Water taking head underwater docking system and construction method

Abstract

The invention discloses an underwater docking system for a water taking head and a construction method, relates to the technical field of underwater engineering and automatic control, and aims to solve the problems of high docking failure rate and low construction efficiency caused by large water flow disturbance, low positioning precision and inaccurate contact force control in an existing underwater docking method. According to the system, relative position information of a water taking head and a butt joint base is obtained in real time through a positioning module, a path calculation module is based on the information and introduces real-time water flow measurement data as compensation amount to calculate a target moving path, and a moving control module controls the water taking head to stably move along the path. The force monitoring and adjusting module monitors the contact force during contact and adjusts movement to complete flexible butt joint. The system is mainly used for installation and maintenance of a water taking head in ocean water taking engineering, high-precision and anti-interference underwater butt joint can be achieved, and construction safety and efficiency are improved.

Description

Technical Field

[0001] This invention belongs to the field of underwater engineering and automatic control technology, specifically relating to an underwater docking system for water intake heads and its construction method. Background Technology

[0002] In the field of underwater engineering, the precise docking between the water intake head and the pre-installed docking base on the seabed is a critical operation. Achieving this process faces several technical challenges, the core of which lies in how to achieve a precise, stable, and safe final approach and connection between the water intake head and the docking base in the complex and dynamic underwater environment.

[0003] First, obtaining high-precision and high-reliability relative position information is fundamental to successful docking. Traditional underwater positioning methods, such as those relying solely on long-baseline or short-baseline underwater acoustic positioning systems, are susceptible to underwater acoustic multipath effects, noise interference, and signal transmission delays, leading to fluctuations and errors in the acquired position data. Especially in the final stages near the docking base, when millimeter-level relative pose accuracy is required, the update rate and accuracy of purely acoustic positioning often fall short. While inertial measurement units (IMUs) can provide high-frequency self-motion data, their measurement errors accumulate and diverge over time. Effectively fusing data from sensors with different characteristics to overcome the limitations of a single sensor and obtain stable, reliable, and accurate six-degree-of-freedom relative pose information is a prerequisite for precise control and also a significant technical challenge.

[0004] Secondly, planning and executing a movement path to the docking point after obtaining location information also faces challenges. The underwater environment experiences continuous or abrupt water flow disturbances, which exert significant forces on underwater objects like the intake head, causing it to deviate from its planned course. If the path planning algorithm does not fully consider the impact of real-time water flow, or only employs simple feedback control, it is difficult to proactively compensate for water flow disturbances, leading to oscillations in the intake head's trajectory, prolonged docking time, and even accidental collisions with the base during approach. Furthermore, the hydrodynamic parameters of the intake head (such as added mass and damping coefficient) are difficult to accurately obtain in the actual operating environment. The inaccuracy of the model further reduces the performance of model-based control algorithms, decreasing the accuracy and stability of path tracking.

[0005] Furthermore, the docking process enters its most critical stage when the water intake head and the docking base make physical contact. At this point, simple displacement control may cause a sharp increase in contact force due to minute positional and angular deviations, resulting in a so-called "jamming" phenomenon or damaging the guiding and sealing mechanisms on the contact surface. A key issue is how to perceive subtle changes in contact force and torque in real time and adjust the water intake head's attitude quickly and smoothly to guide it into the final locking position. Without an effective force control strategy, relying solely on position control is insufficient to adapt to unavoidable docking deviations, potentially leading to docking attempt failures or damage to the equipment.

[0006] In summary, existing technologies for underwater docking missions mainly suffer from insufficient accuracy and reliability in relative position sensing, limited ability to plan and control against dynamic disturbances such as water currents, and a lack of effective management and compliant control of contact forces during the contact phase. These problems restrict the success rate, efficiency, and safety of underwater docking operations. Summary of the Invention

[0007] One object of the embodiments of the present invention is to solve at least the above-mentioned problems and / or defects, and to provide at least the advantages described below.

[0008] Another objective of this invention is to provide an underwater docking system for water intake heads and a construction method thereof.

[0009] To achieve the above-mentioned objectives, the present invention employs the following technical solution: This invention addresses the issues of low docking success rate and susceptibility to collision damage caused by the lack of high-precision, interference-resistant closed-loop control between the water intake head and docking base in complex underwater environments. It also solves the problem that traditional path planning methods do not fully consider real-time water flow disturbances and the dynamic characteristics of the controlled object, resulting in unstable water intake head trajectory, long docking time, and easy deviation from the predetermined path.

[0010] This addresses the issues of insufficient accuracy and susceptibility to interference in close-range docking of single underwater acoustic positioning systems, as well as the inability to provide stable and reliable six-degree-of-freedom relative pose information due to the accumulation of errors in inertial navigation systems.

[0011] This study solved the path planning challenges caused by water flow disturbances and model uncertainties during underwater docking.

[0012] This addresses the problem that inaccurate parameters in the intake head dynamics model caused by changes in the underwater environment and model simplification lead to a decline in the performance of model-based control algorithms and an inability to guarantee path tracking accuracy.

[0013] This addresses the issue that during path tracking, external disturbances and internal system coupling cause deviations between the actual trajectory of the water intake head and the desired path, as well as attitude instability, which affects the final docking accuracy.

[0014] This addresses the issue of a sudden increase in contact force due to slight positional and angular deviations at the moment of contact between the water intake head and the base, which can lead to "jamming" or equipment damage, and the lack of a compliant guiding mechanism to achieve precise locking.

[0015] This addresses the problem of inaccurate parameter estimation caused by measurement noise and data anomalies during online model parameter identification, which affects the stability and reliability of the control system.

[0016] This addresses the problem that online parameter identification algorithms lack robustness when faced with abnormal sensor data or sudden strong disturbances, which may lead to erroneous parameter updates and deterioration of system control performance.

[0017] This addresses the problem of poor system fault tolerance caused by the unreasonable distribution of thrust from the remaining thrusters when some thrusters in an underwater propulsion system malfunction, leading to loss of control of the water intake head or inability to complete basic docking tasks.

[0018] Therefore, the present invention provides the following technical solution: An underwater docking system for water intake heads includes: The positioning module is used to obtain the relative position information between the water intake head and the docking base in real time; A path calculation module is used to calculate the target movement path of the water intake head based on the relative position information. When calculating the target movement path, a flow compensation amount based on real-time flow measurement data is introduced. The path calculation module includes: a dynamic model storage unit for storing the dynamic model of the water intake head; a flow data acquisition unit for acquiring real-time flow measurement data; and a model prediction control unit for predicting the impact of the flow on the trajectory of the water intake head based on the dynamic model and the real-time flow measurement data. With the goal of minimizing docking time and maintaining stable movement, a model predictive control algorithm is used for rolling optimization to calculate the target movement path containing sequential control commands for a future period. The movement control module is used to control the water intake head to move along the target movement path, so that the water intake head moves closer to the docking base; The force monitoring and adjustment module is used to monitor the contact force when the water intake head contacts the docking base, and adjust the movement of the water intake head according to the contact force to complete the docking.

[0019] Preferably, the positioning module includes: a long-baseline underwater acoustic positioning system arranged on the docking base, an acoustic transponder installed on the water intake head, and an inertial measurement unit carried by the water intake head itself; the positioning module is configured to measure distance through the long-baseline underwater acoustic positioning system and the acoustic transponder, and to calculate the relative position information between the water intake head and the docking base by fusing the data from the inertial measurement unit.

[0020] Preferably, the path calculation module includes: a dynamic model storage unit for storing the dynamic model of the water intake head; a water flow data acquisition unit for acquiring real-time water flow measurement data; and a model prediction control unit for predicting the impact of water flow on the trajectory of the water intake head based on the dynamic model and the real-time water flow measurement data, and using a model prediction control algorithm for rolling optimization with the goal of minimizing docking time and maintaining stable movement, to calculate the target movement path containing the sequence control commands for a future period of time.

[0021] Preferably, the path calculation module further includes an online parameter identification unit, which is configured to compare the difference between the predicted motion state of the water intake head and the actual measured motion state in real time, and use the recursive least squares method to estimate and correct the key parameters in the dynamic model online based on the difference, and feed the corrected parameters back to the model prediction control unit for subsequent rolling optimization calculations.

