Intelligent control system and method for sailing posture of multi-sensor yacht

Abstract

The application relates to the technical field of navigation attitude control, and discloses a multi-sensor yacht navigation attitude intelligent regulation and control system and method, which realizes high-precision real-time attitude regulation and control through multi-source sensor fusion and multi-degree-of-freedom decoupling control algorithm; an extended Kalman fusion algorithm is adopted to complete state estimation of multi-source data, effectively reduces measurement noise and environmental interference of a single sensor, and accurately outputs three-degree-of-freedom attitude parameters and disturbance torque estimation values; in combination with a time-varying nonlinear dynamic model and a multi-dimensional mapping matrix, a self-adaptive decoupling control torque is output through a self-adaptive decoupling control algorithm, coupling interference between multi-degree-of-freedom motions is effectively inhibited; meanwhile, closed-loop regulation and control is triggered based on an attitude deviation degree, boundary checking of an actuator is combined to guarantee control safety, and an online self-adaptive iteration of the model is realized through a smoothing update mechanism, so that the attitude control precision and system robustness of the yacht under a steady-state navigation working condition are significantly improved.

Description

Technical Field

[0001] This invention relates to the technical field of navigation attitude control, specifically to a multi-sensor intelligent control system and method for yacht navigation attitude. Background Technology

[0002] In the realm of luxury yachts and high-performance surface vessels, the stability of a vessel's sailing attitude directly determines its passenger comfort, navigation safety, and fuel economy. Yachts, as typical underactuated surface vessels, typically rely on a limited set of actuators, including the rudder system, anti-roll fins, and main propeller, to simultaneously control attitude in three degrees of freedom: roll, pitch, and yaw. This system inherently possesses strongly coupled dynamic characteristics. With real-time changes in speed, draft, and the marine environment, key dynamic parameters such as hydrodynamic damping and added mass exhibit strong time-varying nonlinear characteristics, further exacerbating the difficulty of attitude control and placing higher demands on the environmental adaptability and control precision of yacht attitude control systems. Currently, most mainstream yacht sailing attitude control systems are designed based on linear, steady hydrodynamic models, employing traditional feedback control strategies and utilizing a single inertial measurement unit for attitude sensing. Existing yacht attitude control systems have several shortcomings: First, the control model is idealized and does not fully consider the time-varying nonlinear characteristics of hydrodynamic parameters with speed and sea state. The underactuated system has poor multi-degree-of-freedom coupling suppression effect, and the control accuracy is significantly reduced under high sea state conditions. Second, the sensing system is singular, and the multi-source heterogeneous sensors lack a unified spatiotemporal reference. The state estimation is easily affected by sea surface disturbances, making it difficult to accurately perceive environmental disturbances and the vessel's own attitude. Third, it is mainly based on feedback control, and the control response lags behind wave and wind disturbances. Especially in irregular wave and sudden change scenarios, the compensation delay leads to large attitude fluctuations. Fourth, it has poor adaptability across the entire speed range. The control performance fluctuates greatly between low and high speed conditions, the model has weak adaptive ability, and parameters need to be readjusted for different ship types and load conditions, resulting in high adaptation costs and long cycles.

[0003] Chinese invention patent application CN121879361A, published on April 17, 2026, discloses a multi-source information fusion intelligent navigation control system suitable for unmanned self-propelled vessels. It includes a data acquisition module for collecting data within a radius around the self-propelled vessel; a data processing module for fusing the acquired data to create a real-time map with the self-propelled vessel as the origin; and an attitude adjustment unit for adjusting the vessel's navigation attitude based on the real-time map data. This system achieves cloud-edge-device computation of the data, significantly reducing the computational demands on the vessel itself for high-precision attitude control. However, this technical solution cannot solve the coupling problem of multi-degree-of-freedom attitude control (roll, pitch, yaw), thus reducing the safety of the vessel's attitude control. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] To address the shortcomings of existing yacht attitude control systems: First, the control model is idealized and does not fully consider the time-varying nonlinear characteristics of hydrodynamic parameters with speed and sea state. This results in poor multi-degree-of-freedom coupling suppression in underactuated systems and significant attenuation of control accuracy under high sea states. Second, the sensing system is limited; the lack of a unified spatiotemporal reference for multi-source heterogeneous sensors makes state estimation susceptible to sea surface disturbances, hindering accurate perception of environmental disturbances and the yacht's attitude. Third, the reliance on feedback control leads to a lag in control response compared to wave and wind disturbances. This compensation delay, especially in irregular wave and abrupt change scenarios, negatively impacts attitude control. The problems include: large fluctuations in attitude; poor adaptability across the entire speed range; large fluctuations in control performance under low and high speed conditions; weak model adaptability; and the need to readjust parameters for different ship types and load conditions, resulting in high adaptation costs and long cycles. To address these issues, the above measures are implemented by using time-varying nonlinear hydrodynamic modeling and multi-degree-of-freedom decoupled control, combined with multi-source sensor fusion state estimation, to effectively improve the multi-degree-of-freedom attitude control accuracy of underactuated yachts. Furthermore, a multi-dimensional mapping matrix is ​​constructed to achieve closed-loop self-learning, and a working condition feedforward compensation mechanism is introduced to significantly enhance the stability of yacht navigation attitude and the adaptability of working conditions across the entire speed range and sea conditions.

[0006] (II) Technical Solution

[0007] This invention is achieved through the following technical solution: a multi-sensor intelligent control method for yacht navigation attitude, the method comprising the following steps:

[0008] S1. Under gradient sea state and speed conditions, collect time-varying hydrodynamic coefficients and attitude response data of the yacht under different attitude motions, and establish a hydrodynamic coefficient-attitude disturbance reference database; combine the multi-degree-of-freedom coupling characteristics of the underactuated system to construct a time-varying nonlinear hydrodynamic model, and complete the spatiotemporal calibration and synchronous calibration of multi-source heterogeneous sensors.

[0009] S2. Under typical steady-state navigation conditions, collect real-time data from multiple sensors, attitude motion parameters, and actuator control data. Combined with the time-varying nonlinear hydrodynamic model, identify the quantitative correlation between sensing features, attitude state, and control torque, and construct a multi-dimensional mapping matrix of sensing, attitude, and control.

[0010] S3. During continuous navigation, multi-source sensors collect yacht motion data and environmental disturbance data online in real time, perform multi-source fusion state estimation, obtain the current accurate attitude parameters and disturbance characteristic parameters, call the time-varying nonlinear hydrodynamic model and the sensing-attitude-control multidimensional mapping matrix, and calculate the multi-degree-of-freedom decoupled control torque.

