A multi-modal automatic steering control method for a full rotation propeller

Abstract

The present application belongs to the field of automatic control of ships, and particularly relates to a multi-mode automatic steering control method suitable for a full-reversing propeller, comprising: S1, obtaining the Nomoto first-order model parameters of a ship; obtaining equivalent conversion coefficients through real ship calibration; S2, identifying the current sea state level, calculating the adaptive PID control parameters of the current control period, and determining the pipe width; S3, calculating the heading deviation, determining the control gain, performing discrete PID control operation on the heading deviation, and generating an equivalent rudder angle command; S4, solving the predicted heading angle; generating a lead compensation control amount, superimposing it with the equivalent rudder angle command, and obtaining the final equivalent rudder angle; S5, decomposing the required yawing moment into the rotation angle adjustment amount of the propeller relative to the current state and the screw speed adjustment amount; S6, outputting to the actuator after steering mode switching processing, and driving the full-reversing propeller to perform steering and speed adjustment actions. The method can realize reasonable moment distribution and high-precision retention of the heading control command.

Description

Technical Field

[0001] This invention belongs to the field of ship automatic control, and in particular relates to a multi-modal automatic steering control method adapted to azimuth thrusters. Background Technology

[0002] An autopilot is an important navigational device used to automatically adjust a ship's course or track. It collects feedback signals such as course and rudder angle, calculates control parameters, and drives the propellers to turn the rudder, enabling the ship to automatically maintain its course or route. Autopilots can reduce speed loss, shorten sailing time, save fuel, and simultaneously save manpower and reduce the workload of the crew.

[0003] Azimuth thrusters represent a significant development direction in modern ship propulsion technology. Unlike traditional ship propulsion methods, the propeller shaft of an azimuth thruster can rotate 360° around its axis of rotation in the horizontal plane, integrating propulsion and rudder functions into one unit without the need for an additional rudder system. Azimuth thrusters offer advantages such as high propulsion efficiency, small turning diameter, and flexible maneuverability, enabling complex maneuvers such as turning on the spot and lateral movement. They are widely used in various special-purpose vessels, including tugboats, engineering vessels, research vessels, and port operation vessels.

[0004] However, existing automatic steering technology still faces many challenges when adapted to azimuth thrusters, specifically in the following four aspects:

[0005] 1) Lack of specific adaptation for the characteristics of azimuth thrusters. Traditional autopilots can achieve heading control by outputting a single rudder angle control command, while azimuth thrusters require simultaneous control of two variables: propeller speed and azimuth angle. These two variables are coupled and jointly determine the ship's heading and speed. Existing autopilots cannot directly decompose the heading control command into the speed and azimuth angle control quantities required by the azimuth thruster, resulting in poor control performance;

[0006] 2) Control parameters cannot be adaptively adjusted. The maneuverability of ships varies significantly under different speeds, loads, and sea states. Most existing automatic steering systems have control parameters that are fixed at the factory or require manual adjustment. When the ship's operating status or environmental conditions change, the control parameters cannot be automatically adjusted to ensure optimal control performance.

[0007] 3) Lack of predictive control capability. Most existing automatic steering systems use feedback control based on the current heading deviation, which is an ex-post adjustment. When the ship has large inertia and a slow response, it is difficult to achieve high-precision heading maintenance.

[0008] 4) Switching between multiple steering modes presents shocks and discontinuities. Ship steering is typically configured with three modes: automatic, servo, and overriding in emergency situations. In existing technologies, switching between these three modes often employs simple switchovers or mechanical conversion devices, which are prone to sudden changes in control quantities during the switching process, leading to propeller shocks and ship attitude fluctuations.

[0009] Therefore, how to achieve reasonable torque distribution and high-precision maintenance of heading control commands through the synergistic effect of adaptive parameter optimization and model prediction compensation under the combined constraints of the dual-variable coupling characteristics of the azimuth thruster and the variable sea conditions, and ensure smooth and shock-free switching of multiple steering modes, has become an urgent problem to be solved. Summary of the Invention

[0010] To address the aforementioned shortcomings of existing technologies, the present invention aims to provide a multi-modal automatic steering control method adapted to azimuth thrusters. Under the combined constraints of the dual-variable coupling characteristics of azimuth thrusters and varying sea conditions, this method achieves reasonable torque distribution and high-precision maintenance of heading control commands through the synergistic effect of adaptive parameter optimization and model prediction compensation, while ensuring smooth and shock-free switching between multiple steering modes.

[0011] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0012] A multi-modal automatic steering control method adapted to a full-radius thruster includes the following steps:

[0013] S1. Obtain the parameters of the Nomoto first-order model of the ship, which describes the dynamic relationship between the bow roll rate and the equivalent rudder angle; obtain the equivalent conversion coefficient through actual ship calibration, which is the bow roll moment generated per unit equivalent rudder angle;

[0014] S2. In each control cycle, acquire the ship's current heading angle, current speed, bow rate, roll angle, and current slewing angle and speed of the propeller; identify the current sea state level based on the statistical characteristics of the roll angle; calculate the adaptive PID control parameters for the current control cycle based on the sea state level and the current speed through gain scheduling, and determine the duct width, which increases as the sea state level increases;

[0015] S3. Calculate the heading deviation based on the target heading and the current heading angle obtained in S2; compare the heading deviation with the pipe width determined in S2, and determine the control gain used when performing discrete PID control on the heading deviation based on the comparison result. The control gain used when the heading deviation is within the pipe width range is lower than the control gain used when it exceeds the pipe width. Perform discrete PID control on the heading deviation based on the determined control gain to generate an equivalent rudder angle command.

