. For example, at different depths such as 10 meters, 20 meters, and 30 meters underwater, a current meter is arranged every 100 meters, and the sampling frequency is 1 Hz. Using signal processing methods such as wavelet analysis, the noise of the flow velocity data is removed and the features are extracted to obtain the variation curves of the flow velocity at different depths and positions over time. By comparing the magnitudes and variation trends of the flow velocities at different depths, the distribution characteristics of the flow velocity in the vertical direction are analyzed. The three-dimensional flow velocity vector decomposition algorithm is used to decompose the flow velocity into horizontal and vertical components. For example, the flow velocity vector (1, 8, 5) m / s at a depth of 10 meters is decomposed into a horizontal component (1, 8) m / s and a vertical component 5 m / s. This calculation is repeated at different depths and positions to obtain a series of horizontal and vertical flow velocity components. Then, using the vector synthesis algorithm, the components at each point are synthesized into the resultant vector at that position. The direction of the vector is the direction of the ocean current, and the magnitude is the ocean current speed. By arranging tide gauges, the variation of the tide level is continuously observed at a sampling interval of 1 minute. The Fourier transform is performed on the tide level data to obtain the frequency spectrum diagram of the tide level, and the frequency component with the maximum energy is extracted from it. The corresponding period is the main tidal period, usually 12 hours or 24 hours. By analyzing the maximum and minimum values of the tide level data, the range of the tide level rise and fall is obtained, generally about 2 - 3 meters. Considering comprehensively the spatio-temporal distribution characteristics of the ocean current speed and direction, and the tidal period and the law of the tide level rise and fall, the genetic algorithm is used for optimization and solution to obtain an optimal adjustment strategy for the buoyancy and propulsion system parameters. For example, when the ocean current speed is relatively large, the thrust of the thruster can be appropriately increased, and the pitch angle of the underwater robot can be adjusted to reduce the resistance; when the tide level is rising, the buoyancy can be appropriately increased to offset the influence of the tidal force. The optimized strategy can enable the underwater robot to maintain the best navigation attitude and energy efficiency in the complex marine environment.

[0052] Combined with the total available energy and average power limit of the underwater robot, as well as the obtained real-time ocean current and tidal data, the propulsion efficiency under different ocean current and tidal conditions is calculated.

[0053] Obtain parameters such as the battery capacity and discharge characteristics of the underwater robot. Combining with the actual mission requirements, calculate the total available energy and average power limit during the entire mission through the energy budget model, which serves as the basic constraint conditions for the optimization of the propulsion system. The energy budget model is based on parameters such as battery capacity, discharge rate, and power demand, estimating the navigation time and distance under different working conditions, providing a decision-making basis for subsequent energy optimization. Through methods such as pool tests or CFD numerical simulations, obtain the resistance characteristics and propeller characteristics of the underwater robot, getting the resistance curve and propeller efficiency curve at different speeds, providing a basis for the analysis of propulsion efficiency. According to ocean current and tide data, combining with the resistance characteristics and propeller characteristics of the underwater robot, use the computational fluid dynamics (CFD) method to simulate the changes in shaft power and thrust under different combinations of ocean current speed, tidal current speed, and navigation speed, obtaining detailed propulsion efficiency curves and response surfaces. By fitting the propeller characteristic surface, establish the mapping relationship between thrust demand and shaft power. Use machine learning algorithms, such as support vector machine (SVM) or neural network (NN), to establish the non-linear mapping relationship between ocean current, tide, speed, and propulsion efficiency. Through the training and optimization of historical data, obtain the propulsion efficiency prediction model to achieve the rapid estimation of real-time propulsion efficiency. Embed the propulsion efficiency prediction model into the objective function of energy optimization. Taking the minimization of energy consumption or the maximization of the voyage range as the optimization goal, with the navigation speed, heading, and propulsion power as the optimization variables, establish a non-linear programming model including constraint conditions such as speed range, heading angle range, and power upper limit. Use algorithms such as sequential quadratic programming (SQP) or interior point method to solve it, searching for the optimal combination of navigation strategy parameters with the highest energy utilization efficiency. Convert the optimal navigation strategy parameters into control instructions for the propulsion system. Through methods such as model predictive control (MPC) or adaptive control, adjust the propeller speed, pitch angle, and rudder angle in real time, and combine feedback information and observer algorithms to ensure the smoothness and robustness of the control instructions, enabling the underwater robot to autonomously adapt to environmental conditions in complex and variable ocean current and tide environments, select the navigation mode with the highest energy utilization efficiency, and maximize the endurance time and expand the operation range on the premise of ensuring safe and reliable operation.