[0011] S4. Calculate the relative attitude deviation of the current attitude parameters relative to the preset attitude threshold; when the relative attitude deviation exceeds the dynamic adaptive deviation threshold, call the sensor-attitude-control multidimensional mapping matrix, optimize and allocate the control torque compensation of each actuator based on actuator constraints, send it to the navigation actuator for closed-loop control, and synchronously update the sensor-attitude-control multidimensional mapping matrix and hydrodynamic model parameters; when the relative attitude deviation does not exceed the dynamic adaptive deviation threshold, smoothly update the current detection and control data to the sensor-attitude-control multidimensional mapping matrix.

[0012] S5. When an intervention action is detected in the yacht's navigation condition, the sensor-attitude-control multidimensional mapping matrix and the time-varying nonlinear hydrodynamic model are invoked to calculate the expected attitude fluctuation caused by the intervention action. The expected attitude fluctuation is then converted into a feedforward compensation signal and superimposed onto the attitude control system.

[0013] Preferably, the specific steps for establishing the hydrodynamic coefficient-attitude disturbance reference database in step S1 include:

[0014] S101 covers the entire speed range of zero-speed anchoring, low-speed cruising, medium-speed navigation, and high-speed navigation, as well as typical sea states from level 1 to level 5, and prepares multiple sets of gradient working condition test samples.

[0015] S102. Perform dynamic response tests on all samples, record the amplitude and frequency of roll, pitch, and yaw angles under each working condition, as well as the response characteristics of the corresponding control torque;

[0016] S103. The operating parameters, hydrodynamic coefficients, attitude response parameters, and control torque parameters are stored in a structured manner to form a hydrodynamic coefficient-attitude disturbance reference database.

[0017] Preferably, the specific steps for constructing the time-varying nonlinear hydrodynamic model and multi-sensor calibration in step S1 include:

[0018] S11. Set up gradient speed and gradient sea state test groups, each group corresponding to constant speed and sea state level, and conduct towing pool test and sea sailing test on the target yacht. Simultaneously collect hydrodynamic damping coefficient, added mass coefficient and three-degree-of-freedom attitude response data of roll, pitch and yaw under different speeds, different drafts and different wave levels.

[0019] S12. Based on experimental data, identify the time-varying characteristics of hydrodynamic coefficients, fit the nonlinear variation law of hydrodynamic coefficients with speed, draft and sea state, and combine the actuator configuration constraints of underactuated system to establish a three-degree-of-freedom coupled dynamic model of roll-pitch-bow.

[0020] S13. Perform time synchronization calibration and spatial coordinate registration on multi-source heterogeneous sensors such as inertial measurement unit, satellite positioning unit, anemometer, wave height meter, rudder angle sensor, and thruster speed sensor to establish a sensor data acquisition system under a unified spatiotemporal reference.

[0021] Preferably, when constructing the sensing-attitude-control multidimensional mapping matrix in step S2, an underactuated actuator constraint optimization step is also incorporated:

[0022] Based on the maximum output torque, response bandwidth, and operating degrees of freedom of each actuator, the performance constraint boundary of the actuator is established. Combining the attitude control priority of each degree of freedom, the torque distribution weight under multi-objective optimization is solved. The control quantity benchmark value in the sensing-attitude-control multidimensional mapping matrix is ​​constrained and corrected, and then the attitude control effect is matched accordingly.

[0023] Preferably, the specific steps for calculating the multi-degree-of-freedom decoupling control torque in step S3 include:

[0024] S31. During the current control cycle, collect data on the yacht's angular acceleration, linear acceleration, position and heading, wind speed and direction, wave height and direction, rudder angle, and thrust through multi-source sensors.

[0025] S32. Perform spatiotemporal registration, noise filtering, and feature extraction on multi-source sensor data, and perform state estimation using the extended Kalman fusion algorithm to obtain real-time attitude angles, angular velocities, and disturbance torque estimates for the three degrees of freedom of roll, pitch, and yaw.

[0026] S33. Input the real-time attitude parameters and disturbance characteristic parameters into the time-varying nonlinear hydrodynamic model, combine the sensing-attitude-control multidimensional mapping matrix, and output the three-degree-of-freedom decoupled control torque through the adaptive decoupling control algorithm.

[0027] Preferably, the specific steps of S4 include:

[0028] S41. Using the preset attitude stability threshold as the reference value, calculate the ratio of the difference between the current real-time attitude parameters and the reference value to the reference value to obtain the attitude relative deviation.

[0029] S42. Determine whether the relative deviation of the posture is greater than the dynamic adaptive deviation threshold; if yes, then execute S43; if no, then execute S44.

[0030] S43. Based on the control weights of each actuator in the sensor-attitude-control multidimensional mapping matrix, prioritize adjusting the output of the high-weight core actuator, calculate the control torque fine-tuning compensation amount of each actuator based on the attitude relative deviation value and weight ratio, send it to the corresponding navigation actuator for closed-loop control, and simultaneously update the sensor-attitude-control multidimensional mapping matrix and hydrodynamic model parameters.

[0031] S44. Keeping the current control parameters unchanged, use a preset exponential smoothing formula to smoothly update the current sensing data, attitude parameters and control data to the sensing-attitude-control multidimensional mapping matrix.

[0032] Preferably, step S43 further includes an actuator boundary verification step before issuing the control torque fine-tuning compensation amount:

[0033] S431. Read the target actuator output value after superimposed control torque fine-tuning compensation;

[0034] S432. Determine whether the output value of the target actuator falls within the preset allowable output range of the actuator; if it exceeds the upper limit of the range, then forcibly cut off the control torque fine-tuning compensation amount to the output value corresponding to the upper limit of the range; if it is lower than the lower limit of the range, then forcibly cut off the control torque fine-tuning compensation amount to the output value corresponding to the lower limit of the range.

[0035] Preferably, the operational intervention action in S5 includes a sudden change in speed, and the specific steps include:

[0036] S51. Monitor the main engine throttle adjustment signal and speed setting command in real time through the ship's power control system;

[0037] S52. Extract the hydrodynamic characteristic coefficients corresponding to the target speed from the sensing-attitude-control multidimensional mapping matrix, and calculate the expected pitch angle change and roll damping change caused by the speed change as the expected attitude fluctuation.

[0038] S53. Generate a corresponding feedforward compensation signal based on the expected attitude fluctuation amount, and adjust the angle of attack of the anti-roll fin and the rudder angle in advance to counteract the attitude fluctuation caused by sudden changes in speed.