[0016] S4. Based on the Nomoto first-order model of S1, using the bow angular velocity obtained in S2 as the initial condition and the equivalent rudder angle command generated in S3 as the model input, solve the predicted bow angular velocity trajectory within a preset step size, and integrate the predicted bow angular velocity trajectory within the preset step size. The integration result is superimposed with the current heading angle obtained in S2 to obtain the predicted heading angle. The predicted heading deviation is obtained from the difference between the target heading and the predicted heading angle. An advance compensation control quantity is generated based on the predicted heading deviation, and the advance compensation control quantity is superimposed with the equivalent rudder angle command generated in S3 to obtain the final equivalent rudder angle.

[0017] S5. Based on the equivalent conversion coefficient of S1, convert the final equivalent rudder angle obtained in S4 into the required yaw torque; calculate the grading coefficient based on the heading deviation and the preset deviation threshold in S3, the grading coefficient being used to characterize the adjustment relationship between the magnitude of the heading deviation and the torque distribution weight between the turning angle control and the speed control; determine the torque distribution ratio between the turning angle control channel and the speed control channel based on the grading coefficient; using the current turning angle and current speed of the propeller obtained in S2 as the starting point for incremental calculation, decompose the required yaw torque into the turning angle adjustment amount and the propeller speed adjustment amount of the propeller relative to the current state according to the distribution ratio;

[0018] S6. The slewing angle control quantity and propeller speed control quantity obtained from S5 are processed by the steering mode switching and then output to the actuator to drive the azimuth thruster to perform steering and speed regulation actions, and return to S2 in the next control cycle to form a closed-loop control; the steering mode switching process includes: direct output when no mode switching occurs, gradual transition output with exponential decay when switching between automatic steering mode and servo steering mode, and immediate takeover output when switching to emergency over-control mode.

[0019] Compared with the prior art, the present invention has the following advantages:

[0020] 1. This method achieves a reasonable torque distribution for heading control commands under dual-variable coupling of a full-aperture propeller. Unlike traditional autopilots that only output a single rudder angle command, this method maps the final equivalent rudder angle to a uniform pitch torque using an equivalent conversion coefficient. It then dynamically calculates a grading coefficient based on the heading deviation, adaptively adjusting the torque distribution ratio between the yaw angle channel and the speed channel. Compared to existing fixed or simple proportional distribution methods, this method primarily uses speed fine-tuning under small deviation conditions and yaw angle adjustment under large deviation conditions, avoiding unnecessary frequent rudder turns and balancing response speed and actuator lifespan.

[0021] 2. This method achieves adaptive optimization of control parameters based on sea state and speed. In existing technologies, the PID parameters of autopilots are mostly factory-fixed values ​​or rely on manual tuning based on experience, making them unable to cope with changes in ship operating conditions. This method identifies sea state levels online using statistical characteristics of the roll angle, and uses this as a scheduling variable to calculate the PID gain and pipe width for the current cycle in real time. This allows the control parameters to automatically become more conservative as sea state deteriorates and more sensitive as sea state stabilizes, effectively overcoming the contradiction that fixed parameters become sluggish when conservative and oscillating when aggressive under varying operating conditions.

[0022] 3. Model-based predictive compensation significantly improves course-keeping accuracy. Existing autopilots adjust based on current deviations, essentially a reactive correction, which is prone to overshoot and oscillation when the ship has high inertia and a delayed response. This method uses a Nomoto first-order model to predict the bow roll rate forward and integrate it to obtain the predicted course angle. Based on this, a predictive compensation control quantity is generated and superimposed on the equivalent rudder angle command. This shifts the control action from responding only after the deviation has occurred to intervening when the deviation is about to occur, effectively suppressing course overshoot in ships with high inertia and improving steady-state keeping accuracy.

[0023] 4. Ensures smooth and shock-free switching between multiple steering modes. Existing technologies often employ hard switching between automatic, servo, and emergency overrun modes, where sudden changes in control input can easily cause propeller shocks and ship attitude fluctuations. This method designs an exponential decay gradual transition strategy for switching between automatic and servo modes, ensuring smooth convergence of control output; for emergency overrun, an immediate takeover mechanism is adopted to ensure safety priority, balancing the smoothness of the switching process with the immediacy of emergency response.

[0024] In summary, this method can achieve reasonable torque distribution and high-precision maintenance of heading control commands through the synergistic effect of adaptive parameter optimization and model prediction compensation under the combined constraints of the bivariate coupling characteristics of the azimuth thruster and the variable sea conditions, and ensure smooth and shock-free switching of multiple steering modes.

[0025] Preferably, in S1, the Nomoto first-order model is:

[0026] ;

[0027] Where T is the ship's maneuverability time constant; K is the yaw rate index; and r is the bow roll rate. Equivalent rudder angle;

[0028] The equivalent conversion coefficient satisfy:

[0029] ;

[0030] in, The yaw torque generated by the azimuth thruster.

[0031] This setup, using a simple first-order linear model and equivalent transformation coefficients, constructs a standardized mapping path from control commands to physical torque, ensuring real-time performance while laying a reliable model foundation for multi-stage collaborative control.

[0032] Preferably, in S2, identifying the current sea state level based on the statistical characteristics of the roll angle includes:

[0033] Calculate the variance of the roll angle :

[0034] ;

[0035] in, This is the sampled value of the i-th roll angle; The mean roll angle is N; N is the number of sampling points.