[0054] Exemplarily, the underwater robot uses a lithium-ion battery pack as the main power source, with a battery capacity of 10 kWh, a rated voltage of 48 V, and a maximum discharge rate of 2C. By analyzing historical mission data, an energy budget model based on parameters such as battery SOC (state of charge), average power, and ambient temperature was established. The multi-linear regression method was used to fit the energy consumption curve, and the average prediction error was less than 5%. Using this model, the total available energy for the current mission was estimated to be 8 kWh, and the average power limit was 2 kW. A resistance test was carried out in a towing tank. By measuring the traction force at different speeds (0.5 - 3 m / s), the resistance curve of the robot was obtained, and the resistance coefficient was fitted to be 0.35 by the least squares method. At the same time, the CFD software ANSYS Fluent was used to numerically simulate the propeller, and the thrust coefficient and torque coefficient curves at different advance coefficients were obtained, and the grid independence of the model was verified. Through data cleaning and feature engineering, 8 characteristic parameters such as speed, heading, flow velocity, flow direction, tide level, and water depth were selected to construct an SVM regression model, and the hyperparameters of the model were optimized by grid search and cross-validation methods. The average relative error on the test set was 3%. The propulsion efficiency model was embedded into the energy optimization objective function. With the goal of maximizing the voyage, the propeller speed, pitch angle, and heading angle were used as optimization variables to establish a non-linear programming model with 5 inequality constraints including speed, power, and maneuverability. The SQP algorithm was used to solve the model, and the optimal navigation strategy parameters for different voyage requirements were obtained, and the effectiveness of the strategy was verified by simulation, with an average energy saving rate of 15%. Finally, the optimal strategy parameters were converted into control commands, and the model predictive control (MPC) algorithm was used to adjust the propeller speed and pitch angle in real time. The control period was 1 s, and the prediction and control steps were 10 steps and 5 steps respectively. The propeller speed was controlled by PID feedback, and the propeller thrust and torque were observed by the Kalman filter algorithm to achieve stable and robust control of the propulsion system, significantly improving the motion efficiency and energy utilization rate of the robot.

[0055] According to the mission points or recovery positions after deployment and real-time ocean current and tide data, thresholds are set for classification. Based on whether the thresholds are exceeded, it is divided into favorable ocean currents and unfavorable ocean currents, and the buoyancy and propulsion systems are adjusted.

[0056] Analyze the ocean current and tidal characteristics for each category, make judgments from both favorable and unfavorable perspectives, and set thresholds for parameters such as flow velocity, flow direction, and tidal level. The setting of the thresholds requires the establishment of a multi-objective optimization model, with the energy utilization efficiency and the speed retention rate as the objective functions, combined with the propeller characteristic curve, energy consumption model, historical data, etc. of the underwater robot. By solving the model, the optimal speed and course angle thresholds under various environmental conditions are obtained. When the measured value exceeds the threshold range, it is considered that the ocean current or tide in this state has a significant impact on the movement of the underwater robot, and the movement strategy needs to be adjusted in a timely manner. If favorable ocean current conditions are detected, appropriately increase the cruising speed of the underwater robot and optimize the course angle to make it consistent with the ocean current direction. According to the working characteristic curve and maneuverability index of the propeller, determine the maximum speed change rate under different working conditions, increase the cruising speed, and keep the deviation of the course angle from the ocean current direction within the threshold. At the same time, adjust the state of the buoyancy device or ballast tank to make the underwater robot dive to a deeper layer with a greater ocean current speed to obtain a greater propulsion gain. If unfavorable ocean current conditions are detected, promptly reduce the cruising speed of the underwater robot and adjust the course angle to form an angle with the ocean current direction. According to the working characteristic curve and maneuverability index of the propeller, determine the maximum speed change rate and course change rate under different working conditions, reduce the cruising speed, and control the course angle change rate within the threshold. At the same time, adjust the state of the buoyancy device or ballast tank to make the underwater robot float to the surface layer with a relatively small ocean current speed, avoid swimming against the current for a long time, and reduce the resistance effect of the ocean current.

[0057] Exemplarily, the underwater robot starts from the deployment point (122.4°E, 29.8°N), needs to pass through two mission points A (122.6°E, 30.1°N) and B (122.9°E, 30.3°N), and finally arrives at the recovery point (123.2°E, 30.6°N), with a total voyage of about 150 km. The A* algorithm is used to globally plan the route. With the minimum total energy consumption as the objective function, considering the influence of environmental factors such as water depth, terrain, and ocean current, an optimal path consisting of 15 waypoints is generated, and the route is divided into 5 sub-segments according to the significant differences in ocean current and tide. The ocean current data received by AIS and deep-sea mooring buoys shows that the northeast ocean current prevails in this sea area all year round, with an average flow velocity of 0.8 m / s, and the coincidence degree between the background field and the measured data reaches more than 90%. Using the DBSCAN clustering algorithm to analyze the measured data, when the parameter ε is taken as 0.2 m / s and minPts is taken as 50, 4 typical ocean current patterns can be obtained, corresponding to forward stable flow, reverse stable flow, vortex flow, and mixed flow respectively. For the forward ocean current, when the flow velocity is greater than 1.2 m / s, the cruising speed is increased from 1.5 m / s to 1.8 m / s, and the heading angle is adjusted to be less than 3° different from the ocean current; for the reverse ocean current, when the flow velocity is greater than 1.5 m / s, the cruising speed is reduced to 1.0 m / s, and the heading angle is adjusted to form an angle of 30° with the ocean current. At the same time, by adjusting the liquid level of the ballast tank, the underwater robot floats up 50 m to reduce the resistance against the reverse current.