[0039] Preferably, the condition intervention action in S5 also includes a steering maneuver action;

[0040] When a yacht steering command is detected, the expected heave moment and yaw coupling moment caused by the steering centrifugal force are calculated based on the steering angular velocity and the target rudder angle. The expected moment is then substituted into the sensor-attitude-control multidimensional mapping matrix to obtain the expected attitude fluctuation, and the corresponding anti-roll fin and thruster feedforward compensation signals are generated and sent to the attitude control system.

[0041] A multi-sensor yacht navigation attitude intelligent control system is used to implement the multi-sensor yacht navigation attitude intelligent control method; the system includes:

[0042] The yacht multi-source data acquisition and preprocessing module is used to acquire hydrodynamic data, multi-source sensor data, attitude response data and actuator control data during pool tests and actual ship navigation, and to perform feature extraction and noise filtering, supporting the construction of hydrodynamic coefficient-attitude disturbance benchmark database and time-varying nonlinear hydrodynamic model;

[0043] The attitude dynamics benchmark model and multidimensional mapping matrix management module is used to build, store and dynamically update the hydrodynamic coefficient-attitude disturbance benchmark database, time-varying nonlinear hydrodynamic model and sensing-attitude-control multidimensional mapping matrix;

[0044] The multi-degree-of-freedom attitude calculation and closed-loop control module is used to calculate the yacht's multi-degree-of-freedom attitude parameters and relative attitude deviation in real time, and to perform control torque optimization and closed-loop control based on the sensor-attitude-control multi-dimensional mapping matrix.

[0045] The navigation condition feedforward compensation and execution drive module is used to monitor the yacht's operating condition intervention actions, generate feedforward compensation signals, and work with the navigation actuators to complete the final attitude control output.

[0046] (III) Beneficial Effects

[0047] This invention provides a multi-sensor intelligent control system and method for yacht navigation attitude. It has the following beneficial effects:

[0048] I. By constructing a hydrodynamic coefficient-attitude disturbance benchmark database and a time-varying nonlinear hydrodynamic model, a dynamic foundation for precise control was laid. Through gradient operating condition tests, the time-varying nonlinear laws of the hydrodynamic coefficients were identified, and a three-degree-of-freedom coupled dynamic model of roll, pitch, and yaw was established. This model overcomes the adaptation limitations of traditional linear steady models and can accurately characterize the yacht's motion characteristics under different speeds and sea states. Simultaneously, the spatiotemporal calibration and synchronous calibration of multi-source heterogeneous sensors were completed. A multi-dimensional mapping matrix of sensing, attitude, and control was constructed and actuator constraint optimization was integrated. This achieved a quantitative correlation between sensing characteristics, attitude state, and control torque, providing a reliable mapping basis and constraint boundary for subsequent online yacht navigation attitude control.

[0049] Second, high-precision real-time attitude control was achieved through multi-source sensor fusion and multi-degree-of-freedom decoupled control algorithms. An extended Kalman fusion algorithm was used to perform state estimation of multi-source data, effectively reducing measurement noise and environmental interference from single sensors and accurately outputting estimated values ​​of three-degree-of-freedom attitude parameters and disturbance torques. Combining a time-varying nonlinear dynamic model and a multi-dimensional mapping matrix, an adaptive decoupled control algorithm was used to output decoupled control torques, effectively suppressing coupling interference between multi-degree-of-freedom motions. Simultaneously, closed-loop control was triggered based on attitude deviation, and actuator boundary verification ensured control safety. A smooth update mechanism enabled online adaptive iteration of the model, significantly improving the attitude control accuracy and system robustness under steady-state navigation conditions of the yacht.

[0050] Third, the attitude stability under sudden changes in operating conditions is greatly improved through the working condition feedforward compensation mechanism. For typical working condition intervention actions such as sudden changes in speed and steering maneuvers during navigation, the expected attitude fluctuation caused by changes in operating conditions is calculated in advance by using a multi-dimensional mapping matrix and a time-varying hydrodynamic model. Within the same cycle of the working condition execution, the expected fluctuation is converted into a feedforward signal to actively compensate the control system, thereby mitigating the attitude impact caused by changes in operating conditions in advance. This solves the problem of lag in traditional feedback control of yacht navigation attitude and ensures the stability of yacht navigation attitude control throughout the entire navigation process. Attached Figure Description

[0051] Figure 1 A flowchart of the multi-sensor intelligent control method for yacht navigation attitude provided by the present invention;

[0052] Figure 2 This is a schematic diagram of the modules of the multi-sensor yacht navigation attitude intelligent control system provided by the present invention. Detailed Implementation

[0053] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] An embodiment of the multi-sensor yacht navigation attitude intelligent control system and method is as follows:

[0055] Example 1:

[0056] Please see Figure 1 A multi-sensor intelligent control method for yacht navigation attitude, comprising the following steps:

[0057] This embodiment is based on a 12-meter luxury motor yacht, with a target navigation attitude control accuracy of roll angle ≤ 2° and pitch angle ≤ 1.5°. The core hardware configuration is as follows:

[0058] Navigation actuators include the main thruster, hydraulic steering gear, two pairs of retractable anti-roll fins, and bow thrusters, which respectively provide thrust output, heading control, roll reduction, and low-speed attitude adjustment; the anti-roll fins have a maximum output torque of 2500 Nm. The maximum steering torque of the servo motor is 1800. The bow thruster has a maximum thrust of 12 kN.

[0059] Multi-source sensor data acquisition equipment:

[0060] Inertial Measurement Unit (IMU): Employs a high-precision fiber optic gyroscope, installed at the center of gravity of the yacht, with a sampling frequency of 100Hz, outputting three-axis attitude angles, angular velocity, and linear acceleration, with an angle measurement accuracy of ±0.05°;

[0061] Satellite positioning unit: Beidou + GPS dual-mode positioning, outputs position, speed and heading information, sampling frequency 10Hz;

[0062] Environmental sensing sensors: wind speed and direction meter, ultrasonic wave height meter, which collects wind speed, wind direction, wave height and wave direction data in real time, with a sampling frequency of 5Hz;

[0063] Actuator feedback sensors: rudder angle sensor, fin angle sensor, thruster speed sensor, which provide feedback on the actual output status of the actuator, with a sampling frequency of 20Hz;

[0064] Data processing and control unit: Shipborne industrial control computer, equipped with a real-time operating system, with a control cycle of 50ms, responsible for data fusion, algorithm calculation and control command output.

[0065] S1. Under gradient sea state and speed conditions, collect time-varying hydrodynamic coefficients and attitude response data of the yacht under different attitude motions, and establish a hydrodynamic coefficient-attitude disturbance reference database; combine the multi-degree-of-freedom coupling characteristics of the underactuated system to construct a time-varying nonlinear hydrodynamic model, and complete the spatiotemporal calibration and synchronous calibration of multi-source heterogeneous sensors.