[0036] when At this time, the current sea state level is determined to be S = 1, representing calm sea state;

[0037] when At that time, the current sea state level is determined to be S = 2, representing a moderate sea state;

[0038] when At that time, the current sea state level is determined to be S = 3, representing severe sea state;

[0039] in, and These are the preset first standard deviation threshold and the second standard deviation threshold, respectively. .

[0040] This setup eliminates the need for external weather sensors or complex wave spectrum estimation. It enables online, low-computational-consumption sea state identification using only the ship's existing roll angle signal, providing a reliable and easily implemented condition criterion for subsequent gain scheduling.

[0041] Preferably, in S2, the calculation of the adaptive PID control parameters for the current control cycle through gain scheduling includes:

[0042] ;

[0043] ;

[0044] ;

[0045] in, , , These are the preset baseline PID control parameters; Sea state correction factor; For speed correction factor; , , These are the adaptive PID control parameters for the current control cycle.

[0046] Compared to single-variable scheduling or lookup table methods, this setup, with its simple product structure, simultaneously addresses the dual needs of suppressing oscillations when sea conditions deteriorate and compensating for maneuverability differences when speed changes, while also facilitating engineering calibration and online real-time calculations.

[0047] Preferably, the sea state correction factor The values ​​are 1.0, 0.7, and 0.4 for sea state grades S = 1, S = 2, and S = 3, respectively.

[0048] The speed correction factor for:

[0049] ;

[0050] in, The preset reference speed; This is the current speed.

[0051] With this setup, the sea state correction factor decreases in a stepwise manner as the sea state level increases, allowing the control gain to automatically decay in adverse sea conditions to suppress frequent rudder movements and heading oscillations caused by wave disturbances. The speed correction factor adopts a monotonically decreasing rational function form, so that the gain naturally decreases at high speeds to match the ship's sensitive maneuverability and easy overshoot characteristics at high speeds, while the gain increases at low speeds to compensate for insufficient rudder effectiveness.

[0052] Preferably, in S2, the pipe width W(S) is determined according to the sea state class S:

[0053] When S = 1, the pipe width W(S) = ;

[0054] When S = 2, the pipe width W(S) = ;

[0055] When S = 3, the pipe width W(S) = ;

[0056] in, , , This is a preset pipe width threshold corresponding to the sea state level, and < < .

[0057] This setup, with its sea state-driven graded pipeline width adjustment mechanism, ensures that the tolerance boundary for course control matches the actual sea state, balancing accurate tracking in adverse sea conditions with robustness and equipment protection in calm sea conditions.

[0058] Preferably, in S4, the step of solving the predicted trajectory of the bow angular velocity within a preset step size and integrating the predicted trajectory of the bow angular velocity within the preset step size includes:

[0059] Assuming the equivalent rudder angle remains constant, the Nomoto first-order model is discretized and solved to obtain a prediction step size of T. p Predicted bow roll rate at time :

[0060] ;

[0061] in, T is the current bow roll rate; p The prediction step size; T is the ship maneuvering time constant; The reversibility index; δ eq (k) represents the current equivalent rudder angle;

[0062] Integrate the predicted bow roll rate and add it to the current heading angle. By superimposing the values, the predicted heading angle is obtained. :

[0063] ;

[0064] The predicted heading deviation for:

[0065] ;

[0066] in, For the target course;

[0067] The advance compensation control quantity for:

[0068] ;

[0069] in, To predict the compensation gain coefficient;

[0070] The final equivalent rudder angle for:

[0071] ;

[0072] in, The equivalent rudder angle command generated for S3.

[0073] This setup amplifies the deviation between the predicted and target headings and adds it as a lead compensation to the feedback rudder angle command. This allows the controller to detect the heading deviation trend caused by the ship's inertial lag in advance and actively apply corrections. It effectively alleviates the overshoot and oscillation problems caused by response lag in the heading control of ships with large inertia. Moreover, it uses a closed-loop analytical solution instead of rolling optimization throughout the process, resulting in extremely low computational burden and meeting the requirements for real-time embedded deployment.

[0074] Preferably, in S5, the calculation of the grading coefficient based on the heading deviation in S3 and the preset deviation threshold includes:

[0075] when hour, = 1;

[0076] when hour, ;

[0077] when hour, = 0.

[0078] Where λ(eψ) is the grading coefficient; For heading deviation; To preset a large deviation threshold, To preset a small deviation threshold, and < .

[0079] This setup, using a dual-threshold piecewise linear grading coefficient, achieves continuous, shock-free grading adjustment of control intensity, organically unifying stable maintenance under small deviations with rapid correction under large deviations, and taking into account steady-state accuracy, dynamic response, and servo protection.

[0080] Preferably, in S5, the step of decomposing the required yaw torque into the thruster's rotation angle adjustment and the propeller speed adjustment relative to the current state includes:

[0081] The required bow roll torque for:

[0082] ;

[0083] in, This is the final equivalent rudder angle obtained from S4; These are the equivalent conversion factors;

[0084] The rotation angle adjustment amount for:

[0085] ;

[0086] in, This is the current rotation angle; This represents the maximum change in rotation angle in a single step. This refers to the gyration gain under small deviations. It is a symbolic function;

[0087] The propeller speed adjustment amount for:

[0088] ;

[0089] in, The current rotational speed; For speed maintenance compensation; This is a steering speed compensation item;

[0090] Δn speed = K v ·(V ref - V(k));

[0091] Δn turn = K n ·|α cmd (k) - α(k)|·V(k);

[0092] Among them, K v V is the speed compensation coefficient; ref V(k) represents the target speed; V(k) represents the current speed; K represents the target speed. n This is the steering compensation coefficient;

[0093] when|e ψ | < θ deadzone When the heading deviation is determined to have reached a steady state, turn angle zeroing control is activated:

[0094] α cmd (k) = α(k)·(1 - K return ·Δt);

[0095] Where, θ deadzone K is the preset dead zone threshold;return Δt is the rotation angle attenuation coefficient; Δt is the control period.