[0022] Preferably, the motion control module includes: a thrust generation unit, used to generate a reference thrust force of the water intake head in the forward direction and a torque for attitude adjustment based on the target movement path; and a trajectory tracking unit, used to acquire the actual movement trajectory and attitude of the water intake head in real time, compare it with the expected value of the target movement path, and dynamically adjust the thrust distribution of each thruster based on the generated trajectory deviation and attitude deviation, so that the water intake head tracks the target path and maintains a stable attitude as it approaches the docking base.

[0023] Preferably, the force monitoring and adjustment module includes: a six-dimensional force sensor, installed on the contact surface between the water intake head and the docking base, for acquiring contact force and torque information; and a force / position hybrid control unit, for controlling the water intake head to perform compliant movement in two translational degrees of freedom perpendicular to the docking direction and three rotational degrees of freedom around the axis based on the contact force and torque information, so as to offset the position and angle deviation during docking, while maintaining a constant micro-force propulsion in the docking axis direction until mechanical locking is completed.

[0024] Preferably, the online parameter identification unit is configured to: collect in real time the actual acceleration information of the water intake head measured by the inertial measurement unit, and the theoretical resultant force calculated by the thruster thrust command; construct an observation equation with the hydrodynamic coefficient in the dynamic model as the parameter to be identified, and use the difference between the theoretical resultant force and the inertial force calculated based on the actual acceleration as the observation vector; use the recursive least squares method with a forgetting factor to estimate the hydrodynamic coefficient online, and feed the real-time updated parameter value back to the prediction model of the model prediction control unit.

[0025] Preferably, the online parameter identification unit further includes a reliability assurance subunit, configured to: perform data consistency verification on the actual acceleration information and the theoretical resultant force; if the difference in magnitude between the two exceeds a preset threshold, it is determined to be data abnormal, the parameter update of the current period is paused, and the parameter estimate of the previous period is used; after the data verification is passed, the innovation sequence of the current observation vector is calculated, and the covariance of the innovation sequence is monitored; when the covariance of the innovation sequence suddenly increases, the forgetting factor of the recursive least squares method is automatically increased to reduce the forgetting rate of old data and enhance the robustness of the algorithm to instantaneous interference.

[0026] Preferably, the motion control module further includes a fault-tolerant control unit configured to: monitor the commanded thrust and actual thrust feedback of each thruster in real time; if the thrust deviation of a certain thruster continuously exceeds the safety threshold and its rotational speed feedback is abnormal, then the thruster is determined to have malfunctioned; based on the position and thrust capability of the remaining normally functioning thrusters, reconstruct the thrust allocation matrix; with the goal of getting as close as possible to the original resultant force and torque, use the weighted least squares method to solve the reconstructed thrust allocation matrix, set higher weights for the thrust and torque allocation priority on the core degrees of freedom, and generate fault-tolerant control commands.

[0027] A method for underwater connection of water intake heads includes the following steps: S1. Construction Preparation and System Deployment: The integration and testing of the water intake head, docking base, and underwater docking control system are completed on the construction mother ship; subsequently, the docking base is hoisted and fixed at the predetermined seabed target position; the underwater docking control system includes a positioning unit, a control unit, and a power unit; wherein, the positioning unit includes a long-baseline acoustic array deployed on the docking base and an acoustic transponder and inertial measurement unit (IMU) installed on the water intake head; the control unit is deployed on the construction mother ship and is communicatively connected to the water intake head via a watertight cable. S2. Relative Position Dynamic Sensing: The positioning system is activated to acquire, in real time, the relative position information between the lowered water intake head and the fixed docking base. Specifically, the method for acquiring this relative position information in real time involves: ranging via the long-baseline acoustic array and acoustic transponder; fusing the water intake head attitude and acceleration data provided by the inertial measurement unit (IMU); and using a filtering algorithm to calculate high-precision three-dimensional relative position and attitude information. S3. Flow Compensation Path Planning: Based on the relative position information and incorporating flow compensation based on real-time marine environmental monitoring data, the target movement path from the intake head to the docking base is calculated. The method for calculating the target movement path adopts the Model Predictive Control (MPC) algorithm. This algorithm is based on the dynamic model of the intake head, with the optimization objective of minimizing docking time and maintaining motion stability. It also incorporates real-time flow data as feedforward compensation and continuously calculates the optimal movement path sequence in the future time domain. Simultaneously, an online model parameter identification process is executed in parallel: the motion data of the water intake head is collected in real time, compared with the predicted values ​​of the dynamic model, the hydrodynamic parameters in the model are estimated and corrected online using the recursive least squares method, and the corrected parameters are fed back to the model predictive control (MPC) algorithm in real time to adapt to changes in the actual marine environment; S4. Adaptive Path Tracking Control: Controls the water intake head to move along the target movement path, and dynamically adjusts the thrust of the thrusters according to the actual movement state of the water intake head during the movement to resist water flow interference and stably approach the docking base; the specific process of dynamically adjusting the thruster thrust is as follows: the actual movement trajectory of the water intake head is compared with the expected value of the target movement path to generate trajectory and attitude deviation; based on this deviation, the thrust command of each thruster is calculated and adjusted in real time through a thrust distribution algorithm, so that the water intake head has the ability to resist current and stabilize direction, and accurately track the target path; The adaptive path tracking control process also includes a thruster fault tolerance step: real-time monitoring of the operating status of each thruster; if a thruster failure is detected, the thrust distribution matrix is ​​immediately recalculated based on the configuration of the remaining normal thrusters, prioritizing control of the core motion degree of freedom and maintaining stable head approximation. S5. Active, compliant, and precise docking: When the water intake head contacts the docking base, the contact force and torque are monitored in real time, and the attitude and position of the water intake head are actively adjusted based on the contact force information to compensate for the final docking deviation until mechanical locking is completed. The specific method of active adjustment based on contact force information is a force / position hybrid control strategy: the contact force and torque are obtained through a six-dimensional force sensor installed on the end face of the water intake head; the water intake head is controlled to make compliant position adjustments in the direction perpendicular to the docking axis and compliant attitude adjustments in the rotational degree of freedom to eliminate misalignment; at the same time, a constant micro-force is maintained in the docking axis direction to ensure smooth engagement of the locking mechanism. S6. Confirmation of docking status and equipment recovery: After confirming that the water intake head and docking base are successfully locked, recover the hoisting device connected to the water intake head and some of the above-water auxiliary equipment to complete the underwater docking construction.

[0028] Compared with the prior art, the advantages and beneficial technical effects of the present invention are: This invention significantly improves the automation and accuracy of underwater intake docking by constructing a complete closed-loop control system integrating high-precision positioning, intelligent path planning, adaptive tracking, and compliant force control. The system effectively overcomes various adverse factors such as water flow disturbances, model uncertainties, and contact impacts, ensuring a smooth and precise docking process. Ultimately, it greatly improves the success rate and reliability of underwater docking operations, reducing reliance on operator experience and the risk of potential equipment damage. By employing a model predictive control algorithm and introducing real-time water flow data as feedforward compensation, path planning is not only based on the current position but also proactively predicts and compensates for the impact of water flow disturbances on the trajectory. The resulting motion path sequence, generated by this optimization method, balances speed and stability, enabling the intake head to approach the docking base more intelligently and smoothly, reducing oscillations and adjustments during the movement process.

[0029] This invention integrates data from a long-baseline underwater acoustic positioning system and an inertial measurement unit, comprehensively utilizing the absolute accuracy of acoustic positioning and the high-frequency characteristics of inertial measurement, effectively suppressing errors and noise from a single sensor. This provides continuous, stable, and high-precision six-degree-of-freedom relative pose information for subsequent control, laying a solid data foundation for the success of the entire docking mission and avoiding control errors caused by inaccurate positioning information.

[0030] The model predictive control unit of this invention effectively solves the problems of unstable intake head trajectory and low docking efficiency caused by the lack of consideration for real-time water flow disturbances and the lack of foresight in traditional path planning methods by introducing a model predictive control algorithm and performing rolling optimization. Based on a dynamic model and real-time water flow data, this unit can predict the impact of water flow in advance and perform multi-objective collaborative optimization with the goal of minimizing docking time and maintaining stable movement. By dynamically generating and executing sequential control commands in each control cycle, it achieves active feedforward compensation for water flow disturbances, enabling the intake head to smoothly and stably approach the docking base, significantly improving the accuracy of path tracking, anti-interference ability, and the level of intelligence of the entire docking process.