[0066] The specific steps for establishing the hydrodynamic coefficient-attitude disturbance reference database in S1 include:

[0067] S101 covers the entire speed range of zero-speed anchoring, low-speed cruising (8kn), medium-speed navigation (18kn), and high-speed navigation (28kn), as well as typical sea states from level 1 to level 5, and prepares multiple sets of gradient operating condition test samples.

[0068] Specifically, this embodiment sets up a basic operating condition combination of 4 speed levels × 5 sea states, totaling 20 sets of gradient operating condition samples; each operating condition is repeated 3 times in parallel to reduce random errors, where the sea state levels correspond to the following significant wave heights: 0.1m for level 1, 0.3m for level 2, 0.6m for level 3, 1.0m for level 4, and 1.5m for level 5; the test simultaneously covers three draft states: full load (draft 1.8m), half load (draft 1.5m), and unloaded (draft 1.2m), ultimately forming a total of 60 sets of test samples, covering the typical operating condition range of actual yacht navigation.

[0069] S102. Perform dynamic response tests on all samples, and record the amplitude and frequency of roll, pitch, and yaw angles under each working condition, as well as the response characteristics of the corresponding control torque.

[0070] Specifically, the experiment was conducted in a towing tank, where a wave generator simulated the target sea state, and a trailer propelled the yacht to a set speed. The attitude response characteristics were tested using the uncontrolled free decay method: an external excitation device was used to induce an initial attitude deflection in the yacht, which was then released. Attitude decay curve data were continuously collected using an IMU and motion capture system. Under sea state 3, 18 knots cruising conditions, and full load draft, the uncontrolled roll angle amplitude was measured to be approximately 6.5° with a natural roll frequency of 0.42Hz, the pitch angle amplitude to be approximately 3.2° with a natural pitch frequency of 0.58Hz, and the bow angle amplitude to be approximately 2.8°. Simultaneously, the output roll torque per unit angle of attack of the anti-roll fins was measured to be 185 mm / s under these conditions. / °.

[0071] S103. The operating parameters, hydrodynamic coefficients, attitude response parameters, and control torque parameters are stored in a structured manner to form a hydrodynamic coefficient-attitude disturbance reference database.

[0072] Specifically, each record in the database contains structured fields: speed, sea state, draft, roll damping coefficient, pitch damping coefficient, yaw damping coefficient, roll added mass, pitch added mass, roll angle amplitude, pitch angle amplitude, yaw angle amplitude, and unit control torque response; a total of 60 sets of baseline operating condition data are stored, and the data sampling density and accuracy meet the input requirements for subsequent dynamic model identification.

[0073] The specific steps for constructing a time-varying nonlinear hydrodynamic model and performing multi-sensor calibration in S1 include:

[0074] S11. Set up gradient speed and gradient sea state test groups, each corresponding to constant speed and sea state level, and conduct towing tank tests and live sea sailing tests on the target yacht. Simultaneously collect hydrodynamic damping coefficient, added mass coefficient and three-degree-of-freedom attitude response data of roll, pitch, and bow under different speeds, different drafts and different wave levels.

[0075] Specifically, five speed gradients (8kN, 12kN, 18kN, 22kN, and 28kN) were set, corresponding to three sea state gradients (2, 3, and 4), for towing tank tests. Simultaneously, live-ship navigation tests in sea states 1-5 were conducted in a real-world marine environment to verify the validity of the tank test results. During the tests, hydrodynamic loads on the hull were measured using a six-component balance on the hull, and three-degree-of-freedom attitude response data were acquired using a high-precision IMU. All acquired data were aligned using a unified timestamp, with a time synchronization accuracy better than 5ms.

[0076] S12. Based on experimental data, identify the time-varying characteristics of hydrodynamic coefficients, fit the nonlinear variation law of hydrodynamic coefficients with speed, draft and sea state, and combine the actuator configuration constraints of underactuated system to establish a three-degree-of-freedom coupled dynamic model of roll, pitch and yaw.

[0077] Specifically, by using the least squares system identification method, the experimental data under gradient conditions are fitted to obtain the time-varying nonlinear expressions for the hydrodynamic coefficients: Roll damping coefficient: ,in This is the roll damping coefficient, in units of... ; The speed of the yacht, in units of (Festival).

[0078] Pitch damping coefficient: ,in This is the pitch damping coefficient, in units of... ; Significant wave height, in units of ; It is a natural constant with a value of approximately 2.71828.

[0079] Simultaneously, a three-degree-of-freedom coupling coefficient matrix is ​​introduced to characterize the mutual influence between roll, pitch, and yaw motions. The resulting time-varying nonlinear coupled dynamic model achieves a fitting accuracy of ≥92% for the yacht's attitude response, accurately describing the yacht's motion characteristics under different operating conditions.

[0080] S13. Perform time synchronization calibration and spatial coordinate registration on multi-source heterogeneous sensors such as inertial measurement unit, satellite positioning unit, anemometer, wave height meter, rudder angle sensor, and thruster speed sensor to establish a sensor data acquisition system under a unified spatiotemporal reference.

[0081] Specifically, using the IMU's 100Hz clock as the global time reference, all sensors are time-synchronized and calibrated using the PTP precise time protocol. After synchronization, the timestamp error of each sensor's data is ≤1ms. For spatial coordinate registration, a right-handed coordinate system is established with the ship's center of gravity as the origin. Through the calibration of each sensor's installation position and coordinate transformation matrix, the measurement data of all sensors are uniformly transformed to the ship's center of gravity coordinate system. The coordinate transformation error is ≤0.02m, ensuring spatial consistency of multi-source data.

[0082] S2. Under typical steady-state navigation conditions, collect real-time data from multiple sensors, attitude motion parameters and actuator control data, and combine them with a time-varying nonlinear hydrodynamic model to identify the quantitative correlation between sensor features, attitude state and control torque, and construct a multi-dimensional mapping matrix of sensing, attitude and control.

[0083] Specifically, the typical steady-state operating conditions were selected as 18kn cruise, sea state 3, and full load draft. Under these conditions, the average angle of attack of the anti-roll fin was 8°, the rudder angle was 0°, and the main engine speed was 2400 r / min. The system continuously collected 10 minutes of steady-state operating data. Combined with a time-varying nonlinear hydrodynamic model, the control weights of each actuator for different degrees of freedom were identified through correlation analysis: the roll control weight was 0.85 for the anti-roll fin and 0.15 for the rudder; the pitch control weight was 0.7 for the anti-roll fin and 0.3 for the main thruster thrust adjustment; and the bow roll control weight was 0.9 for the rudder and 0.1 for the bow thruster.