[0058] S104. Determine it as a favorable ocean current, predict the changing trend of the flow velocity in the future for a period of time through the autoregressive moving average model, and adjust the buoyancy of the underwater robot in advance according to the prediction result;

[0059] Specifically, when it is determined as a favorable ocean current, control the drainage volume of the ballast tank of the underwater robot to balance the buoyancy and gravity of the underwater robot; at the same time, control the heading angle of the robot to be consistent with the ocean current direction; combine the predicted ocean current speed, direction, and duration information, and obtain the optimal speed and heading control sequence through the objective function of energy optimization, while ensuring the track tracking accuracy, minimizing the energy consumption; monitor the attitude and position of the robot in real time, introduce a trajectory tracking algorithm based on model predictive control, and adjust the control quantity in advance according to the prediction of the motion state in the future for a period of time to reduce the heading deviation caused by ocean current interference and maintain smooth and efficient motion;

[0060] More specifically, for the favorable ocean current, use the autoregressive moving average model to predict the changing trend of the flow velocity in the future for a period of time, and adjust the buoyancy device in advance according to the prediction result to make the robot in a neutral buoyancy state and flow downstream, using the natural force to assist the motion.

[0061] According to the obtained favorable ocean current information, adjust the buoyancy device of the underwater robot in advance. By controlling the drainage volume of the ballast tank, make the buoyancy of the robot balance with the gravity, and be in a neutral buoyancy state. The adjustment range and speed of the drainage volume of the ballast tank need to be estimated and optimized according to the design parameters of the robot, such as the drainage volume range, the volume of the liquid tank, etc. At the same time, optimize the heading angle of the robot to make it consistent with the ocean current direction, and maximize the use of the driving force of the favorable ocean current. The adjustment of the heading angle can be achieved by means of rudder surface control or thrusters. During the favorable ocean current period, appropriately reduce the rotation speed and thrust output of the propeller, reduce the energy consumption of the propulsion system, and use the natural force of the ocean current to assist the robot to move forward. Combining the predicted ocean current speed, direction and duration information, through the objective function of energy optimization, use methods such as dynamic programming and genetic algorithms to solve the optimal speed and heading control sequence, and minimize the energy consumption while ensuring the track tracking accuracy. Real-time monitor the attitude and position of the robot. By introducing a trajectory tracking algorithm based on model predictive control (MPC), according to the prediction of the motion state in the next period of time, adjust the control amount in advance, reduce the heading deviation caused by ocean current interference, and maintain smooth and efficient motion.

[0062] Exemplarily, continuously predict the ocean current speed in the next 6 hours. When the speed at the 4th hour is greater than 0.6 m / s and the confidence level is higher than 90%, it is determined as a favorable ocean current. Control the ballast tank to drain 10 L, make the robot dive 5 m, and adjust the heading angle to an angle less than 5° with the ocean current direction. According to the predicted ocean current speed, reduce the rotation speed of the propeller from 1500 rpm to 1000 rpm, and reduce the thrust by 20%. Introduce MPC-based trajectory control, track the reference trajectory with an accuracy of 0.1 m / s and 1°. The prediction time domain is the next 20 s, and the control time domain is 5 s. By solving the quadratic objective function, obtain the optimal control sequence of the propeller rotation speed and rudder angle in the next 5 s, and execute it iteratively to achieve smooth and efficient trajectory tracking. After 12 hours of continuous navigation, the actual travel distance of the robot is 38 km, the average speed is 0.9 m / s, and the energy consumption is 1.5 kWh. Compared with not using the ocean current, the voyage is increased by 20% and the energy consumption is reduced.

[0063] S105. Judge it as an unfavorable ocean current, and adjust the buoyancy and propulsion system power output of the underwater robot through a fuzzy control algorithm; through tidal data, use the artificial potential field method to dynamically optimize the path;

[0064] Specifically, through the fuzzy control algorithm, a fuzzy control rule table for the buoyancy adjustment amount is obtained; through the Kalman filter combined with the model predictive control (MPC) algorithm, the buoyancy adjustment amount and the propulsion system power in a future period of time are dynamically adjusted to synchronize with the ocean current changes; through the non-linear optimization algorithm, the optimal combination of the buoyancy adjustment amount and the propulsion power is obtained; through the artificial potential field method, according to the deployment position, the predetermined task path and the recovery position of the robot, combined with the real-time ocean current distribution data measured by the ADCP, the global path is segmented and optimized. Using the gradient information of the artificial potential field function as the input of the fuzzy control and the Kalman filter, the dynamic weight adjustment is carried out according to the state error of the underwater robot and the magnitude of the environmental interference, and the weight coefficients of the artificial potential field method and the feedback control law are adjusted in real time.

[0065] More specifically, for adverse ocean currents, the buoyancy is adjusted through the fuzzy control algorithm to make the robot in a positive buoyancy state. The Kalman filter algorithm is used to monitor the flow velocity changes in real time, dynamically adjust the buoyancy magnitude and the propulsion system power output, use the tidal data to calculate the combination of the buoyancy and the propulsion system power, and adopt the artificial potential field method to dynamically optimize the path.