[0031] This invention, by introducing an online parameter identification unit, can continuously estimate and correct key parameters in the dynamic model in real time based on the deviation between the actual motion state of the water intake head and the model prediction. This enables the model used by the control algorithm to adaptively approximate the actual dynamic characteristics of the water intake head, significantly improving the accuracy of model predictive control and its adaptability to environmental changes, and ensuring the effectiveness of path planning.

[0032] This invention achieves precise closed-loop tracking of the intake head's trajectory and attitude by comparing the desired path with the actual motion state and dynamically adjusting the thrust distribution of each thruster. This control strategy can quickly respond to external disturbances (such as uneven water flow) and internal coupling effects, ensuring that the intake head always moves stably along the optimal path and maintains a good motion attitude, creating favorable conditions for subsequent precise docking.

[0033] This invention employs a force / position hybrid control strategy, combining force sensing and position control during the contact phase. It allows the intake head to make compliant position and attitude adjustments in non-docking directions to eliminate deviations, while maintaining stable micro-force propulsion along the docking axis. This active compliance mechanism effectively avoids rigid collisions and equipment damage, guiding the intake head smoothly into place and completing mechanical locking, significantly improving the success rate and safety of the final docking stage.

[0034] This invention achieves online real-time estimation of hydrodynamic coefficients by constructing observation equations based on dynamic principles and utilizing recursive least squares with a forgetting factor. This method can continuously refine model parameters using routine measurement data from system operation, making model predictions more closely reflect reality. This provides a more accurate model foundation for advanced control algorithms and improves the overall control performance of the system.

[0035] This invention significantly improves the robustness of the parameter identification process in the face of data anomalies and transient strong disturbances by adding reliability assurance measures such as data consistency verification and adaptive adjustment of the forgetting factor. It can effectively filter unreliable data, prevent erroneous parameter updates, and protect existing identification results when disturbances occur, ensuring the stable operation and reliability of the control system in complex underwater environments.

[0036] This invention achieves fault-tolerant control of the propulsion system by real-time monitoring of the thruster status and rapid reconstruction of the thrust distribution matrix in the event of a failure. It maximizes the utilization of the remaining operational thrusters, prioritizes control of the core motion degrees of freedom, and ensures that the intake head maintains basic stability and controllability even after some thrusters fail, allowing it to continue docking or safely evacuate, thus significantly improving the system's mission reliability and survivability.

[0037] Other advantages, objectives, and features of the embodiments of the present invention will be apparent in part from the following description, and in part will be understood by those skilled in the art through study and practice of the embodiments of the present invention. Detailed Implementation

[0038] To further illustrate the technical means and effects of this invention, the following embodiments are provided for further explanation. The specific embodiments described herein are merely illustrative and not intended to limit the scope of the invention.

[0039] It should be noted that, unless otherwise specified, the experimental methods described in the following implementation plan are all conventional methods, and the reagents and materials described are all commercially available unless otherwise specified.

[0040] According to one embodiment of the present invention, an underwater docking system for a water intake head includes: The positioning module is used to obtain the relative position information between the water intake head and the docking base in real time; The path calculation module is used to calculate the target movement path of the water intake head based on the relative position information. When calculating the target movement path, a water flow compensation amount based on real-time water flow measurement data is introduced. The path calculation module includes: a dynamic model storage unit for storing the dynamic model of the water intake head; a water flow data acquisition unit for acquiring real-time water flow measurement data; and a model prediction control unit for predicting the impact of water flow on the movement trajectory of the water intake head based on the dynamic model and the real-time water flow measurement data. With the goal of minimizing docking time and maintaining stable movement, a model predictive control algorithm is used for rolling optimization to calculate the target movement path containing sequential control commands for a future period. The movement control module is used to control the water intake head to move along the target movement path, so that the water intake head moves closer to the docking base; the force monitoring and adjustment module is used to monitor the contact force when the water intake head contacts the docking base, and adjust the movement of the water intake head according to the contact force to complete the docking.

[0041] Specifically, this system is used in marine water intake projects for the installation and maintenance of intake heads, achieving high-precision docking through the integration of multiple modules. The positioning module employs a combination of a long-baseline underwater acoustic positioning system and an inertial measurement unit (IMU) to acquire the relative position information between the intake head and the docking base in real time. Acoustic beacons of the long-baseline underwater acoustic positioning system are deployed around the docking base, and the acoustic transponders on the intake head respond to the signals. The position is calculated through ranging, while the IMU provides high-frequency attitude and acceleration data. After data fusion, stable six-degree-of-freedom relative pose information is output.

[0042] The path calculation module, based on the aforementioned relative position information, incorporates real-time water flow measurement data as compensation to calculate the target movement path. This module includes a dynamic model storage unit that stores the dynamic model of the water intake head; a water flow data acquisition unit that acquires real-time water flow data from an acoustic Doppler current profiler; and a model prediction control unit that performs rolling optimization using a model predictive control algorithm. The model prediction control unit uses the current state as initial conditions, combined with the dynamic model and real-time water flow data, to predict the impact of water flow on the trajectory over a future period. With minimizing docking time and maintaining stable movement as the objective function, it generates a sequence of control commands as the target movement path. Each control cycle involves re-predicting and optimizing to ensure the path adapts to the dynamic environment.

[0043] The motion control module controls the movement of the intake head according to the target movement path. This module decomposes the path command into a reference thrust and attitude adjustment torque, and dynamically adjusts the thrust of each thruster through a thrust distribution algorithm, so that the intake head stably approaches the docking base along the path. The force monitoring and adjustment module is activated when the intake head contacts the docking base. It monitors the contact force and torque through a six-dimensional force sensor, and uses a force-position hybrid control strategy to adjust the position and attitude of the intake head. It performs compliant movement perpendicular to the docking direction to eliminate deviation, while maintaining a small force propulsion in the docking axis direction until mechanical locking is completed.

[0044] Traditional underwater docking systems rely on single-sensor positioning and simple control strategies. The positioning module uses only a long-baseline underwater acoustic positioning system. Due to the susceptibility of acoustic signals to multipath effects and noise interference, position data updates are slow and inaccurate, failing to provide high-precision six-degree-of-freedom information during close-range docking. Path planning employs preset fixed paths or proportional-integral-derivative (PID) control, neglecting real-time water flow disturbances, leading to deviations in the intake head's trajectory and the risk of oscillations or collisions. The motion control module assigns fixed roles to each thruster, lacking dynamic thrust distribution and failing to effectively resist water flow interference. During the contact phase, there is no dedicated force monitoring module, relying solely on the passive compliance of the mechanical structure or manual adjustments by the operator. This makes it prone to sudden increases in contact force due to positional deviations, potentially causing equipment damage or docking failure.

[0045] Compared with existing technologies, the embodiments of this invention have significant improvements. In terms of positioning, the embodiments provide continuous, high-precision relative pose information by fusing underwater acoustic positioning and inertial measurement data, while existing technologies suffer from unreliable positioning due to the limitations of a single sensor. Regarding path planning, the embodiments employ a model predictive control algorithm to introduce real-time water flow compensation and proactively optimize the path, ensuring smooth and efficient motion, while existing technologies suffer from large path tracking errors due to neglecting water flow disturbances and lacking rolling optimization. In terms of motion control, the embodiments dynamically adjust thrust distribution to enhance anti-interference capabilities, while existing control strategies are rigid and slow to respond. During the contact phase, the embodiments achieve active compliant control through a force monitoring and adjustment module to avoid rigid collisions, while existing technologies lack force sensing and active adjustment, relying on manual intervention and resulting in low success rates.

[0046] This invention's system, through modular integration, solves the problems of low relative position sensing accuracy, poor anti-interference capability of path planning, and insufficient contact force control in underwater docking. The system can acquire high-precision position information in real time and generate an optimized path using water flow compensation and model predictive control, enabling the intake head to smoothly and efficiently approach the docking base. During the contact phase, compliant docking is achieved through force monitoring and adjustment, reducing the risk of collision. Overall, this system improves the automation level, success rate, and safety of underwater docking, while reducing reliance on operator experience.