[0084] When constructing the sensing-attitude-control multidimensional mapping matrix in S2, the underactuated actuator constraint optimization step is also incorporated:

[0085] Based on the maximum output torque, response bandwidth, and operating degrees of freedom of each actuator, the performance constraint boundary of the actuator is established. Combining the attitude control priority of each degree of freedom, the torque distribution weight under multi-objective optimization is solved. The control quantity benchmark value in the sensing-attitude-control multi-dimensional mapping matrix is ​​constrained and corrected, and then the attitude control effect is matched accordingly.

[0086] Specifically, the actuator constraint boundaries are set as follows: the angle of attack range of the anti-roll fin is [-15°, +15°], and the maximum output torque is 2500. The response bandwidth is 2Hz; the servo angle range is [-35°, +35°], the maximum output torque is 1800 N·m, and the response bandwidth is 1.5Hz; the bow thruster has a maximum thrust of 12 kN and a response bandwidth of 1Hz. Attitude control priority is set as: roll > pitch > yaw. With minimizing attitude deviation and control energy consumption as dual objectives, a quadratic programming algorithm is used to solve for the optimal torque allocation weights. After correction, all control outputs in the mapping matrix satisfy the actuator physical boundary constraints, and the roll control torque allocation priority is improved by 12%.

[0087] By constructing a hydrodynamic coefficient-attitude disturbance benchmark database and a time-varying nonlinear hydrodynamic model, a dynamic foundation for precise control was laid. Through gradient condition tests, the time-varying nonlinear laws of the hydrodynamic coefficients were identified, and a three-degree-of-freedom coupled dynamic model of roll, pitch, and yaw was established. This model overcomes the adaptation limitations of traditional linear steady models and can accurately characterize the yacht's motion characteristics under different speeds and sea states. At the same time, the spatiotemporal calibration and synchronous calibration of multi-source heterogeneous sensors were completed, a multi-dimensional mapping matrix of sensing, attitude, and control was constructed, and actuator constraint optimization was integrated. This achieved a quantitative correlation between sensing characteristics, attitude state, and control torque, providing a reliable mapping basis and constraint boundary for subsequent online yacht navigation attitude control.

[0088] S3. During continuous navigation, multi-source sensors collect yacht motion data and environmental disturbance data online in real time, perform multi-source fusion state estimation, obtain the current accurate attitude parameters and disturbance characteristic parameters, call the time-varying nonlinear hydrodynamic model and the sensing-attitude-control multidimensional mapping matrix, and calculate the multi-degree-of-freedom decoupled control torque.

[0089] The specific steps for calculating the multi-degree-of-freedom decoupling control torque in S3 include:

[0090] S31. During the current control cycle, collect data on the yacht's angular acceleration, linear acceleration, position and heading, wind speed and direction, wave height and direction, rudder angle, and thrust through multi-source sensors.

[0091] Specifically, the control cycle is 50ms, and the following data are collected synchronously within each cycle: raw data of three-axis angular acceleration, linear acceleration, and attitude angle output by the IMU; airspeed, heading, and position data output by satellite positioning; wind speed and direction data at a height of 10m output by the anemometer; effective wave height and wave direction data output by the wave height meter; actual rudder angle data output by the rudder angle sensor; and main engine speed and thrust data output by the thruster speed sensor.

[0092] S32. Perform spatiotemporal registration, noise filtering, and feature extraction on multi-source sensor data. Then, use the extended Kalman fusion algorithm to perform state estimation and obtain real-time attitude angles, angular velocities, and disturbance torque estimates for the three degrees of freedom of roll, pitch, and yaw.

[0093] Specifically, the data from each sensor is first resampled via time interpolation and spatial coordinate transformation to achieve spatiotemporal registration. A first-order low-pass filter is used to remove high-frequency measurement noise from the sensors. Then, using a time-varying nonlinear hydrodynamic model as the system state equation and multi-source sensor data as the observation equation, an extended Kalman filter is constructed for state estimation. The state vector includes three degrees of freedom: attitude angle, angular velocity, and disturbance torque. For a given control cycle, the following values ​​are obtained through fusion estimation: roll angle 3.2°, roll velocity 1.8° / s, pitch angle 1.7°, pitch velocity 0.9° / s, and yaw angle 1.2°. Simultaneously, the roll component of the wave disturbance torque is estimated to be approximately 1200°. Pitch component approximately 580 .

[0094] S33. Input the real-time attitude parameters and disturbance characteristic parameters into the time-varying nonlinear hydrodynamic model, combine the sensing-attitude-control multidimensional mapping matrix, and output the three-degree-of-freedom decoupled control torque through the adaptive decoupling control algorithm.

[0095] Specifically, the attitude angle, angular velocity, and disturbance torque are input into the adaptive decoupling controller. Feedforward decoupling compensation is performed based on the coupled dynamics model to counteract coupling interference between degrees of freedom. Combined with the control weight allocation of the multi-dimensional mapping matrix, the final output is: roll control torque 1150. Pitch control torque 420 Bow roll control torque 180 Verification showed that the coupling effect between the three degrees of freedom was reduced by more than 85% after decoupling, avoiding the problem of "controlling roll to introduce pitch deviation".

[0096] S4. Calculate the relative attitude deviation of the current attitude parameters relative to the preset attitude threshold; when the relative attitude deviation exceeds the dynamic adaptive deviation threshold, call the sensor-attitude-control multidimensional mapping matrix, optimize and allocate the control torque compensation of each actuator based on actuator constraints, send it to the navigation actuator for closed-loop control, and synchronously update the sensor-attitude-control multidimensional mapping matrix and hydrodynamic model parameters; when the relative attitude deviation does not exceed the dynamic adaptive deviation threshold, smoothly update the current detection and control data to the sensor-attitude-control multidimensional mapping matrix.

[0097] The specific steps of S4 include:

[0098] S41. Using a preset attitude stability threshold as a reference value, calculate the ratio of the difference between the current real-time attitude parameters and the reference value to the reference value, thus obtaining the attitude relative deviation. Specifically, the formula for calculating the attitude relative deviation is: ,in This represents the relative deviation of the posture. This represents the current real-time attitude angle; This is the preset attitude stability threshold.

[0099] In this embodiment, the preset roll stability threshold is 2° and the pitch stability threshold is 1.5°. The current real-time roll angle is 3.2°. Substituting into the formula, the relative deviation of the roll attitude is calculated to be |3.2-2| / 2×100%=60%; the pitch angle is 1.7°, and the relative deviation of the pitch attitude is |1.7-1.5| / 1.5×100%≈13.3%.