[0096] This configuration decouples the required yaw torque into adjustments for two types of actuators: the slewing angle and the propeller speed. This adjustment is achieved through the aforementioned grading coefficients. In swerve control, a smooth transition is achieved by rapidly correcting large deviations with the maximum step size and finely adjusting proportionally for small deviations. The speed command is simultaneously superimposed with a speed maintenance compensation term and a turning energy consumption compensation term, so that the thrust loss caused by changes in swerve during the turning process is actively compensated to maintain stable speed. After the heading deviation enters the dead zone, the swerve exponential decay zeroing mechanism is activated to automatically return the thruster to the correct position to eliminate parasitic drag and energy consumption in steady state.

[0097] Preferably, in S6, the steering mode switching process includes:

[0098] When switching between automatic steering mode and servo steering mode, the control quantity is adjusted according to the following exponential decay function. Perform a gradual transition:

[0099] ;

[0100] in, This is the control quantity used to switch the start time. The control variable for the target mode; t is the preset transition time constant; t is the time elapsed after the switch.

[0101] when When the handover is complete, it is determined that the handover is finished; among them, The preset threshold for completing the switchover;

[0102] When switching to emergency over-control steering mode, the ongoing gradual transition is immediately interrupted, and the control quantity of emergency over-control steering mode is directly output.

[0103] When switching from emergency over-control steering mode to automatic steering mode or follow-up steering mode, the switch should be performed only after the operator confirms the cancellation.

[0104] This setup, with its layered switching strategy of smooth transition in normal operation, instantaneous emergency response, and manual confirmation upon exit, achieves an optimal balance between operational comfort, emergency response speed, and a closed-loop human-machine safety system. Attached Figure Description

[0105] To make the objectives, technical solutions, and advantages of the invention clearer, the invention will now be described in further detail with reference to the accompanying drawings, wherein:

[0106] Figure 1 This is a flowchart of the method;

[0107] Figure 2 This is a diagram of the overall system architecture used in the implementation of this method;

[0108] Figure 3 This is a schematic diagram of the predictive heading compensation principle in Example 1;

[0109] Figure 4 The heading deviation classification coefficient in Example 1 Line graph;

[0110] Figure 5 This is a schematic diagram of the mode switching state machine logic in Example 1. Detailed Implementation

[0111] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0112] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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 a part of the embodiments of the present invention, not all of them. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to represent selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0113] Example 1

[0114] like Figure 1 As shown, this invention provides a multi-modal automatic steering control method adapted to a full-radius thruster. Specifically, the overall architecture block diagram of the system used to implement this method is shown below. Figure 2 As shown.

[0115] This method includes the following steps:

[0116] S1. Obtain the parameters of the Nomoto first-order model of the ship, which describes the dynamic relationship between the bow roll rate and the equivalent rudder angle; obtain the equivalent conversion coefficient through actual ship calibration, which is the bow roll moment generated per unit equivalent rudder angle.

[0117] In specific implementation, the Nomoto first-order model is as follows:

[0118] ;

[0119] Where T is the ship's maneuverability time constant; K is the yaw rate index; and r is the bow roll rate. Equivalent rudder angle;

[0120] The equivalent conversion coefficient satisfy:

[0121] ;

[0122] in, The yaw torque generated by the azimuth thruster.

[0123] To facilitate better understanding, the following explanation is provided.

[0124] Establish a ship-based coordinate system with the ship's center of gravity as the origin. Considering the three degrees of freedom of pitch, sway, and yaw, the ship's maneuvering motion equations are as follows:

[0125] ;

[0126] ;

[0127] ;

[0128] Where: m is the ship's mass (kg); I z denoted as , where is the moment of inertia of the ship about the z-axis (kg·m²); u is the sway velocity (m / s); v is the roll velocity (m / s); r is the bow roll rate, i.e., the turning rate (rad / s); X, Y, and N are the longitudinal force (N), transverse force (N), and bow roll moment (N·m) acting on the ship, respectively.

[0129] Based on the structural characteristics of the azimuth thruster, a mathematical relationship between the thrust direction angle and the forces acting on the ship is established.

[0130] For ships equipped with azimuth thrusters, let the thrust of the thruster be... The thrust direction angle (i.e., the rotation angle) is α. p The thruster installation location coordinates are (x p , y p If the azimuth thruster exerts a force and torque on the ship, then the force and torque exerted by the azimuth thruster on the ship are:

[0131] ;

[0132] ;

[0133] ;

[0134] When a ship is equipped with n azimuth thrusters, the resultant force and resultant torque are the superposition of the forces from each thruster:

[0135] ;

[0136] ;

[0137] ;

[0138] Linearizing the three-degree-of-freedom equations and neglecting the changes in oscillation motion, we obtain the classic Nomoto first-order model:

[0139] ;

[0140] in:

[0141] T is the ship's maneuverability time constant (s), K is the turning index (1 / s), and δeq is the equivalent rudder angle (rad).