[0047] According to one embodiment of the present invention, for an underwater docking system for a water intake head, preferably, the positioning module includes: a long-baseline underwater acoustic positioning system arranged on the docking base, an acoustic transponder installed on the water intake head, and an inertial measurement unit carried by the water intake head itself; the positioning module is configured to perform distance measurement through the long-baseline underwater acoustic positioning system and the acoustic transponder, and fuse the data of the inertial measurement unit to calculate the relative position information between the water intake head and the docking base.

[0048] A set of acoustic beacons forming a long-baseline underwater acoustic positioning system with a known geometric configuration can be precisely deployed around a pre-installed docking base on the seabed. An acoustic transponder is installed on the intake head where docking is required. When the system is operational, these seabed beacons sequentially emit acoustic signals, which are received and responded to by the transponders on the intake head. Multiple distance values ​​are calculated by measuring the round-trip time of the signals. Simultaneously, an inertial measurement unit integrated within the intake head continuously measures its own triaxial acceleration and triaxial angular velocity. All this raw measurement data is transmitted to a central processing unit. This unit runs a data fusion algorithm, such as an extended Kalman filter. This algorithm combines the absolute position reference provided by acoustic ranging with the high-frequency, continuous relative motion information provided by the inertial measurement unit. The acoustic data helps correct for drift errors that accumulate over time in inertial navigation, while the inertial data provides smooth motion estimates during the intervals between acoustic data updates and assists in attitude calculation. Through this deep data fusion, the system ultimately outputs a high-update-rate and stable six-degree-of-freedom relative pose information of the intake head relative to the docking base, including three-dimensional position and three-dimensional attitude.

[0049] Existing technologies may rely on only a single type of sensor. For example, a system might use only a long-baseline underwater acoustic positioning system for positioning. Because underwater acoustic signals are susceptible to multipath effects, temperature stratification, and noise interference, the measurement data may exhibit jumps and significant random errors, and cannot provide continuous position information within the data update cycle. Another approach might use only a high-precision inertial measurement unit (IMU) for dead reckoning. Although its data output frequency is extremely high, the sensor's bias and inaccuracies cause position and attitude calculation errors to accumulate and diverge rapidly over time. During docking processes that can last for several minutes or even longer, its positioning accuracy will be completely lost, failing to meet the final centimeter-level docking accuracy requirements. These single positioning methods are all ill-suited to providing continuous, reliable, and high-precision attitude feedback in complex real-world underwater environments.

[0050] This invention cleverly leverages the long-term absolute accuracy of acoustic positioning and the short-term relative accuracy and high update rate of inertial measurement by deeply integrating a long-baseline underwater acoustic positioning system with an inertial measurement unit. This combination effectively suppresses the inherent limitations of a single sensor; acoustic data corrects drift in inertial navigation, and inertial data smooths and compensates for intermittent and abrupt changes in acoustic measurements. This provides a stable, continuous, and reliable relative state estimate for the entire control system, laying a crucial information foundation for subsequent precise path planning and control.

[0051] According to one embodiment of the present invention, for an underwater docking system for a water intake head, preferably, the model prediction control unit in the path calculation module is configured to: perform rolling optimization through a model prediction control algorithm based on the dynamic model and the real-time water flow measurement data, with the objective function of minimizing docking time and maintaining smooth motion, predict the influence of water flow on the trajectory of the water intake head, and calculate the target movement path containing a sequence of control commands for a future period of time; wherein, the model prediction control algorithm executes the following steps in each control cycle: Using the current state of the water intake head as the initial condition, and combining the dynamic model and real-time water flow measurement data, the motion trajectory of the water intake head in the future multiple time steps is predicted; An optimization objective function is constructed, which simultaneously considers minimizing docking time and motion stability index, wherein motion stability is quantified by minimizing motion acceleration or energy consumption; Under constraints, the objective function is solved to generate a series of optimal control command sequences for future time steps; The first or first few instructions in the control instruction sequence are output to the motion control module, and in the next cycle, prediction and optimization are performed again based on the updated state to achieve rolling optimization control.

[0052] Specifically, this unit is the intelligent core of the underwater docking system. Its operation relies on a pre-established dynamic model of the water intake head, which describes the physical relationship between thrust, water flow force, and the acceleration of the water intake head. The unit continuously receives the current status information of the water intake head from the positioning module, including its position, attitude, and velocity, while also receiving real-time water flow data measured by an acoustic Doppler current profiler.

[0053] At the start of each control cycle, the model predictive control unit uses the current state of the intake head as the initial condition for prediction. It combines stored dynamic models with the latest acquired flow measurement data to simulate and predict the possible trajectory of the intake head over multiple consecutive time steps. This predictive process can proactively reveal how flow disturbances will affect the future position and attitude of the intake head.

[0054] Next, this unit constructs a multi-objective optimization function. This function explicitly sets minimizing docking time and motion stability as the core optimization objectives. Motion stability is quantified mathematically, for example, by minimizing changes in acceleration during motion or minimizing overall energy consumption. Solving this objective function requires satisfying physical constraints such as the intake thrust limit and the attitude safety range.

[0055] Under the given constraints, the unit solves the optimization problem, generating a sequence of optimal control commands for future time steps. These commands collectively constitute the target movement path. However, the unit does not execute the entire command sequence; instead, it outputs only the first one or a few commands to the motion control module in real time to drive the thrusters. Upon entering the next control cycle, the unit collects the latest actual state of the intake head and repeats the prediction, optimization, and command output process. This cyclical optimization mechanism based on the latest feedback is called rolling optimization. This allows path planning to continuously adapt to the dynamic changes in the intake head's own state and the external water flow environment, achieving a combination of forward-looking and adaptive intelligent control.

[0056] Traditional path planning methods typically employ preset trajectory tracking or simple proportional-integral-derivative (PID) control strategies. In the preset trajectory method, the intake head's movement path is a fixed curve planned before operation based on ideal conditions. This path cannot incorporate real-time flow measurement data, so when the actual flow direction or velocity deviates from the preset conditions, the intake head will continuously deviate from the course due to a lack of proactive compensation. Another common approach uses a PID controller, which calculates control commands based solely on the instantaneous deviation between the intake head's current position and the target point. Its behavior is lag-dependent and reactive. When continuous flow disturbances exist, this type of controller can only passively compensate after the position deviation occurs, easily leading to repeated oscillations of the intake head near the target path, or requiring extremely high control gain to resist strong disturbances, thus causing motion instability or even instability. Both of these traditional methods lack the ability to predict the future dynamics of the system and do not explicitly optimize docking time and motion stability as multiple objectives.

[0057] The model predictive control employed in this embodiment is an active predictive control strategy combining feedforward and feedback. It utilizes a model to anticipate the future impact of water flow disturbances and issues compensation commands in advance, thereby making the intake head's movement more intelligent and smoother. In contrast, existing technologies, whether fixed-path or proportional-integral-derivative (PID) control, are passively reactive and cannot overcome the inherent lag of the control system, resulting in weak anti-interference capabilities. Furthermore, this embodiment, through a carefully designed objective function, simultaneously pursues a balance between docking efficiency and motion quality, while existing technologies typically only focus on eliminating instantaneous position errors, neglecting the smoothness of the motion process itself, potentially leading to excessively long docking times or motion shocks. Finally, this embodiment employs a rolling optimization mechanism to continuously update the planned path, adapting to the complex and ever-changing underwater environment, while existing fixed-path or simple feedback control technologies lack this online replanning capability and have poor adaptability.

[0058] The model prediction control unit defined in this invention provides powerful intelligent path planning capabilities for underwater docking systems. Its core function lies in significantly improving the accuracy and stability of the intake head's movement when facing dynamic water flow disturbances through forward-looking prediction and rolling optimization. This unit uses real-time water flow data as a feedforward, enabling the system to actively counteract rather than passively respond to environmental disturbances. Simultaneously, its multi-objective optimization characteristics ensure that the intake head can quickly approach the docking base while maintaining a stable motion attitude, avoiding unnecessary oscillations and abrupt adjustments. The sequential control commands generated by this optimization calculation make the intake head's movement more coherent and controllable. Ultimately, the implementation of this unit greatly enhances the adaptability, robustness, and intelligence of the entire docking system in complex underwater environments, laying a solid foundation for subsequent precise docking.