[0100] S42. Determine whether the relative deviation of the attitude is greater than the dynamic adaptive deviation threshold; if yes, execute S43; if no, execute S44.

[0101] Specifically, the system presets a dynamic adaptive deviation threshold of 15%, which is calculated from attitude fluctuation data over 30 consecutive days under normal navigation conditions using the 3σ criterion. If the relative roll deviation is 60% or greater than 15%, the attitude control procedure is triggered; if the relative pitch deviation is 13.3% or less than 15%, independent control is not triggered, and compensation is performed synchronously with the roll control.

[0102] S43. Based on the control weights of each actuator in the sensor-attitude-control multidimensional mapping matrix, prioritize adjusting the output of the high-weight core actuator. Calculate the control torque fine-tuning compensation amount of each actuator based on the relative attitude deviation value and weight ratio, and send it to the corresponding navigation actuator for closed-loop control. Simultaneously update the sensor-attitude-control multidimensional mapping matrix and hydrodynamic model parameters.

[0103] Specifically, based on the parameter weights of the mapping matrix, the output of the anti-roll fin with the highest weight is adjusted first; based on a roll deviation of 60% and an anti-roll fin weight of 0.85, it is calculated that the angle of attack of the anti-roll fin needs to be increased by 5°, corresponding to an additional roll torque of approximately 925. Simultaneously fine-tune the rudder angle by 1.2°, supplementing the roll torque by approximately 225. Total compensation torque 1150 Control commands are sent to the anti-roll fin hydraulic servo system and the steering gear system, with an execution cycle of 50ms. After the adjustment is completed, the current operating condition data, control variables and attitude response data are synchronously updated to the multi-dimensional mapping matrix and hydrodynamic model parameter library.

[0104] S431, Read the target actuator output value after superimposed control torque fine-tuning compensation.

[0105] Specifically, with the anti-roll fin angle of attack set to 8° before adjustment, and a fine-tuning amount of +5° added, the target anti-roll fin angle of attack is 13°; with the rudder angle set to 0° before adjustment, and a fine-tuning amount of +1.2° added, the target rudder angle is 1.2°.

[0106] S432. Determine whether the target actuator output value falls within the preset actuator allowable output range; if it exceeds the upper limit of the range, force the control torque fine-tuning compensation amount to be truncated to the output value corresponding to the upper limit of the range; if it is lower than the lower limit of the range, force the control torque fine-tuning compensation amount to be truncated to the output value corresponding to the lower limit of the range.

[0107] Specifically, the preset allowable angle of attack range for the anti-roll fin is [-15°, +15°], with a target angle of attack of 13° falling within this range; the allowable rudder angle range is [-35°, +35°], with a target rudder angle of 1.2° falling within this range; once the verification is successful, control commands are issued normally. If the calculated target angle of attack is 16°, it is forcibly truncated to 15°, and a saturation mark is recorded to prevent actuator overload damage.

[0108] S44. Keep the current control parameters unchanged, and use the preset exponential smoothing formula to smoothly update the current sensor data, attitude parameters and control data to the sensor-attitude-control multidimensional mapping matrix.

[0109] Specifically, the parameter is updated using an exponential smoothing formula: ,in These are the updated mapping matrix parameter values; These are the parameter values ​​measured in the current period; To update the historical parameter values ​​in the mapping matrix before updating; This is a smoothing coefficient, with a value range of 0 < <1, This embodiment takes =0.02.

[0110] If the relative deviation of the roll attitude in a certain period is 8%, which does not exceed the 15% threshold, it is judged as normal navigation fluctuation. The system uses the above formula to slowly update the currently detected attitude parameters, disturbance characteristics and control data to the multi-dimensional mapping matrix. This achieves adaptive iteration of the model, which ensures model stability and gradually adapts to the drift of the operating conditions.

[0111] High-precision real-time attitude control was achieved through multi-source sensor fusion and multi-degree-of-freedom decoupled control algorithms. An extended Kalman fusion algorithm was used to perform state estimation of multi-source data, effectively reducing measurement noise and environmental interference from single sensors and accurately outputting estimated values ​​of three-degree-of-freedom attitude parameters and disturbance torques. Combining a time-varying nonlinear dynamic model and a multi-dimensional mapping matrix, an adaptive decoupling control algorithm was used to output decoupling control torque, effectively suppressing coupling interference between multi-degree-of-freedom motions. Simultaneously, closed-loop control was triggered based on attitude deviation, and actuator boundary checks ensured control safety. A smooth update mechanism enabled online adaptive iteration of the model, significantly improving the attitude control accuracy and system robustness under steady-state navigation conditions for the yacht.

[0112] S5. When an intervention action is detected in the yacht's navigation condition, the sensor-attitude-control multidimensional mapping matrix and the time-varying nonlinear hydrodynamic model are invoked to calculate the expected attitude fluctuation caused by the intervention action. The expected attitude fluctuation is then converted into a feedforward compensation signal and superimposed onto the attitude control system.

[0113] S51. The main engine throttle adjustment signal and speed setting command are monitored in real time through the ship's power control system.

[0114] Specifically, it communicates with the ship's main engine control system via the CAN bus to read the main engine throttle opening, target speed, and set speed command in real time; when the change in the set speed value is detected to exceed 2kn, it is determined to be an intervention action for sudden speed change.

[0115] S52. Extract the hydrodynamic characteristic coefficients corresponding to the target speed from the sensing-attitude-control multidimensional mapping matrix, and calculate the expected pitch angle change and roll damping change caused by the speed change as the expected attitude fluctuation.

[0116] Specifically, when the system receives an acceleration command to increase the speed from 18 knots to 24 knots, it extracts the hydrodynamic coefficients corresponding to 24 knots from the mapping matrix: the pitch damping coefficient increases by 22% and the roll damping coefficient increases by 18%. At the same time, the increase in speed leads to an increase in the pitch head pitch moment. It is calculated that the expected pitch angle will increase by 1.2°, the roll attenuation capability will be improved, but the initial impact will cause the roll angle to increase by 0.5° instantaneously. The total expected attitude fluctuation is pitch +1.2° and roll +0.5°.

[0117] S53. Generate corresponding feedforward compensation signals based on the expected attitude fluctuations, and adjust the angle of attack of the anti-roll fins and the rudder angle in advance to counteract the attitude fluctuations caused by sudden changes in speed.