[0066] For adverse ocean currents, a fuzzy control algorithm is adopted. According to the membership functions of ocean current speed and direction, a fuzzy control rule table for buoyancy adjustment amount is obtained. The control rules are in the form of IF-THEN, and the membership functions are in the form of triangles or trapezoids to clarify the value ranges and corresponding relationships of various variables. Through fuzzy inference and defuzzification, an accurate control command for the ballast tank drainage volume is output to achieve rapid buoyancy adjustment of the robot, making it in a positive buoyancy state to offset the resistance of the reverse ocean current. The Kalman filter algorithm is used to filter and estimate the flow velocity measurement data of the ADCP. By establishing a state space model of the ocean current speed, the changing trend of the flow velocity is monitored, predicted, and corrected in real time. The state space model adopts a constant coefficient linear differential equation, such as dx / dt = Ax + Bu + w, where x is the ocean current speed state vector, A is the state transition matrix, B is the control input matrix, u is the buoyancy and thrust control amount, and w is the process noise. The Kalman filter dynamically adjusts the buoyancy adjustment amount and propulsion system power in the next period of time in synchronization with the ocean current changes through prediction and update steps, in combination with the model predictive control (MPC) algorithm. Tide level data provided by the tide table or a nearby tide gauge station is obtained. Through tidal harmonic analysis, characteristic parameters such as the tidal period and tidal range are obtained. Regarding the tide as a periodic environmental disturbance, through advance prediction and active adaptation, on the basis of adverse ocean current control, the energy consumption is further reduced and the voyage range is expanded. Combining the resistance characteristics and propulsion efficiency curve of the robot, a non-linear optimization algorithm, such as the genetic algorithm or particle swarm algorithm, is used to solve the optimal combination of buoyancy adjustment amount and propulsion power. On the premise of meeting the dynamic constraints of the underwater robot, the energy consumption of the entire navigation process is minimized. According to the deployment position, predetermined mission path, and recovery position of the robot, combined with the real-time ocean current distribution data measured by the ADCP, the artificial potential field method is used to segmentally optimize the global path. An artificial potential field function with the current position of the robot as the starting point and the recovery position as the end point is constructed. By superimposing the gravitational potential field and the repulsive potential field, the influence of environmental factors such as tides, ocean currents, and terrain on the potential field is considered at the same time. The gravitational potential field points to the target point, and the repulsive potential field is perpendicular to the obstacle boundary. Through the analytical expression and gradient calculation formula of the potential field function, the expected speed and heading angle of the robot at each position are obtained. Using the gradient information of the artificial potential field function as the input of fuzzy control and Kalman filter, through dynamic weight adjustment, smooth switching between global path planning and local navigation control is achieved. The dynamic weight adjustment adjusts the weight coefficients of the artificial potential field method and the feedback control law in real time according to the state error of the underwater robot and the magnitude of environmental disturbances, ensuring that the robot moves efficiently along the predetermined path.

[0067] Exemplarily, in actual deployment, an ADCP with a working frequency of 600 kHz and 4 acoustic transducers is placed on both sides of an underwater glider, with a measurement range of 30 - 50 m and a speed resolution better than 1 cm / s. When the reverse ocean current speed is continuously detected to be greater than 20% of the glider's nominal speed of 1 knot, i.e., 0.2 knots (0.1 m / s), and the duration exceeds 10% of the glider's endurance time of 6 hours, i.e., 36 minutes, it is determined as a significant adverse ocean current. The input variables of the fuzzy control algorithm are the reverse ocean current speed (0 - 0.5 m / s) and direction (0 - 180°), and the output variable is the buoyancy adjustment percentage (-20% - 20%). A triangular membership function is used, the inference adopts the Mamdani model, and the defuzzification adopts the centroid method. The control rules are in the form of: "IF the ocean current speed is medium AND the ocean current direction is lateral, THEN the buoyancy adjustment amount is small", with a total of 9 rules. Through fuzzy inference, the buoyancy adjustment amount is obtained as 8%, and an instruction is issued to increase the volume of the ballast tank, causing the glider to float up by 0.5 m to offset the influence of the reverse ocean current. In the Kalman filter state space model, the state variables are the ocean current speed and direction, the observation variable is the output of the ADCP, a constant acceleration model is adopted, the state transition matrix is [1, Δt; 0, 1], the observation matrix is [1, 0; 0, 1], the diagonal elements of the initial state covariance matrix are 1, and the diagonal elements of the observation noise covariance matrix are 0.01. The filter is updated once per second to predict the ocean current changes in the next 10 seconds. Combining with the MPC algorithm, the optimal control sequence in the next 30 seconds is calculated every 5 seconds, including the buoyancy adjustment amount and the propeller speed. The objective function is to minimize the weighted sum of the vertical speed and the heading angle deviation, and the constraint conditions are the state variable range, the control variable range, and the glider dynamics equation. According to the current time and position, the tide table is queried to extract the tide level data for the previous and next 1 hour, and harmonic analysis is performed using the fast Fourier transform. The amplitude of the M2 tidal component is obtained as 0.8 m, the period is 12.42 hours, and the average tidal range is 1.5 m. The genetic algorithm is used to optimize the buoyancy adjustment amount and the propeller speed. The chromosome length is 10, the crossover probability is 0.8, the mutation probability is 0.1, and it is iterated 50 times. The fitness function is to minimize the energy consumption, and 10% of the energy can be saved under the action of the tide. In the artificial potential field method, the gravitational potential field function is: Ua(x, y) = 0.5Ka[(x - xt)^2 + (y - yt)^2], and the repulsive potential field function is: Ur(x, y) = 0.5Kr(1 / ρ - 1 / ρ0)^2, where Ka and Kr are the gravitational and repulsive coefficients, xt and yt are the target position coordinates, ρ is the distance to the obstacle, and ρ0 is the radius of the obstacle influence range. Let Ka = 1, Kr = 100, ρ0 = 5 m, and the obstacle position is known. The global path is discretized into 50 waypoints, and the gradient direction of the resultant potential field is calculated for each waypoint as the desired heading.According to the deviation between the underwater robot and the desired heading, dynamically adjust the weights Wapf (0 - 1) of the artificial potential field method and the weight Wfb (0 - 1) of the feedback control law, where Wapf + Wfb = 1. The greater the deviation, the greater Wfb. The USBL system uses a frequency of 18 - 36 kHz, the transducer baseline length is 0.2 m, and the positioning accuracy is better than 0.5% of the water depth.