[0059] According to one embodiment of the present invention, for an underwater docking system for a water intake head, preferably, the path calculation module further includes an online parameter identification unit, configured to compare in real time the difference between the predicted motion state and the actual measured motion state of the water intake head, and based on the difference, use the recursive least squares method to estimate and correct the key parameters in the dynamic model online, and feed the corrected parameters back to the model prediction control unit for subsequent rolling optimization calculations.

[0060] For example, this unit operates continuously as part of the path calculation module. It receives real-time acceleration measurements from the inertial measurement unit, records thrust commands sent to each thruster by the controller, and calculates the theoretical resultant force and torque acting on the intake head based on the thruster model. Internally, this unit constructs an observation model whose parameters to be identified are key hydrodynamic coefficients in the intake head's dynamics model, such as added mass and damping coefficients. The input to the observation model is the difference between the theoretical resultant force and the inertial force calculated from the actual acceleration; this difference is considered to be primarily contributed by hydrodynamic forces. The unit employs a recursive least squares method with a forgetting factor, iteratively updating the estimates of these hydrodynamic coefficients using continuously inputting fresh observation data. These real-time updated, more accurate parameter estimates are immediately fed back to the model prediction control unit to update its internal dynamics model used for prediction, thus ensuring that the next rolling optimization calculation is based on a model that better reflects the current true motion characteristics of the intake head.

[0061] Existing technologies may use a fixed set of hydrodynamic parameters identified under ideal laboratory conditions in model predictive control algorithms. However, when a water intake is operating underwater, its hydrodynamic characteristics vary with factors such as water depth, proximity to the seabed, and attached organisms. Using fixed, inaccurate model parameters leads to distortion of the predictive model, resulting in a systematic deviation between the model-predicted control trajectory and the actual dynamic response of the water intake. This deviation significantly reduces the effectiveness of model-based optimization control and may even degrade control performance due to model mismatch, causing the planned path to still produce significant errors during actual tracking.

[0062] This invention introduces a parallel online parameter identification process, endowing the control system with self-learning and self-correction capabilities. It can automatically and continuously adjust its internal model to approximate the dynamic characteristics of the intake head in a real operating environment. This adaptive mechanism effectively overcomes the control performance degradation caused by model mismatch, ensuring that the prerequisite for model predictive control algorithms—a relatively accurate predictive model—is always maintained. This significantly improves the robustness and control accuracy of the entire path planning system in complex, time-varying underwater environments, allowing the advantages of intelligent path planning to be fully realized. According to one embodiment of the present invention, for an underwater docking system for a water intake head, preferably, the movement control module includes: a thrust generation unit, used to generate a reference thrust force of the water intake head in the forward direction and a torque for attitude adjustment based on the target movement path; and a trajectory tracking unit, used to acquire the actual movement trajectory and attitude of the water intake head in real time, compare it with the expected value of the target movement path, and dynamically adjust the thrust distribution of each thruster according to the generated trajectory deviation and attitude deviation, so that the water intake head tracks the target path and maintains a stable attitude as it approaches the docking base.

[0063] A specific implementation of a motion control module in an underwater docking system for a water intake head is as follows: This module receives target motion path information from a path calculation module. This information includes a series of positions and attitude points that the water intake head is expected to reach within a future period. The thrust generation unit first analyzes these path points, calculating the reference thrust required by the water intake head in the forward direction and the adjustment torque used to counteract rotational deviations in order to achieve the current cycle's motion target. Subsequently, the trajectory tracking unit begins operation, continuously acquiring the actual motion trajectory and real-time attitude of the water intake head provided by the positioning module and comparing them with the expected path points. This comparison process generates trajectory and attitude deviations. Based on these deviations, the control algorithm does not directly command individual thrusters. Instead, it uses a thrust allocation algorithm that comprehensively considers the position, thrust direction, and thrust capability of all thrusters, dynamically decomposing the required resultant force and torque into specific thrust commands for each thruster with certain optimization objectives, such as minimizing total energy consumption or avoiding thruster saturation. This allocation is dynamic and can be adjusted in real time according to the actual movement of the water intake head, ensuring that it can closely track the target path while maintaining its own stable attitude, thus smoothly approaching the docking base.

[0064] Existing technologies may employ relatively simple control allocation strategies. For example, the system might design an independent controller for each degree of motion and assign a fixed control role to each thruster, such as one thruster specifically controlling forward movement and another specifically controlling steering. This rigid allocation method cannot fully utilize the combined capabilities of all thrusters. When encountering uneven water flow disturbances or performance differences between thrusters, it is difficult to achieve accurate trajectory tracking and stable attitude maintenance, easily leading to motion incoordination and energy waste. Another approach might lack fine-grained attitude feedback and control, focusing only on position tracking. This could cause unnecessary tilting or deflection of the intake head during approach, introducing additional angular deviations for the final docking.

[0065] This invention achieves coordinated and precise control of the six-degree-of-freedom motion of the intake head by tightly integrating path tracking with dynamic thrust distribution. It not only ensures that the intake head follows a predetermined spatial trajectory but also actively maintains its stable attitude. This integrated control method effectively addresses underwater environmental disturbances and internal system coupling effects, making the intake head's approach process more precise and smooth, and providing a stable and reliable final state for subsequent docking operations.

[0066] According to one embodiment of the present invention, for an underwater docking system for a water intake head, preferably, the force monitoring and adjustment module includes: a six-dimensional force sensor, installed on the contact surface between the water intake head and the docking base, for acquiring contact force and torque information; and a force / position hybrid control unit, for controlling the water intake head to perform compliant movement in two translational degrees of freedom perpendicular to the docking direction and three rotational degrees of freedom about three axes based on the contact force and torque information, so as to offset the position and angle deviations during docking, while maintaining a constant micro-force propulsion in the docking axis direction until mechanical locking is completed.

[0067] A specific implementation of a force monitoring and adjustment module for an underwater docking system for water intake heads is as follows: When the water intake head initially contacts the docking base under the drive of the motion control module, a six-dimensional force sensor installed on the docking end face of the water intake head begins to function. It accurately measures the contact force components in three directions and the contact torque components in three directions. This force and torque information is transmitted in real time to the force-position hybrid control unit. Based on this, the control unit determines the relative pose deviation between the water intake head and the docking base. Its control strategy is that, in the two translational degrees of freedom perpendicular to the docking axis, a fixed position command is no longer forcibly maintained; instead, contact force feedback is allowed to guide the water intake head to make compliant lateral movements to eliminate center misalignment. Similarly, in the rotational degrees of freedom about the three axes, contact torque feedback is also allowed to guide the water intake head to make compliant attitude rotations to correct angular deviations. Through this compliant adjustment in five degrees of freedom, the water intake head can adaptively align and slide into the guide mechanism of the docking base. Meanwhile, in the crucial docking axis direction, the system maintains a constant and small thrust, which is sufficient to overcome the sealing resistance and ensure that the locking mechanism can continue to engage until the final mechanical locking is completed.

[0068] Existing technologies may still employ a purely position-control mode during the contact phase. When the sensor detects an increase in contact force, the system may simply attempt to fine-tune its position setpoint. However, this sluggish, position-loop-based adjustment is cumbersome in dealing with rapidly changing contact forces, easily generating force oscillations, causing repeated start-stop cycles during the docking process, and even damaging guide components due to excessive contact forces. Another approach may completely lack active force control, relying solely on the passive compliance of the mechanical structure and the operator's experience. Its docking success is highly accidental and carries a high risk of equipment collision.

[0069] This invention achieves proactive, compliant guidance during the docking contact phase by introducing a force-position hybrid control strategy based on a six-dimensional force sensor. It enables the water intake head to sense and intelligently respond to contact forces, proactively eliminating minute deviations in position and angle, thereby transforming a potential rigid collision into a compliant sliding process. This mechanism significantly reduces reliance on the precision of mechanical structures and the experience of operators, substantially improving the success rate and safety of the final docking phase, and effectively protecting the equipment from damage.

[0070] According to one embodiment of the present invention, for an underwater docking system for a water intake head, preferably, the online parameter identification unit is configured to: collect in real time the actual acceleration information of the water intake head measured by the inertial measurement unit, and the theoretical resultant force calculated by the thruster thrust command; construct an observation equation with the hydrodynamic coefficients in the dynamic model as the parameters to be identified, and use the difference between the theoretical resultant force and the inertial force calculated based on the actual acceleration as the observation vector; use a recursive least squares method with a forgetting factor to estimate the hydrodynamic coefficients online, and feed back the real-time updated parameter values ​​to the prediction model of the model prediction control unit.