[0118] Specifically, based on the expected pitch fluctuation, a +2.5° feedforward compensation for the anti-roll fin angle of attack is generated to provide pitch recovery torque to counteract the pitch; based on the expected roll fluctuation, a +0.5° feedforward compensation for the rudder angle is generated; the feedforward signal is issued synchronously with the main engine acceleration command, actively compensating for attitude fluctuations while increasing speed. Tests have shown that this feedforward mechanism can reduce attitude fluctuations during sudden speed changes by more than 70%.

[0119] The operating condition intervention actions in S5 also include steering maneuvers;

[0120] When a yacht steering command is detected, the expected heave moment and yaw coupling moment caused by the steering centrifugal force are calculated based on the steering angular velocity and the target rudder angle. The expected moments are then substituted into the sensor-attitude-control multidimensional mapping matrix to obtain the expected attitude fluctuation, and the corresponding anti-roll fin and thruster feedforward compensation signals are generated and sent to the attitude control system.

[0121] Specifically, when the system receives a full left turn command (target rudder angle -35°, turn angular velocity 0.8° / s), it calculates that the centrifugal force generated during the turn will be approximately 800. The heeling moment, coupled with the bow roll motion, generates a rolling moment of approximately 220. The total predicted roll angle will increase by 2.1°; a feedforward compensation signal is then generated to drive the anti-roll fins to output a reverse roll torque of 1020 in advance. This corresponds to an angle of attack adjustment of +5.5°; simultaneously, the bow thrusters output auxiliary compensation torque to suppress bow roll oscillations during steering. Testing has shown that the maximum roll angle during steering can be controlled within 2.5°.

[0122] By employing a working condition feedforward compensation mechanism, the attitude stability under sudden changes in working conditions is significantly improved. For typical intervention actions such as sudden changes in speed and steering maneuvers during navigation, the expected attitude fluctuation caused by changes in working conditions is calculated in advance using a multi-dimensional mapping matrix and a time-varying hydrodynamic model. Within the same cycle of the working condition execution, the expected fluctuation is converted into a feedforward signal to actively compensate the control system, thus mitigating the attitude impact caused by changes in working conditions in advance. This solves the lag problem of traditional feedback control of yacht navigation attitude and ensures the stability of yacht navigation attitude control throughout the entire navigation process.

[0123] Example 2:

[0124] Please see Figure 2 A multi-sensor yacht navigation attitude intelligent control system is used to realize a method for intelligent control of multi-sensor yacht navigation attitude; the system includes:

[0125] The yacht multi-source data acquisition and preprocessing module P1 is used to acquire hydrodynamic data, multi-source sensor data, attitude response data and actuator control data during pool tests and actual ship navigation, and to perform feature extraction and noise filtering, supporting the construction of hydrodynamic coefficient-attitude disturbance benchmark database and time-varying nonlinear hydrodynamic model.

[0126] The attitude dynamics benchmark model and multidimensional mapping matrix management module P2 is used to build, store and dynamically update the hydrodynamic coefficient-attitude disturbance benchmark database, time-varying nonlinear hydrodynamic model and sensor-attitude-control multidimensional mapping matrix;

[0127] The multi-degree-of-freedom attitude calculation and closed-loop control module P3 is used to calculate the yacht's multi-degree-of-freedom attitude parameters and relative attitude deviation in real time, and to perform control torque optimization and closed-loop control based on the sensor-attitude-control multi-dimensional mapping matrix.

[0128] The navigation condition feedforward compensation and execution drive module P4 is used to monitor the yacht's operating condition intervention actions, generate feedforward compensation signals, and work with the navigation actuators to complete the final attitude control output.

[0129] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A multi-sensor intelligent control method for yacht navigation attitude, characterized in that, The method includes the following steps: S1. Under gradient sea state and speed conditions, collect time-varying hydrodynamic coefficients and attitude response data of the yacht under different attitude motions, and establish a hydrodynamic coefficient-attitude disturbance reference database; combine the multi-degree-of-freedom coupling characteristics of the underactuated system to construct a time-varying nonlinear hydrodynamic model, and complete the spatiotemporal calibration and synchronous calibration of multi-source heterogeneous sensors. S2. Under typical steady-state navigation conditions, collect real-time data from multiple sensors, attitude motion parameters, and actuator control data. Combined with the time-varying nonlinear hydrodynamic model, identify the quantitative correlation between sensing features, attitude state, and control torque, and construct a multi-dimensional mapping matrix of sensing, attitude, and control. S3. During continuous navigation, multi-source sensors collect yacht motion data and environmental disturbance data online in real time, perform multi-source fusion state estimation, obtain the current accurate attitude parameters and disturbance characteristic parameters, call the time-varying nonlinear hydrodynamic model and the sensing-attitude-control multidimensional mapping matrix, and calculate the multi-degree-of-freedom decoupled control torque. S4. Calculate the relative attitude deviation of the current attitude parameters relative to the preset attitude threshold; when the relative attitude deviation exceeds the dynamic adaptive deviation threshold, call the sensor-attitude-control multidimensional mapping matrix, optimize and allocate the control torque compensation of each actuator based on actuator constraints, send it to the navigation actuator for closed-loop control, and synchronously update the sensor-attitude-control multidimensional mapping matrix and hydrodynamic model parameters; when the relative attitude deviation does not exceed the dynamic adaptive deviation threshold, smoothly update the current detection and control data to the sensor-attitude-control multidimensional mapping matrix. S5. When an intervention action is detected in the yacht's navigation condition, the sensor-attitude-control multidimensional mapping matrix and the time-varying nonlinear hydrodynamic model are invoked to calculate the expected attitude fluctuation caused by the intervention action. The expected attitude fluctuation is then converted into a feedforward compensation signal and superimposed onto the attitude control system.

2. The intelligent control method for multi-sensor yacht navigation attitude according to claim 1, characterized in that, The specific steps for establishing the hydrodynamic coefficient-attitude disturbance reference database in S1 include: S101 covers the entire speed range of zero-speed anchoring, low-speed cruising, medium-speed navigation, and high-speed navigation, as well as typical sea states from level 1 to level 5, and prepares multiple sets of gradient working condition test samples. S102. Perform dynamic response tests on all samples, record the amplitude and frequency of roll, pitch, and yaw angles under each working condition, as well as the response characteristics of the corresponding control torque; S103. The operating parameters, hydrodynamic coefficients, attitude response parameters, and control torque parameters are stored in a structured manner to form a hydrodynamic coefficient-attitude disturbance reference database.