[0068] S106. Real - time monitor the direction data, velocity data of ocean currents and the periodic change data of tides, and combine with the coordinates of the underwater robot deployment task point and recovery point to generate a path planning report;

[0069] Specifically, obtain the coordinates of the underwater robot deployment task point and recovery point, and preliminarily plan a basic navigation path according to the Euclidean distance and azimuth angle between the two points; use the A* algorithm for global path search to generate a collision - free and blind - spot - free global optimal path; superimpose and analyze the global optimal path with the predicted ocean current field to evaluate the influence degree of ocean currents on the movement of the underwater robot; through numerical simulation or pool tests, obtain the optimal heading angle range under different working conditions; according to the results of the ocean current influence analysis, locally optimize the global path, introduce virtual gravity and repulsion using the artificial potential field method to make the underwater robot stay away from the counter - current area and approach the downstream area, and keep parallel to the ocean current direction; construct a potential field function, reasonably set the combined weights of the gravitational potential field and the repulsive potential field, and superimpose the tidal velocity vector into the potential field gradient. The potential field function is used to make the path planning adapt to the periodic changes of tides; based on the optimized path, formulate adjustment strategies for the buoyancy and propulsion systems according to the dynamic characteristics and maneuverability of the underwater robot; establish a dynamic model of the underwater robot considering multiple degrees of freedom such as speed, heading, pitch, and roll; use the model predictive control algorithm to optimize the rotational speed and pitch angle of the thruster in real - time, and use the adaptive neuro - fuzzy inference system to establish a non - linear mapping relationship between the buoyancy adjustment amount and the ship speed, heading, and flow velocity; establish a task risk assessment, identify potential risk factors such as equipment failure, communication interruption, and bad sea conditions, obtain the hazard levels of each risk factor, and propose corresponding prevention, control measures and emergency plans; generate a path planning report according to the above content.

[0070] More specifically, according to the real - time monitored ocean current direction and velocity, as well as the periodic change data of tides, combine with the coordinates of the underwater robot deployment task point and recovery point to generate a detailed path planning report. The report should include the impact analysis of ocean currents and tides on the robot path, adjustment suggestions for the buoyancy and propulsion systems, estimated energy consumption, and operation risks.

[0071] Obtain the coordinates of the deployment task points and recovery points of the underwater robot. Based on the Euclidean distance and azimuth angle between the two points, preliminarily plan a basic navigation path. Use the A* algorithm for global path search, construct a weighted sum type or product type heuristic function that comprehensively considers multiple factors such as route length, energy consumption, and risk, and make full use of ocean current forecast information to dynamically adjust the navigation costs in different directions and speeds. With minimizing the navigation time and energy consumption as the optimization goal and environmental factors such as ocean current speed, direction, and water depth as the constraint conditions, generate a collision-free and blind-zone-free global optimal path. Superimpose and analyze the global optimal path with the predicted ocean current field to evaluate the influence degree of ocean currents on the movement of the underwater robot. Combine factors such as the maneuvering performance and energy consumption characteristics of the underwater robot, and through numerical simulation or pool tests, obtain the optimal heading angle range under different working conditions. If the included angle between the heading and the ocean current direction is within the optimal range and the forward ocean current speed is greater than a specific threshold, it is considered that there is a favorable propulsion effect, and the speed of the underwater robot can be appropriately increased; conversely, if the included angle is within the most unfavorable range and the reverse ocean current speed is greater than a specific threshold, it is considered that there is a significant resistance effect, and the speed needs to be reduced and the heading adjusted. According to the results of the ocean current influence analysis, locally optimize the global path. Use the artificial potential field method to introduce virtual gravitational and repulsive forces, so that the underwater robot stays away from the countercurrent area, approaches the downstream area, and keeps parallel to the ocean current direction. Construct an exponential or Gaussian potential field function, reasonably set the combined weights of the gravitational potential field and the repulsive potential field, so that the path reaches a balance between obstacle avoidance and trend following. At the same time, superimpose the tidal current velocity vector on the potential field gradient, so that the path planning can adapt to the periodic changes of the tide, and different strategies are adopted during the ebb tide period and the flood tide period, using the tidal current energy to assist the movement of the underwater robot. Based on the optimized path, formulate adjustment strategies for the buoyancy and propulsion systems according to the dynamic characteristics and maneuvering performance of the underwater robot. Establish a dynamic model of the underwater robot considering multiple degrees of freedom such as speed, heading, pitch, and roll, set optimization objective functions such as minimum energy consumption or shortest time, as well as physical constraint conditions such as speed and angle, and use the model predictive control (MPC) algorithm to optimize the rotational speed and pitch angle of the thruster in real time, and minimize the energy consumption on the premise of meeting the speed requirement. Use the adaptive neuro-fuzzy inference system (ANFIS) to establish a non-linear mapping relationship between the buoyancy adjustment amount and the speed, heading, and flow velocity, take the ocean current and tide forecast information as the input of the premise layer, and realize the intelligent control of the buoyancy by training and optimizing the membership function and rule base parameters. Comprehensively consider factors such as voyage, speed, load, and efficiency, estimate the energy demand for the underwater robot to complete the entire task, give the energy consumption prediction values and confidence intervals under different working conditions, and compare them with the available power of the vehicle-mounted battery to judge whether the energy is sufficient.Conduct a comprehensive task risk assessment, identify potential risk factors such as equipment failures, communication interruptions, and adverse sea conditions, and use qualitative and quantitative methods such as fault tree analysis and fuzzy comprehensive evaluation to obtain the hazard levels of each risk factor, and propose corresponding prevention and control measures and emergency plans. Finally, generate a detailed path planning report, which includes necessary information such as the task background, TT&C communication, and logistics support, as well as intuitive decision-making assistance information such as 3D paths, energy consumption curves, and risk radar charts for the review and reference of the command personnel.