[0071] A specific implementation of an online parameter identification unit for an underwater intake docking system is as follows: This unit continuously acquires data, including the actual acceleration vector of the intake head in three-dimensional space measured by the inertial measurement unit, and the theoretical resultant force vector calculated based on thrust commands sent to all thrusters and known thruster models. The unit constructs a system observation equation, using hydrodynamic coefficients to be identified in the dynamic model, such as velocity-related damping coefficients, as parameters to be solved. The observation vector of the observation equation, i.e., one side of the equation, is the difference between the theoretical resultant force vector and the inertial force vector calculated based on actual acceleration and mass. This difference should theoretically be equal to the hydrodynamic force. The unit uses a recursive least squares method with a forgetting factor to solve this observation equation. This algorithm can recursively update the estimates of the hydrodynamic coefficients using a continuous stream of new measurement data, while gradually diminishing the influence of old data through the forgetting factor, making the estimation results more reflective of the current dynamic characteristics of the system. These real-time updated and more accurate hydrodynamic coefficients are immediately fed back into the predictive model of the model predictive control unit for rolling optimization calculations in the next control cycle.

[0072] Existing technologies may rely entirely on fixed model parameters identified offline. These parameters may be obtained through specialized system identification experiments in test tanks, but they cannot reflect the changes in the dynamic characteristics of the intake head in a real marine environment caused by biological attachment, proximity to the seabed, or changes in water flow conditions. Using these invariant parameters will lead to persistent errors in the predictive model, causing the optimization of model predictive control to be based on a distorted system model. The optimality of the planned paths and control commands cannot be guaranteed, and it may even lead to stability problems in actual execution.

[0073] This invention achieves online, real-time identification of key hydrodynamic parameters by constructing observation equations based on dynamic principles and applying a recursive estimation algorithm. This method can automatically and continuously correct the model using routine data generated during normal system operation, enabling the predictive model to continuously approximate the actual dynamic behavior of the intake head. It elevates model predictive control from a controller dependent on a prior fixed model to an adaptive controller with online self-learning capabilities, thereby significantly enhancing the control accuracy and robustness of the system in complex time-varying environments.

[0074] According to one embodiment of the present invention, for an underwater docking system for water intake heads, preferably, the online parameter identification unit further includes a reliability assurance subunit, configured to: perform data consistency verification on the actual acceleration information and the theoretical resultant force; if the difference in magnitude between the two exceeds a preset threshold, it is determined to be data abnormal, the parameter update of the current cycle is paused, and the parameter estimate of the previous cycle is used; after the data verification is passed, the innovation sequence of the current observation vector is calculated, and the covariance of the innovation sequence is monitored; when the covariance of the innovation sequence suddenly increases, the forgetting factor of the recursive least squares method is automatically increased to reduce the forgetting rate of old data and enhance the robustness of the algorithm to instantaneous interference.

[0075] A specific implementation of a reliability assurance subunit for an online parameter identification unit in a water intake underwater docking system is as follows: Before each parameter update calculation, the subunit first performs a data consistency check. It compares the actual acceleration information to be used with the theoretical resultant force information by orders of magnitude. If an extreme anomaly is found in the data of a certain axis, such as a very small theoretical resultant force but a very large measured acceleration, or vice versa, and the difference exceeds a preset reasonable threshold based on the system's physical characteristics, then the data for this cycle is determined to be unreliable. At this time, the subunit will pause the parameter update process for this cycle, and the control unit will continue to use the parameter estimates obtained from the previous control cycle for model prediction to avoid erroneous data contaminating the identification results. After the data passes the preliminary verification, the subunit will calculate the innovation sequence of the current observation vector, that is, the difference between the actual observed value and the model predicted value. It continuously monitors the statistical characteristics of this innovation sequence, especially its covariance. When a sudden increase in the innovation covariance is detected, this usually indicates that the system has been subjected to strong instantaneous external disturbances or unmodeled dynamics, and the subunit will automatically increase the forgetting factor in the recursive least squares algorithm. Increasing the forgetting factor means that the algorithm will give higher weight to historical data when updating parameters, thereby reducing the sensitivity to recent abnormal data. This enhances the robustness to transient disturbances and prevents drastic changes in parameter estimates.

[0076] Existing technologies may lack such proactive reliability assurance mechanisms during online parameter identification. Their recursive least squares method may always operate with a fixed, small forgetting factor, continuously and rapidly tracking changes in all input data. When instantaneous anomalies or strong disturbances occur in sensor data, this algorithm indiscriminately absorbs the abnormal information into the parameter estimation, leading to significant and erroneous parameter drift. These erroneous parameters are immediately fed back to the controller, potentially triggering abrupt changes in control commands, thereby affecting the stability of the intake head movement and even inducing system oscillations.

[0077] This invention constructs a safety barrier for the online parameter identification process by adding data consistency verification and an adaptive forgetting factor adjustment mechanism. It can effectively identify and isolate unreliable data, and intelligently adjust the algorithm's learning rate when sensing drastic environmental changes, thereby significantly improving the robustness and reliability of the parameter estimation process in complex and uncertain underwater environments. This ensures the stable operation of the entire adaptive control system and avoids the risk of control performance degradation due to instantaneous data anomalies.

[0078] According to one embodiment of the present invention, for an underwater docking system for a water intake head, preferably, the motion control module further includes a fault-tolerant control unit configured to: monitor the commanded thrust and actual thrust feedback of each thruster in real time; if the thrust deviation of a certain thruster continuously exceeds a safety threshold and its rotational speed feedback is abnormal, then determine that the thruster has malfunctioned; reconstruct the thrust allocation matrix based on the position and thrust capability of the remaining normally functioning thrusters; and solve the reconstructed thrust allocation matrix using a weighted least squares method with the goal of approximating the original resultant force and torque as closely as possible, setting higher weights for the priority of thrust and torque allocation in the core degrees of freedom, and generating fault-tolerant control commands.

[0079] This unit continuously monitors the operating status of all thrusters. It compares the commanded thrust sent to each thruster with the actual thrust fed back by that thruster, while also monitoring the thruster's rotational speed signal. If a thruster's thrust deviation exceeds the set safety threshold for multiple control cycles, and its rotational speed feedback also shows abnormalities, such as jamming or overspeeding, then that thruster is considered to have malfunctioned. Once the fault is confirmed, the fault-tolerant control unit immediately initiates a reconfiguration procedure. It first reconstructs a thrust allocation matrix based on the spatial position and thrust direction of the remaining normally functioning thrusters. This matrix describes how the remaining thrusters combine to generate resultant force and torque. Subsequently, using the original desired resultant force and torque to be achieved as the objective, it solves for this new thrust allocation matrix using the weighted least squares method. During the solution process, higher weights are assigned to thrust and torque in the core degrees of freedom, such as ensuring that the forward direction and basic attitude stability are maintained first, thereby generating a set of fault-tolerant control commands for the remaining normally functioning thrusters, so that the intake head can maintain the best possible controllability even if the thrust is impaired.

[0080] Current technology may lack this proactive fault-tolerant control capability. When the system detects a thruster malfunction, it might simply shut it down and continue using the existing thrust distribution logic designed for a full-function propulsion system. This would cause the resulting resultant force and torque to deviate significantly from expectations, leading to the water intake head yawing, rotating, or even reversing uncontrollably, making the docking mission completely impossible. Operators might then have to abruptly terminate the mission, retrieving the water intake head to the surface for repairs, resulting in significant operational disruption and time costs.

[0081] This invention, by integrating real-time fault diagnosis and thrust distribution reconfiguration capabilities, endows underwater docking systems with the ability to continue operation even in the event of partial thruster failure. It can quickly identify faults and intelligently reconfigure the output of the remaining thrusters to preserve core motion functions to the greatest extent possible. This fault-tolerant mechanism significantly improves the system's mission reliability and survivability, enabling docking operations to continue even in the face of partial equipment failure, or at least maintain a stable state for safe evacuation, avoiding mission failure due to a single component malfunction.