3. The intelligent control method for multi-sensor yacht navigation attitude according to claim 1, characterized in that, The specific steps for constructing the time-varying nonlinear hydrodynamic model and multi-sensor calibration in S1 include: S11. Set up gradient speed and gradient sea state test groups, each group corresponding to constant speed and sea state level, and conduct towing pool test and sea sailing test on the target yacht. Simultaneously collect hydrodynamic damping coefficient, added mass coefficient and three-degree-of-freedom attitude response data of roll, pitch and yaw under different speeds, different drafts and different wave levels. S12. Based on experimental data, identify the time-varying characteristics of hydrodynamic coefficients, fit the nonlinear variation law of hydrodynamic coefficients with speed, draft and sea state, and combine the actuator configuration constraints of underactuated system to establish a three-degree-of-freedom coupled dynamic model of roll-pitch-bow. S13. Perform time synchronization calibration and spatial coordinate registration on multi-source heterogeneous sensors such as inertial measurement unit, satellite positioning unit, anemometer, wave height meter, rudder angle sensor, and thruster speed sensor to establish a sensor data acquisition system under a unified spatiotemporal reference.

4. The intelligent control method for multi-sensor yacht navigation attitude according to claim 1, characterized in that, When constructing the sensing-attitude-control multidimensional mapping matrix in S2, the underactuated actuator constraint optimization step is also incorporated: Based on the maximum output torque, response bandwidth, and operating degrees of freedom of each actuator, the performance constraint boundary of the actuator is established. Combining the attitude control priority of each degree of freedom, the torque distribution weight under multi-objective optimization is solved. The control quantity benchmark value in the sensing-attitude-control multidimensional mapping matrix is ​​constrained and corrected, and then the attitude control effect is matched accordingly.

5. The intelligent control method for multi-sensor yacht navigation attitude according to claim 1, characterized in that, The specific steps for calculating the multi-degree-of-freedom decoupling control torque in S3 include: S31. During the current control cycle, collect data on the yacht's angular acceleration, linear acceleration, position and heading, wind speed and direction, wave height and direction, rudder angle, and thrust through multi-source sensors. S32. Perform spatiotemporal registration, noise filtering, and feature extraction on multi-source sensor data, and perform state estimation using the extended Kalman fusion algorithm to obtain real-time attitude angles, angular velocities, and disturbance torque estimates for the three degrees of freedom of roll, pitch, and yaw. S33. Input the real-time attitude parameters and disturbance characteristic parameters into the time-varying nonlinear hydrodynamic model, combine the sensing-attitude-control multidimensional mapping matrix, and output the three-degree-of-freedom decoupled control torque through the adaptive decoupling control algorithm.

6. The intelligent control method for multi-sensor yacht navigation attitude according to claim 1, characterized in that, The specific steps of S4 include: S41. Using the preset attitude stability threshold as the reference value, calculate the ratio of the difference between the current real-time attitude parameters and the reference value to the reference value to obtain the attitude relative deviation. S42. Determine whether the relative deviation of the posture is greater than the dynamic adaptive deviation threshold; if yes, then execute S43; if no, then execute S44. S43. Based on the control weights of each actuator in the sensor-attitude-control multidimensional mapping matrix, prioritize adjusting the output of the high-weight core actuator, calculate the control torque fine-tuning compensation amount of each actuator based on the attitude relative deviation value and weight ratio, send it to the corresponding navigation actuator for closed-loop control, and simultaneously update the sensor-attitude-control multidimensional mapping matrix and hydrodynamic model parameters. S44. Keeping the current control parameters unchanged, use a preset exponential smoothing formula to smoothly update the current sensing data, attitude parameters and control data to the sensing-attitude-control multidimensional mapping matrix.

7. The intelligent control method for multi-sensor yacht navigation attitude according to claim 6, characterized in that, Before issuing the control torque fine-tuning compensation amount, S43 also includes an actuator boundary verification step: S431. Read the target actuator output value after superimposed control torque fine-tuning compensation; S432. Determine whether the target actuator output value falls within the preset actuator allowable output range; If the value exceeds the upper limit of the range, the control torque fine-tuning compensation will be forcibly truncated to the output value corresponding to the upper limit of the range. If the value is below the lower limit of the range, the control torque fine-tuning compensation will be forcibly truncated to the output value corresponding to the lower limit of the range.

8. The intelligent control method for multi-sensor yacht navigation attitude according to claim 1, characterized in that, The operational condition intervention actions in S5 include sudden speed changes, and the specific steps include: S51. Monitor the main engine throttle adjustment signal and speed setting command in real time through the ship's power control system; S52. Extract the hydrodynamic characteristic coefficients corresponding to the target speed from the sensing-attitude-control multidimensional mapping matrix, and calculate the expected pitch angle change and roll damping change caused by the speed change as the expected attitude fluctuation. S53. Generate a corresponding feedforward compensation signal based on the expected attitude fluctuation amount, and adjust the angle of attack of the anti-roll fin and the rudder angle in advance to counteract the attitude fluctuation caused by sudden changes in speed.

9. The intelligent control method for multi-sensor yacht navigation attitude according to claim 1, characterized in that, The working condition intervention actions in S5 also include steering maneuver actions; When a yacht steering command is detected, the expected heave moment and yaw coupling moment caused by the steering centrifugal force are calculated based on the steering angular velocity and the target rudder angle. The expected moment is then substituted into the sensor-attitude-control multidimensional mapping matrix to obtain the expected attitude fluctuation, and the corresponding anti-roll fin and thruster feedforward compensation signals are generated and sent to the attitude control system.

10. A multi-sensor intelligent control system for yacht navigation attitude, used to implement the multi-sensor intelligent control method for yacht navigation attitude as described in any one of claims 1-9; characterized in that, The system includes: The yacht multi-source data acquisition and preprocessing module is used to acquire hydrodynamic data, multi-source sensor data, attitude response data and actuator control data during pool tests and actual ship navigation, and to perform feature extraction and noise filtering, supporting the construction of hydrodynamic coefficient-attitude disturbance benchmark database and time-varying nonlinear hydrodynamic model; The attitude dynamics benchmark model and multidimensional mapping matrix management module is used to build, store and dynamically update the hydrodynamic coefficient-attitude disturbance benchmark database, time-varying nonlinear hydrodynamic model and sensing-attitude-control multidimensional mapping matrix; The multi-degree-of-freedom attitude calculation and closed-loop control module is used to calculate the yacht's multi-degree-of-freedom attitude parameters and relative attitude deviation in real time, and to perform control torque optimization and closed-loop control based on the sensor-attitude-control multi-dimensional mapping matrix. The navigation condition feedforward compensation and execution drive module is used to monitor the yacht's operating condition intervention actions, generate feedforward compensation signals, and work with the navigation actuators to complete the final attitude control output.

Images