[0072] Exemplarily, 5 ADCPs and 10 ADCs are used in the target sea area, which are respectively deployed at 1 main measurement point and 4 auxiliary measurement points, covering an area of about 2,000 square kilometers. 20 layers are set vertically, and the maximum depth reaches the seabed. The working frequency of the ADCP is 300 kHz, the ADC frequency is 1.2 MHz, and the sampling interval is 1 minute for both. The fast Fourier transform (FFT) is used to extract the tidal parameters of 4 main tidal components such as M2, S2, K1, and O1, and a three-dimensional sea current field with a spatial resolution of 100 m and a temporal resolution of 1 h is constructed. By performing anomaly detection and mean filtering on the monitoring data, wild points and high-frequency noise are removed. 500 representative measurement points are randomly selected, and the Kriging interpolation method is used to establish the ocean current background field model, with the interpolation error less than 0.1 m / s. Considering the influence of wind stress and tidal force, based on the Navier-Stokes equation and the k-ε turbulence model, the finite volume method is used for numerical solution, and the ocean current prediction for the next 48 h is carried out on a 1 km grid, with the coincidence degree with the measured data reaching 85%. The A* algorithm is used to search and optimize the global path, and f(n)=g(n)+h(n) is selected as the heuristic function, where g(n) is the navigation time from the starting point to the current node, and h(n) is the ratio of the Euclidean distance from the current node to the target point to the flow velocity. The ocean current data is updated every 10 minutes. When the forward ocean current angle is less than 30° and the flow velocity is greater than 0.8 m / s, the cruising speed is increased by 20%; when the reverse ocean current angle is greater than 150° and the flow velocity is greater than 1.5 m / s, the speed is reduced by 50% and the heading is adjusted by 15°. In the local path optimization, Uatt(x)=Katte^(-x^2 / σ^2) is used as the gravitational potential field function, Urep(x)=Krepe^(-1 / x) is used as the repulsive potential field function, and the resultant potential field function is U(x)=Uatt(x)+ΣUrep(x), where Katt is taken as 100, Krep is taken as 50, σ is taken as 1 / 3 of the length of the underwater robot, and the obstacle distance threshold is 10 m. The tidal current velocity vector is calculated by the homogeneous current dynamics model (HIM) and superimposed with the potential field gradient vector as the desired velocity. Based on the 6-degree-of-freedom dynamics model of the underwater robot, ANFIS is used to establish the mapping relationship between the buoyancy adjustment amount Δb and the ship speed v, the heading angle ψ, and the z-axis angular velocity r: Δb=f(v,ψ,r). The membership function is selected as the Gaussian type, and the training samples are historical mission data, which are updated online as new samples increase. The MPC controller is updated every 5 s, with the sum of the absolute values of the ship speed error as the objective function, the change rates of the propeller speed and the rudder angle as the control variables, the prediction time domain as 20 s, and the control time domain as 5 s. The energy consumption prediction model is: E=f(D,V,T,m,η), where D is the voyage, V is the ship speed, T is the operation time, m is the payload mass, and η is the efficiency coefficient. The model parameters are optimized by the genetic algorithm, and the energy consumption prediction error rate is less than 5%.The fuzzy comprehensive evaluation method is used to evaluate the fault risk. The evaluation indicators include failure rate, fault impact degree, fault detection difficulty, etc. The weights are determined by the analytic hierarchy process, and the risk levels are divided into three levels: high, medium, and low. The path planning report is automatically generated, including chapters such as task overview, measurement and control plan, risk prevention and control, energy consumption budget, emergency plan, etc., and is accompanied by a 3D path animation demonstration, energy consumption curve and risk probability distribution map.

[0073] In the description of the present disclosure, it should be understood that the orientation or positional relationship indicated by orientation words such as "front, back, up, down, left, right", "horizontal, vertical, level" and "top, bottom", etc. is usually based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present disclosure and simplifying the description. Without contrary description, these orientation words do not indicate and imply that the device or element referred to must have a specific orientation or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation on the protection scope of the present disclosure.

[0074] For those skilled in the art, various corresponding changes and deformations can be made according to the technical solutions and concepts described above, and all these changes and deformations should fall within the protection scope of the claims of the present disclosure.