[0082] A method for underwater connection of water intake heads includes the following steps: S1. Construction Preparation and System Deployment: The integration and testing of the water intake head, docking base, and underwater docking control system are completed on the construction mother ship; subsequently, the docking base is hoisted and fixed at the predetermined seabed target position; the underwater docking control system includes a positioning unit, a control unit, and a power unit; wherein, the positioning unit includes a long baseline acoustic array deployed on the docking base and an acoustic transponder and inertial measurement unit (IMU) installed on the water intake head; the control unit is deployed on the construction mother ship and is communicatively connected to the water intake head via a watertight cable.

[0083] S2. Relative position dynamic perception: Activate the positioning system to obtain the relative position information between the lowered water intake head and the fixed docking base in real time; wherein, the specific method for obtaining the relative position information in real time is as follows: distance is measured by the long baseline acoustic array and acoustic transponder, and the attitude and acceleration data of the water intake head provided by the inertial measurement unit (IMU) are fused, and high-precision three-dimensional relative position and attitude information is obtained by filtering algorithm.

[0084] S3. Flow Compensation Path Planning: Based on the relative position information and incorporating flow compensation based on real-time marine environmental monitoring data, the target movement path from the intake head to the docking base is calculated. The method for calculating the target movement path adopts the Model Predictive Control (MPC) algorithm. This algorithm is based on the dynamic model of the intake head, with the optimization objective of minimizing docking time and maintaining motion stability. It also incorporates real-time flow data as feedforward compensation and continuously calculates the optimal movement path sequence in the future time domain. Simultaneously, an online model parameter identification process is executed in parallel: the motion data of the water intake head is collected in real time, compared with the predicted values ​​of the dynamic model, the hydrodynamic parameters in the model are estimated and corrected online using the recursive least squares method, and the corrected parameters are fed back to the model predictive control (MPC) algorithm in real time to adapt to changes in the actual marine environment; The model prediction control unit in the path calculation module is configured to: perform rolling optimization based on the dynamic model and the real-time water flow measurement data, using a model prediction control algorithm to minimize docking time and maintain smooth movement as the objective function, predict the impact of water flow on the trajectory of the water intake head, and calculate the target movement path containing a sequence of control commands for a future period; wherein the model prediction control algorithm executes the following steps in each control cycle: Using the current state of the water intake head as the initial condition, and combining the dynamic model and real-time water flow measurement data, the motion trajectory of the water intake head in the future multiple time steps is predicted; An optimization objective function is constructed, which simultaneously considers minimizing docking time and motion stability index, wherein motion stability is quantified by minimizing motion acceleration or energy consumption; Under constraints, the objective function is solved to generate a series of optimal control command sequences for future time steps; The first or first few instructions in the control instruction sequence are output to the motion control module, and in the next cycle, prediction and optimization are re-performed based on the updated state to achieve rolling optimization control. S4. Adaptive Path Tracking Control: Controls the water intake head to move along the target movement path, and dynamically adjusts the thrust of the thrusters according to the actual movement state of the water intake head during the movement to resist water flow interference and stably approach the docking base; the specific process of dynamically adjusting the thruster thrust is as follows: the actual movement trajectory of the water intake head is compared with the expected value of the target movement path to generate trajectory and attitude deviation; based on this deviation, the thrust command of each thruster is calculated and adjusted in real time through a thrust distribution algorithm, so that the water intake head has the ability to resist current and stabilize direction, and accurately track the target path; The adaptive path tracking control process also includes a thruster fault tolerance step: real-time monitoring of the working status of each thruster; if a thruster malfunction is detected, the thrust distribution matrix is ​​immediately recalculated based on the configuration of the remaining normal thrusters, prioritizing control of the core motion degrees of freedom to maintain the stable approach of the intake head; S5. Active compliant and precise docking: when the intake head contacts the docking base, the contact force and torque are monitored in real time, and the attitude and position of the intake head are actively adjusted according to the contact force information to compensate for the final docking deviation until mechanical locking is completed; the specific method of active adjustment according to the contact force information is a force / position hybrid control strategy: the contact force and torque are obtained through a six-dimensional force sensor installed on the end face of the intake head; the intake head is controlled to make compliant position adjustments in the direction perpendicular to the docking axis and compliant attitude adjustments in the rotational degrees of freedom to eliminate misalignment; at the same time, a constant micro-force propulsion is maintained in the docking axis direction to ensure smooth engagement of the locking mechanism; S6. Docking Status Confirmation and Equipment Recovery: After confirming that the intake head and docking base are successfully locked, recover the hoisting device connected to the intake head and some surface auxiliary equipment to complete the underwater docking construction. The docking status confirmation is performed by verifying the status signal of the locking mechanism fed back by the acoustic beacon, combined with the tension monitoring of the docking cables on the mother ship's deck.

[0085] A specific implementation method for underwater docking of a water intake head is as follows: The entire process begins on the construction mother ship, where the integration and functional testing of the water intake head, docking base, and underwater docking control system (including positioning, control, and propulsion units) are completed. After successful testing, the docking base is first hoisted and fixed at the predetermined seabed target position. Subsequently, the water intake head is hoisted and sunk. At this time, the long-baseline acoustic array deployed on the docking base and the acoustic transponder and inertial measurement unit installed on the water intake head are activated to begin acquiring high-precision relative position and attitude information in real time. Based on this information and real-time water flow data, the path calculation module uses a model predictive control algorithm to continuously calculate the target movement path and performs online model parameter identification in parallel to adapt to environmental changes. The movement control module drives the water intake head to move along the path, dynamically adjusting the thruster thrust to track the path and maintain attitude stability. This process also includes a thruster fault-tolerant step to cope with unexpected failures. When the intake head contacts the base, a six-dimensional force sensor monitors the contact force and torque, and a force-position hybrid control strategy is used for active compliant adjustment until mechanical locking is completed. Finally, after confirming successful docking and locking, the hoisting device connected to the intake head and the above-water auxiliary equipment are retrieved, completing the entire underwater docking construction.

[0086] Existing construction methods may rely heavily on manual intervention by operators. Positioning information may be updated slowly and with limited accuracy; path planning may be simple or pre-set, failing to effectively compensate for real-time water flow. The control process may lack model adaptation and fault tolerance. During the contact and docking phase, it may primarily rely on passive guidance from mechanical guide mechanisms and operator experience, lacking proactive force sensing and compliant control. This method results in low automation throughout the construction process; docking success heavily depends on operator skill and site conditions, leading to low construction efficiency and a high risk of failure in complex sea conditions. Equipment damage is easily caused by improper control or docking deviations.

[0087] This invention provides a highly automated, intelligent, and robust underwater docking method for water intakes by systematically integrating a series of advanced technical modules and intelligent control strategies into a complete construction process. From preparation, deployment, docking, and retrieval, this method forms a closed-loop control system, significantly reducing reliance on manual operation. It comprehensively addresses challenges in multiple stages, including precise positioning, water flow disturbance, model uncertainty, component failure, and final smooth docking, thereby significantly improving the overall success rate, construction efficiency, and operational safety of underwater docking operations.

[0088] Although embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. It can be applied to various fields suitable for the embodiments of the present invention. Other modifications can be readily implemented by those skilled in the art. Therefore, without departing from the general concept defined by the claims and their equivalents, the embodiments of the present invention are not limited to the specific details.

Claims

1. A system for underwater docking of a water intake head, characterized in that, include: The positioning module is used to obtain the relative position information between the water intake head and the docking base in real time; The path calculation module is used to calculate the target movement path of the water intake head based on the relative position information. When calculating the target movement path, a water flow compensation amount based on real-time water flow measurement data is introduced. The path calculation module includes: a dynamic model storage unit for storing the dynamic model of the water intake head; a water flow data acquisition unit for acquiring real-time water flow measurement data; and a model prediction control unit for predicting the impact of water flow on the movement trajectory of the water intake head based on the dynamic model and the real-time water flow measurement data. With the goal of minimizing docking time and maintaining stable movement, a model predictive control algorithm is used for rolling optimization to calculate the target movement path containing sequential control commands for a future period. The movement control module is used to control the water intake head to move along the target movement path, so that the water intake head moves closer to the docking base; the force monitoring and adjustment module is used to monitor the contact force when the water intake head contacts the docking base, and adjust the movement of the water intake head according to the contact force to complete the docking.

2. The underwater docking system for water intake heads as described in claim 1, characterized in that, The positioning module includes: a long-baseline underwater acoustic positioning system arranged on the docking base, an acoustic transponder installed on the water intake head, and an inertial measurement unit carried by the water intake head itself; the positioning module is configured to measure distance through the long-baseline underwater acoustic positioning system and the acoustic transponder, and fuse the data of the inertial measurement unit to calculate the relative position information between the water intake head and the docking base.