[0142] Therefore, for a full-radius thruster, the relationship between the yaw torque Np generated by the thruster and the equivalent rudder angle δeq is defined as follows:

[0143] ;

[0144] Right now: ;

[0145] Wherein, Nref is the yaw moment (N·m / rad) generated by a unit equivalent rudder angle, obtained from actual ship calibration tests, and characterizes the equivalent conversion coefficient between the azimuth thruster and the conventional thruster in the torque-rudder angle relationship.

[0146] Integrating the Nomoto first-order model, we obtain the relationship between the heading angle ψ and the equivalent rudder angle: .

[0147] S2. In each control cycle, acquire the ship's current heading angle, current speed, bow angular velocity, roll angle, and the current swivel angle and current speed of the propeller; identify the current sea state level based on the statistical characteristics of the roll angle; calculate the adaptive PID control parameters for the current control cycle through gain scheduling based on the sea state level and the current speed, and determine the duct width, which increases as the sea state level increases.

[0148] In practice, this invention automatically identifies sea state levels by analyzing the statistical characteristics of the roll angle, without the need for additional sensors.

[0149] The step of identifying the current sea state level based on the statistical characteristics of the roll angle includes:

[0150] Calculate the variance of the roll angle :

[0151] ;

[0152] in, This is the sampled value of the i-th roll angle; The mean roll angle is N; N is the number of sampling points.

[0153] Sea state rating determination:

[0154] ;

[0155] in, and These are the preset first standard deviation threshold and the second standard deviation threshold, respectively. In practice, specific values ​​can be determined based on the ship type and actual ship test data.

[0156] In specific implementation, the calculation of adaptive PID control parameters for the current control cycle through gain scheduling includes:

[0157] ;

[0158] ;

[0159] ;

[0160] in, , , These are the preset baseline PID control parameters; Sea state correction factor; For speed correction factor; , , These are the adaptive PID control parameters for the current control cycle.

[0161] Sea state correction factor f S (S) segmentation definition:

[0162] ;

[0163] Reduce gain in rough sea conditions to prevent the controller from overreacting to wave disturbances and reduce unnecessary servo movements.

[0164] The speed correction factor for:

[0165] ;

[0166] in, The preset reference speed; This is the current speed.

[0167] At high speed Reduce the amount of control input to prevent excessive input at high speeds. At low speeds... Approaching 1, a lower limit is added in actual engineering to avoid excessively low gain at low speeds.

[0168] The pipe width W(S) is determined according to the sea state class S:

[0169] ;

[0170] in, , , This is a preset pipe width threshold corresponding to the sea state level, and < < In practice, typical values ​​are W1 = 2°, W2 = 5°, and W3 = 10°.

[0171] To better understand the adaptive adjustment process in response to changes in sea state, the following example illustrates the system's adaptive adjustment process when sea state changes from calm to rough.

[0172] Initial state (calm sea state):

[0173] Sea state rating S = 1, roll angle standard deviation σ φ = 0.8° (below the threshold σ1 = 2°);

[0174] Current speed V = 5 knots, reference speed V0 = 5 knots, speed correction factor f V (V) = V0 / (V + V0) = 0.5;

[0175] Reference PID gain: K p0 = 2.5, K i0 = 0.01, K d0 = 8.0;

[0176] Actual PID gain (calm sea state): K p = K p0 · f S · f V = 2.5 × 1.0 × 0.5 = 1.25, K i = 0.005, K d = 4.0;

[0177] Pipe width W = 2°;

[0178] Heading accuracy: ±0.8°.

[0179] The process of sea state deterioration:

[0180] Sea conditions gradually deteriorated, and the standard deviation of the roll angle increased to σ.φ = 4.5°;

[0181] The parameter adaptive tuning unit identified the sea state level switching from S=1 to S=2 (σ φ > σ1);

[0182] Sea state correction factor f S The value decreased from 1.0 to 0.7, combined with the speed correction factor f. V = 0.5, PID gain automatically decreases: K p = 2.5 × 0.7 × 0.5 = 0.875, K i = 0.01 × 0.7 × 0.5 = 0.0035, K d = 8.0 × 0.7 × 0.5 = 2.8;

[0183] The pipe width was increased from W1=2° to W2=5° to reduce unnecessary steering maneuvers;

[0184] Sea conditions continued to deteriorate, σ φ Increased to 8.2°, the system identified S=3 (severe);

[0185] The sea state correction factor was further reduced to 0.4, and the PID gain was reduced to: K p = 2.5 × 0.4 × 0.5 = 0.5, K i = 0.01 × 0.4 × 0.5 = 0.002, K d = 8.0 × 0.4 × 0.5 = 1.6; The pipe width is increased to W3=10°;

[0186] Throughout the entire sea state change process, the system automatically adjusts parameters without human intervention;

[0187] Under adverse sea conditions, the heading accuracy is ±3°, and the servo motor operation frequency is reduced by approximately 60%.

[0188] S3. Calculate the heading deviation based on the target heading and the current heading angle obtained in S2; compare the heading deviation with the pipe width determined in S2, and determine the control gain used when performing discrete PID control on the heading deviation based on the comparison result. The control gain used when the heading deviation is within the pipe width range is lower than the control gain used when it exceeds the pipe width. Perform discrete PID control on the heading deviation based on the determined control gain to generate an equivalent rudder angle command.

[0189] In practical implementation, when the heading deviation |e ψ | ≤ W (S)In normal conditions, the controller only makes minor adjustments; when the heading deviation exceeds the pipe width, the controller switches to active correction mode, increasing the control output. This mechanism effectively reduces ineffective and frequent actions of the servo motor in adverse sea conditions, extending the servo motor's service life.