Claims

1. A method for deploying and recovering an underwater robot, characterized in that: include: Determine the energy budget based on the underwater robot's battery capacity and mission duration; obtain historical ocean current and tidal data to determine the extent to which ocean currents and tides affect the underwater robot's path planning; and, determine the extent to which ocean currents and tides affect the propulsion system of an underwater vehicle; Obtain the direction data, speed data and tidal periodic change data of ocean currents, and analyze the changes in the vertical profile of current velocity at different locations; Based on the acquired data on the direction and speed of the ocean currents and the periodic variation of the tides, combined with the energy budget, the ocean currents and tides are judged to be favorable or unfavorable. With energy utilization efficiency and speed retention rate as the objective functions, a multi-objective optimization model is established in combination with the propulsion characteristic curve, energy consumption model, and historical data of the underwater robot to obtain the optimal speed and heading angle thresholds under various environmental conditions. When the measured value exceeds the threshold range, it is considered to be an unfavorable ocean current, otherwise it is a favorable ocean current. Adjustment strategies for the propulsion system are made according to the conditions of the ocean currents and tides. If it is determined that the ocean current is favorable, the displacement of the underwater robot's ballast tank is controlled to balance the buoyancy of the underwater robot with gravity; at the same time, the robot's heading angle is controlled to be consistent with the direction of the ocean current; combined with the predicted ocean current speed, direction and duration information, the optimal speed and heading control sequence is obtained through the energy optimization objective function, while ensuring the track tracking accuracy and minimizing energy consumption; the robot's posture and position are monitored in real time, and by introducing a trajectory tracking algorithm based on model predictive control, the control amount is adjusted in advance according to the prediction of the motion state in the future period, so as to reduce the heading deviation caused by the ocean current interference and maintain smooth and efficient movement; If it is judged to be an unfavorable ocean current, the fuzzy control rule table of the buoyancy adjustment amount is obtained through the fuzzy control algorithm; the buoyancy adjustment amount and the propulsion system power in the future are dynamically adjusted to synchronize with the changes in the ocean current through the Kalman filter combined with the model predictive control algorithm; the optimal combination of buoyancy adjustment amount and propulsion power is obtained through the nonlinear optimization algorithm; the global path is segmented optimized through the artificial potential field method according to the robot's deployment position, predetermined task path and recovery position, combined with the real-time ocean current distribution data measured by ADCP, and the gradient information of the artificial potential field function is used as the input of the fuzzy control and Kalman filter. According to the state error of the underwater robot and the size of the environmental interference, the dynamic weight adjustment is performed, and the weight coefficients of the artificial potential field method and the feedback control law are adjusted in real time; Real-time monitoring of ocean current direction data, speed data and tidal periodic change data, combined with the coordinates of the underwater robot's deployment mission point and recovery point, generates a path planning report.

2. The method for deploying and recovering an underwater robot according to claim 1, characterized in that: The energy budget is determined according to the battery capacity and mission duration of the underwater robot; historical ocean current and tide data are obtained to determine the influence of the ocean current and tide on the path planning of the underwater robot, and the influence of the ocean current and tide on the propulsion system of the underwater robot, including: According to the battery capacity of the underwater robot and the estimated mission duration parameters, the energy budget value during the entire mission is calculated through the energy consumption model to obtain the total available energy of the underwater robot; the historical ocean current and tidal data of the target sea area are obtained and preprocessed; the preprocessed historical ocean current and tidal data are subjected to spectral analysis through the fast Fourier transform algorithm to obtain the velocity, direction and periodic change characteristics of the ocean currents and tides at different time scales; based on the results of the spectral analysis, the degree of influence of the ocean currents and tides on the movement of the underwater robot is determined.

3. The method for deploying and recovering an underwater robot according to claim 1, characterized in that: The acquisition of ocean current direction data, velocity data and tidal periodic variation data, and analysis of the vertical profile variation of the velocity at different locations, includes: The Doppler current profiler is used to monitor the ocean currents in the target sea area in real time to obtain the three-dimensional flow field structure data of the entire water body; by measuring the velocity components in different directions, the velocity vectors at each position are obtained, and the ocean current distribution characteristics of the entire area are comprehensively analyzed to determine whether there are significant spatial differences; the tidal period change data is obtained through the tide meter, and the tidal period is spectrally analyzed using the fast Fourier transform (FFT) to determine the main tidal period; the tidal height fluctuations obtained by the tide meter are correlated with the velocity changes measured by the current profiler to obtain the interaction mechanism between the tide and the ocean current; according to the working principle of the Doppler current profiler, the Doppler effect of sound waves propagating in water is used to obtain the velocity data at different depths on the vertical section of the water body, and the velocity distribution characteristics of the vertical section of the water body are obtained by analyzing and processing the velocity data; according to the working principle of the acoustic Doppler current meter, the Doppler frequency shift effect of sound waves is used to obtain the three-dimensional velocity component data of the local water body, and through The three-dimensional velocity component data are synthesized and calculated to obtain the velocity size and direction of the local water body; the vertical profile velocity distribution data measured by the Doppler velocity profiler and the local velocity data measured by the acoustic Doppler current meter are compared and analyzed to determine the relationship between the local velocity and the vertical profile velocity, and to obtain the position and change characteristics of the local velocity in the vertical profile velocity distribution; based on the relationship between the local velocity and the vertical profile velocity, the local velocity is estimated to be extended in the vertical direction by using interpolation or fitting mathematical methods, to obtain the continuous distribution data of the local velocity in the vertical profile, and to determine the change trend of the local velocity in the entire water depth range; the spatiotemporal distribution characteristics of the water body velocity are obtained by comprehensively analyzing the measurement data of the Doppler velocity profiler and the acoustic Doppler current meter, and the change law of the velocity in the vertical and horizontal directions is determined; the velocity data obtained by the velocity profiler and the current meter are vertically layered and analyzed to calculate the average velocity and direction of different depth layers and draw a vertical profile diagram of the velocity changing with depth.