3. The underwater docking system for water intake heads as described in claim 1, characterized in that, The model prediction control unit in the path calculation module is configured to: perform rolling optimization based on the dynamic model and the real-time water flow measurement data, using a model prediction control algorithm, with the objective function of minimizing docking time and maintaining smooth motion, predict the impact of water flow on the trajectory of the water intake head, and calculate the target movement path containing a sequence of control commands for a future period; wherein, the model prediction control algorithm executes the following steps in each control cycle: Using the current state of the water intake head as the initial condition, and combining the dynamic model and real-time water flow measurement data, the motion trajectory of the water intake head in the future multiple time steps is predicted; An optimization objective function is constructed, which simultaneously considers minimizing docking time and motion stability index, wherein motion stability is quantified by minimizing motion acceleration or energy consumption; Under constraints, the objective function is solved to generate a series of optimal control command sequences for future time steps; The first or first few instructions in the control instruction sequence are output to the motion control module, and in the next cycle, prediction and optimization are performed again based on the updated state to achieve rolling optimization control.

4. The underwater docking system for water intake heads as described in claim 1, characterized in that, The path calculation module also includes an online parameter identification unit, which is configured to compare the difference between the predicted motion state of the water intake head and the actual measured motion state in real time. Based on the difference, the recursive least squares method is used to estimate and correct the key parameters in the dynamic model online, and the corrected parameters are fed back to the model prediction and control unit for subsequent rolling optimization calculations.

5. The underwater docking system for water intake heads as described in claim 1, characterized in that, The motion control module includes: a thrust generation unit, used to generate a reference thrust force of the water intake head in the forward direction and a torque for attitude adjustment based on the target movement path; and a trajectory tracking unit, used to acquire the actual movement trajectory and attitude of the water intake head in real time, compare it with the expected value of the target movement path, and dynamically adjust the thrust distribution of each thruster based on the generated trajectory deviation and attitude deviation, so that the water intake head tracks the target path and maintains a stable attitude as it approaches the docking base.

6. The underwater docking system for water intake heads as described in claim 1, characterized in that, The force monitoring and adjustment module includes: a six-dimensional force sensor, installed on the contact surface between the water intake head and the docking base, used to acquire contact force and torque information; and a force / position hybrid control unit, used to control the water intake head to perform compliant movement in two translational degrees of freedom perpendicular to the docking direction and three rotational degrees of freedom around the axis based on the contact force and torque information, so as to offset the position and angle deviation during docking, while maintaining a constant micro-force propulsion in the docking axis direction until mechanical locking is completed.

7. The underwater docking system for water intake heads as described in claim 4, characterized in that, The parameter online identification unit is configured to: collect in real time the actual acceleration information of the water intake head measured by the inertial measurement unit, and the theoretical resultant force calculated by the thruster thrust command; construct an observation equation with the hydrodynamic coefficient in the dynamic model as the parameter to be identified, and use the difference between the theoretical resultant force and the inertial force calculated based on the actual acceleration as the observation vector; The hydrodynamic coefficients are estimated online using the recursive least squares method with a forgetting factor, and the updated parameter values ​​are fed back to the prediction model of the model prediction control unit.

8. The underwater docking system for water intake heads as described in claim 7, characterized in that, The online parameter identification unit also includes a reliability assurance subunit, configured to: perform data consistency verification between the actual acceleration information and the theoretical resultant force; if the difference in magnitude between the two exceeds a preset threshold, it is determined to be data abnormal, the parameter update of the current period is paused, and the parameter estimate of the previous period is used; after the data verification is passed, the innovation sequence of the current observation vector is calculated, and the covariance of the innovation sequence is monitored; when the covariance of the innovation sequence suddenly increases, the forgetting factor of the recursive least squares method is automatically increased to reduce the forgetting rate of old data and enhance the robustness of the algorithm to instantaneous interference.

9. The underwater docking system for water intake heads as described in claim 1, characterized in that, The motion control module also includes a fault-tolerant control unit, configured to: monitor the commanded thrust and actual thrust feedback of each thruster in real time; if the thrust deviation of a certain thruster continuously exceeds the safety threshold and its rotational speed feedback is abnormal, then the thruster is determined to have malfunctioned; based on the position and thrust capability of the remaining normally functioning thrusters, reconstruct the thrust allocation matrix; with the goal of getting as close as possible to the original resultant force and torque, use the weighted least squares method to solve the reconstructed thrust allocation matrix, set higher weights for the thrust and torque allocation priority on the core degrees of freedom, and generate fault-tolerant control commands.

10. A method for underwater connection of water intake heads, characterized in that, Includes the following steps: S1. Construction Preparation and System Deployment: The integration and testing of the water intake head, docking base, and underwater docking control system are completed on the construction mother ship; subsequently, the docking base is hoisted and fixed at the predetermined seabed target position; the underwater docking control system includes a positioning unit, a control unit, and a power unit; wherein, the positioning unit includes a long-baseline acoustic array deployed on the docking base and an acoustic transponder and inertial measurement unit (IMU) installed on the water intake head; the control unit is deployed on the construction mother ship and is communicatively connected to the water intake head via a watertight cable. S2. Relative Position Dynamic Sensing: The positioning system is activated to acquire, in real time, the relative position information between the lowered water intake head and the fixed docking base. Specifically, the method for acquiring this relative position information in real time involves: ranging via the long-baseline acoustic array and acoustic transponder; fusing the water intake head attitude and acceleration data provided by the inertial measurement unit (IMU); and using a filtering algorithm to calculate high-precision three-dimensional relative position and attitude information. S3. Flow Compensation Path Planning: Based on the relative position information and incorporating flow compensation based on real-time marine environmental monitoring data, the target movement path from the intake head to the docking base is calculated. The method for calculating the target movement path adopts the Model Predictive Control (MPC) algorithm. This algorithm is based on the dynamic model of the intake head, with the optimization objective of minimizing docking time and maintaining motion stability. It also incorporates real-time flow data as feedforward compensation and continuously calculates the optimal movement path sequence in the future time domain. Simultaneously, an online model parameter identification process is executed in parallel: the motion data of the water intake head is collected in real time, compared with the predicted values ​​of the dynamic model, the hydrodynamic parameters in the model are estimated and corrected online using the recursive least squares method, and the corrected parameters are fed back to the model predictive control (MPC) algorithm in real time to adapt to changes in the actual marine environment. S4. Adaptive Path Tracking Control: Controls the water intake head to move along the target movement path, and dynamically adjusts the thrust of the thrusters according to the actual movement state of the water intake head during the movement to resist water flow interference and stably approach the docking base; the specific process of dynamically adjusting the thruster thrust is as follows: the actual movement trajectory of the water intake head is compared with the expected value of the target movement path to generate trajectory and attitude deviation; based on this deviation, the thrust command of each thruster is calculated and adjusted in real time through a thrust distribution algorithm, so that the water intake head has the ability to resist current and stabilize direction, and accurately track the target path; The adaptive path tracking control process also includes a thruster fault tolerance step: real-time monitoring of the operating status of each thruster; if a thruster failure is detected, the thrust distribution matrix is ​​immediately recalculated based on the configuration of the remaining normal thrusters, prioritizing control of the core motion degree of freedom and maintaining stable head approximation. S5. Active, compliant, and precise docking: When the water intake head contacts the docking base, the contact force and torque are monitored in real time, and the attitude and position of the water intake head are actively adjusted based on the contact force information to compensate for the final docking deviation until mechanical locking is completed. The specific method of active adjustment based on contact force information is a force / position hybrid control strategy: the contact force and torque are obtained through a six-dimensional force sensor installed on the end face of the water intake head; the water intake head is controlled to make compliant position adjustments in the direction perpendicular to the docking axis and compliant attitude adjustments in the rotational degree of freedom to eliminate misalignment; at the same time, a constant micro-force is maintained in the docking axis direction to ensure smooth engagement of the locking mechanism. S6. Confirmation of docking status and equipment recovery: After confirming that the water intake head and docking base are successfully locked, recover the hoisting device connected to the water intake head and some of the above-water auxiliary equipment to complete the underwater docking construction.