[0190] S4. Based on the Nomoto first-order model of S1, using the bow roll rate obtained in S2 as the initial condition and the equivalent rudder angle command generated in S3 as the model input, solve the predicted bow roll rate trajectory within a preset step size, and integrate the predicted bow roll rate trajectory within the preset step size. The integration result is superimposed with the current heading angle obtained in S2 to obtain the predicted heading angle. The predicted heading deviation is obtained from the difference between the target heading and the predicted heading angle. An advance compensation control quantity is generated based on the predicted heading deviation, and the advance compensation control quantity is superimposed with the equivalent rudder angle command generated in S3 to obtain the final equivalent rudder angle.

[0191] In specific implementation, the step of solving the predicted trajectory of the bow angular velocity within a preset step size and integrating the predicted trajectory of the bow angular velocity within the preset step size includes:

[0192] Assuming the equivalent rudder angle remains constant (zero-order hold), the Nomoto first-order model is discretized and solved to obtain a prediction step size of T. p Predicted bow roll rate at time :

[0193] ;

[0194] in, T is the current bow roll rate; p The prediction step size; T is the ship maneuvering time constant; The reversibility index; δ eq (k) represents the current equivalent rudder angle;

[0195] Integrate the predicted bow roll rate and add it to the current heading angle. By superimposing the values, the predicted heading angle is obtained. :

[0196] ;

[0197] A schematic diagram of the predictive heading compensation principle is shown below. Figure 3 As shown.

[0198] The predicted heading deviation for:

[0199] ;

[0200] in, For the target course;

[0201] The advance compensation control quantity for:

[0202] ;

[0203] in, To predict the compensation gain coefficient;

[0204] The final equivalent rudder angle for:

[0205] ;

[0206] in, The equivalent rudder angle command generated for S3.

[0207] The deviation between the predicted and target headings is amplified and then added to the feedback rudder angle command as a lead compensation. This allows the controller to detect the heading deviation trend caused by the ship's inertial lag in advance and actively apply corrections. This effectively alleviates the overshoot and oscillation problems caused by response lag in the heading control of ships with large inertia. Moreover, the entire process uses a closed-loop analytical solution instead of rolling optimization, resulting in extremely low computational burden and meeting the requirements for real-time embedded deployment.

[0208] S5. Based on the equivalent conversion coefficient of S1, convert the final equivalent rudder angle obtained in S4 into the required bow torque; calculate the grading coefficient based on the heading deviation and the preset deviation threshold of S3. The grading coefficient is used to characterize the adjustment relationship between the magnitude of the heading deviation and the torque distribution weight between the turning angle control and the speed control; determine the torque distribution ratio between the turning angle control channel and the speed control channel based on the grading coefficient; using the current turning angle and current speed of the propeller obtained in S2 as the starting point for incremental calculation, decompose the required bow torque into the turning angle adjustment amount and the propeller speed adjustment amount of the propeller relative to the current state according to the distribution ratio.

[0209] In practice, heading deviation for:

[0210] ;

[0211] Where, ψ ref Let ψ(k) be the target heading, and let ψ(k) be the actual heading at the current time k.

[0212] The classification coefficient is calculated based on the heading deviation of S3 and the preset deviation threshold:

[0213] ;

[0214] Where λ(eψ) is the grading coefficient; To preset a large deviation threshold, To preset a small deviation threshold, and < .

[0215] In specific implementation, order It is 3°. If the angle is 10°, then the heading deviation classification coefficient can be obtained. Curve graph as Figure 4 As shown.

[0216] In specific implementation, the step of decomposing the required yaw torque into the thruster's rotation angle adjustment and propeller speed adjustment relative to the current state includes:

[0217] The required bow roll torque for:

[0218] ;

[0219] in, This is the final equivalent rudder angle obtained from S4; These are the equivalent conversion factors;

[0220] The rotation angle adjustment amount for:

[0221] ;

[0222] in, This is the current rotation angle; This represents the maximum change in rotation angle in a single step. This refers to the gyration gain under small deviations. It is a symbolic function;

[0223] The propeller speed adjustment amount for:

[0224] ;

[0225] in, The current rotational speed; For speed maintenance compensation; This is a steering speed compensation item;

[0226] Δn speed = K v ·(V ref - V(k));

[0227] Δn turn = K n ·|α cmd (k) - α(k)|·V(k);

[0228] Among them, K vV is the speed compensation coefficient; ref V(k) represents the target speed; V(k) represents the current speed; K represents the target speed. n This is the steering compensation coefficient;

[0229] when|e ψ | < θ deadzone When the heading deviation is determined to have reached a steady state, turn angle zeroing control is activated:

[0230] α cmd (k) = α(k)·(1 - K return ·Δt);

[0231] Where, θ deadzone K is the preset dead zone threshold; return Δt is the rotation angle attenuation coefficient, ensuring that the thruster gradually returns to the straight-line position during straight-line navigation; Δt is the control period.

[0232] The required yaw torque is decoupled into adjustment amounts for two types of actuators: slewing angle and propeller speed, and then determined using the aforementioned grading coefficients. In swerve control, a smooth transition is achieved by rapidly correcting large deviations with the maximum step size and finely adjusting proportionally for small deviations. The speed command is simultaneously superimposed with a speed maintenance compensation term and a turning energy consumption compensation term, so that the thrust loss caused by changes in swerve during the turning process is actively compensated to maintain stable speed. After the heading deviation enters the dead zone, the swerve exponential decay zeroing mechanism is activated to automatically return the thruster to the correct position to eliminate parasitic drag and energy consumption in steady state.