4. The method for deploying and recovering an underwater robot according to claim 1, characterized in that: The method comprises: judging whether the ocean current and tide conditions are favorable or unfavorable based on the acquired ocean current direction data, speed data, and tidal periodic change data, combined with the energy budget, and making an adjustment strategy for the propulsion system based on the ocean current and tide conditions, including: Based on the measurement results of the acoustic Doppler current meter, the speed and direction of the ocean current in the target sea area are calculated. Through the statistical analysis of the ocean current vector, the average velocity, mainstream direction and probability distribution characteristics of the ocean current are obtained; through the probability distribution model, the influence of the ocean current on the propulsion efficiency and heading keeping ability of the underwater robot is judged; through the spectral analysis method, the main periodic components of the tide are extracted, and a harmonic model of tidal level changes is established; through the least squares method or Fourier transform method, the tidal fluctuation process in the future is predicted; the fluctuation of the tidal height is statistically analyzed, and the difference between the high tide and the low tide in a tidal cycle is calculated to determine the amplitude and uncertainty of the tidal change; comprehensive analysis of the flow Based on multi-source monitoring data of speed profile, ocean current speed and tidal cycle, an energy optimization model of the underwater robot is established through data fusion and machine learning algorithms. The energy optimization model is used to predict the energy consumption trend under different speeds, headings and depths, and obtain the motion parameter combination with the best energy efficiency. The buoyancy of the underwater robot is dynamically adjusted according to the prediction results of the energy optimization model. The purpose of adjusting the center of buoyancy is to make the underwater robot reach a balanced state in still water and reduce the coupling effect during dynamic motion. Under the premise of ensuring the stability of the underwater robot's posture, the energy loss of the buoyancy adjustment process is minimized. The propulsion system control strategy of the underwater robot is optimized for different ocean currents and tidal conditions. The battery capacity and discharge characteristic parameters of the underwater robot are obtained, and in combination with the actual mission requirements, the total available energy and average power limit during the entire mission are calculated through the energy budget model; the resistance characteristics and propeller characteristics of the underwater robot are obtained to obtain the resistance curve and propeller efficiency curve at different speeds; according to the ocean current and tidal data, combined with the resistance characteristics and propeller characteristics of the underwater robot, the propulsion efficiency curve and response surface are obtained through the computational fluid dynamics method; by fitting the propeller characteristic surface, the mapping relationship between thrust demand and shaft power is established, and the nonlinear mapping relationship between ocean current, tide, speed and propulsion efficiency is established through the machine learning algorithm, and the propulsion efficiency prediction model is obtained through training and optimization of historical data; a nonlinear programming model is established by combining the propulsion efficiency prediction model with the objective function of energy optimization, and the optimal navigation strategy parameters for energy utilization efficiency are obtained through the nonlinear programming model; based on the optimal navigation strategy parameters, the propeller speed, pitch angle and rudder angle are adjusted in real time.

5. The method for deploying and recovering an underwater robot according to claim 1, characterized in that: The real-time monitoring of the direction data, speed data and periodic change data of the ocean current is combined with the coordinates of the deployment task point and the recovery point of the underwater robot to generate a path planning report, including: Obtain the coordinates of the deployment mission point and recovery point of the underwater robot, and preliminarily plan the basic navigation path based on the Euclidean distance and azimuth between the two points; use the A* algorithm to perform global path search to generate a global optimal path without collision and blind spots; superimpose the global optimal path with the predicted ocean current field to evaluate the impact of the ocean current on the movement of the underwater robot; obtain the optimal heading angle range under different working conditions through numerical simulation or pool tests; based on the results of the ocean current impact analysis, locally optimize the global path, and use the artificial potential field method to introduce virtual gravity and repulsion to keep the underwater robot away from the upstream area, close to the downstream area, and keep it parallel to the ocean current direction; construct a potential field function, reasonably set the combined weights of the gravitational potential field and the repulsive potential field, and use the tidal velocity as the weight. The degree vector is superimposed on the potential field gradient, and the potential field function is used to enable path planning to adapt to the periodic changes of tides; on the basis of the optimized path, according to the dynamic characteristics and maneuverability of the underwater robot, an adjustment strategy for the buoyancy and propulsion system is formulated; a dynamic model of the underwater robot considering multiple degrees of freedom such as speed, heading, pitch, and roll is established; the model predictive control algorithm is used to optimize the speed and pitch angle of the thruster in real time, and an adaptive neural fuzzy inference system is used to establish a nonlinear mapping relationship between the buoyancy adjustment amount and the speed, heading, and flow rate; a mission risk assessment is established to identify potential risk factors such as equipment failure, communication interruption, and severe sea conditions, obtain the hazard level of each risk factor, and propose corresponding prevention and control measures and emergency plans; a path planning report is generated based on the above content.

Images