[0233] S6. The slewing angle control quantity and propeller speed control quantity obtained from S5 are processed by the steering mode switching and output to the actuator to drive the slewing thruster to perform steering and speed regulation actions, and return to S2 in the next control cycle to form a closed loop control.

[0234] In specific implementation, the steering mode switching process includes:

[0235] When switching between automatic steering mode and servo steering mode, the control quantity is adjusted according to the following exponential decay function. Perform a gradual transition:

[0236] ;

[0237] in, This is the control quantity used to switch the start time. The control variable for the target mode; t is the preset transition time constant; t is the time elapsed after the switch.

[0238] This function ensures that the control input does not change abruptly during switching, but rather gradually approaches the target value in an exponential decay manner.

[0239] when When the handover is complete, it is determined that the handover is finished; among them, The preset threshold for completing the switchover;

[0240] When switching to emergency over-control steering mode, the ongoing gradual transition is immediately interrupted, and the control quantity of emergency over-control steering mode is directly output.

[0241] When switching from emergency over-control steering mode to automatic steering mode or follow-up steering mode, the switch should be performed only after the operator confirms the cancellation.

[0242] By employing a layered switching strategy that combines smooth transition in normal operation with instantaneous emergency response and manual confirmation upon exit, the optimal balance is achieved between operational comfort, emergency response speed, and a closed-loop human-machine safety system.

[0243] like Figure 5 As shown, for better understanding, the logic of the mode switching state machine is explained as follows:

[0244] Mode switching state machine logic:

[0245] Automatic steering ↔ Follow-up steering: The operator initiates the switch via the mode selection switch, and the transition is completed gradually through the multi-mode smooth switching unit;

[0246] Follow-up steering → Emergency over-control steering: When the over-control button is triggered, the gradual transition is immediately interrupted, and the emergency mode control quantity is directly output (hard switch, prioritizing emergency response).

[0247] Automatic steering → Emergency override steering: Same as follow-up → Emergency override steering, immediate hard switch;

[0248] Emergency overrun steering → Follow-up steering: Operator confirmation of emergency deactivation is required;

[0249] Emergency over-control steering → Autosteer: Operator confirmation required to deactivate the emergency.

[0250] Unlike traditional autopilots that only output a single rudder angle command, this method maps the final equivalent rudder angle to a uniform pitch torque using an equivalent conversion coefficient. It then dynamically calculates a grading coefficient based on the magnitude of the heading deviation, adaptively adjusting the torque distribution ratio between the yaw angle and speed channels. Compared to existing technologies with fixed or simple proportional distributions, this method primarily uses speed fine-tuning under small deviation conditions and yaw angle adjustment under large deviation conditions, avoiding unnecessary frequent rudder turns and balancing response speed and actuator lifespan. Furthermore, this method employs an exponential decay gradual transition strategy for switching between automatic and servo modes, ensuring smooth convergence of control output. For emergency overruns, an immediate takeover mechanism is used to ensure safety priority, balancing the smoothness of the switching process with the immediacy of emergency response. This ensures smooth and shock-free switching between multiple steering modes.

[0251] To better understand the smooth switching of steering modes in this method, the following example illustrates the process of switching from automatic steering mode to servo steering mode.

[0252] State before switching:

[0253] In automatic steering mode, the PID controller outputs the equivalent rudder angle δ cmd = 5.2° (right rudder);

[0254] The two thrusters rotate at angles of α1 = 15° and α2 = -10°, respectively.

[0255] The propeller speeds are n1 = 450 rpm and n2 = 420 rpm.

[0256] Switching process:

[0257] The operator issues the command "Switch to follow mode";

[0258] The multimodal smooth switching unit records the current control quantity u start (Including the rotational speed and slewing angle of the two thrusters);

[0259] In follow-up mode, the operator sets the handwheel to position 0 (direct navigation), and the target control quantity u target Corresponding to direct flight status;

[0260] The gradual transition function uses a time constant τ s = Smooth transition begins after 3 seconds;

[0261] Approximately 15 seconds later (5τ) s When the difference between the control input and the target value is less than the threshold ε = 0.1°, the switching is complete;

[0262] Throughout the entire switching process, the ship's heading fluctuation was less than ±0.5°, with no rudder impact.

[0263] To better illustrate the automatic heading-keeping function of this method, a port tugboat equipped with twin azimuth thrusters is used as an example for the following illustration.

[0264] Ship parameters: Overall length: 32 m, Beam: 10 m, Depth: 4.5 m

[0265] Full load displacement: approximately 500 t

[0266] Equipped with two azimuth thrusters, each with a maximum power of 1200 kW.

[0267] Propeller installation location: symmetrically positioned at the stern, longitudinally from the center of gravity at x p = 12 m, lateral distance y p = ±2.5 m (negative for port, positive for starboard); default is the same-direction cooperative control mode.

[0268] Control parameter settings:

[0269] Nomoto model parameters: T = 8 s, K = 0.051 / s

[0270] Baseline PID parameters: K p0 = 2.5, K i0 = 0.01, K d0 = 8.0

[0271] Heading deviation thresholds: θ1 = 10°, θ2 = 3°

[0272] Prediction step size: T p = 5 s

[0273] Pipe width: W1 = 2°, W2 = 5°, W3 = 10°

[0274] Work process:

[0275] After the system starts up, it enters automatic steering mode and sets the target heading to 090°.

[0276] The heading detection unit collects compass heading angle, GPS speed, and yaw rate in real time;

[0277] Calculate the heading deviation e ψ The PID control law outputs the equivalent rudder angle command δ. cmd ;

[0278] The thrust vector decomposition element will δ cmd It is broken down into commands for the rotational speed and slewing angle of the two thrusters;

[0279] The predictive heading compensator predicts the heading 5 seconds later, calculates the advance compensation amount, and adds it to the control command.

[0280] The actuator interface sends control commands to the propeller inverter and the slewing mechanism;

[0281] The ship maintains a heading accuracy of better than ±1° in calm sea conditions.

[0282] This method identifies sea state levels online using statistical characteristics of roll angles and uses these levels as scheduling variables to calculate the PID gain and pipe width for the current cycle in real time. This allows control parameters to automatically become more conservative as sea state deteriorates and more sensitive as sea state stabilizes, effectively overcoming the contradiction that fixed parameters become sluggish when conservative and oscillating when aggressive under varying operating conditions. Furthermore, this method uses a Nomoto first-order model to predict and integrate the bow roll rate to obtain the predicted heading angle. Based on this, a lead-compensation control quantity is generated and superimposed on the equivalent rudder angle command. This shifts the control action from responding only after a deviation has occurred to intervening as soon as a deviation is about to occur, effectively suppressing heading overshoot in ships with large inertia and improving steady-state holding accuracy.

[0283] This method can achieve reasonable torque distribution and high-precision maintenance of heading control commands through the combined constraints of the bivariate coupling characteristics of the azimuth thruster and the variable sea conditions, and ensure smooth and shock-free switching of multiple steering modes.

[0284] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit the technical solutions. Those skilled in the art should understand that any modifications or equivalent substitutions to the technical solutions of the present invention without departing from the spirit and scope of the present invention should be covered within the scope of the claims of the present invention.

Claims

1. A multi-mode automatic steering control method adapted to a full-radius thruster, characterized in that, Includes the following steps: S1. Obtain the parameters of the Nomoto first-order model of the ship, which describes the dynamic relationship between the bow roll rate and the equivalent rudder angle; obtain the equivalent conversion coefficient through actual ship calibration, which is the bow roll moment generated per unit equivalent rudder angle; S2. In each control cycle, acquire the ship's current heading angle, current speed, bow rate, roll angle, and current slewing angle and speed of the propeller; identify the current sea state level based on the statistical characteristics of the roll angle; calculate the adaptive PID control parameters for the current control cycle based on the sea state level and the current speed through gain scheduling, and determine the duct width, which increases as the sea state level increases; S3. Calculate the heading deviation based on the target heading and the current heading angle obtained in S2; compare the heading deviation with the pipe width determined in S2, and determine the control gain used when performing discrete PID control on the heading deviation based on the comparison result. The control gain used when the heading deviation is within the pipe width range is lower than the control gain used when it exceeds the pipe width. Perform discrete PID control on the heading deviation based on the determined control gain to generate an equivalent rudder angle command. S4. Based on the Nomoto first-order model of S1, using the bow angular velocity obtained in S2 as the initial condition and the equivalent rudder angle command generated in S3 as the model input, solve the predicted bow angular velocity trajectory within a preset step size, and integrate the predicted bow angular velocity trajectory within the preset step size. The integration result is superimposed with the current heading angle obtained in S2 to obtain the predicted heading angle. The predicted heading deviation is obtained from the difference between the target heading and the predicted heading angle. An advance compensation control quantity is generated based on the predicted heading deviation, and the advance compensation control quantity is superimposed with the equivalent rudder angle command generated in S3 to obtain the final equivalent rudder angle. S5. Based on the equivalent conversion coefficient of S1, convert the final equivalent rudder angle obtained in S4 into the required bow torque; calculate the grading coefficient based on the heading deviation and the preset deviation threshold in S3. The grading coefficient is used to characterize the adjustment relationship between the magnitude of the heading deviation and the torque distribution weight between the turning angle control and the speed control; determine the torque distribution ratio between the turning angle control channel and the speed control channel based on the grading coefficient. Using the current slewing angle and current rotation speed of the thruster obtained in S2 as the starting point for incremental calculation, the required yaw torque is decomposed into the slewing angle adjustment amount and propeller speed adjustment amount of the thruster relative to the current state according to the allocation ratio. S6. The slewing angle control quantity and propeller speed control quantity obtained from S5 are processed by the steering mode switching and output to the actuator to drive the slewing thruster to perform steering and speed regulation actions, and return to S2 in the next control cycle to form a closed loop control. The steering mode switching process includes: direct output when no mode switching occurs, gradual transition output with exponential decay when switching between automatic steering mode and servo steering mode, and immediate takeover output when switching to emergency over-control mode.

2. The multi-mode automatic steering control method adapted to a full-radius thruster according to claim 1, characterized in that, In S1, the Nomoto first-order model is: ； Where T is the ship's maneuverability time constant; K is the yaw rate index; and r is the bow roll rate. Equivalent rudder angle; The equivalent conversion coefficient satisfy: ； in, The yaw torque generated by the azimuth thruster.

3. The multi-mode automatic steering control method adapted to a full-radius thruster according to claim 1, characterized in that, In S2, identifying the current sea state level based on the statistical characteristics of the roll angle includes: Calculate the variance of the roll angle : ； in, This is the sampled value of the i-th roll angle; The mean roll angle is N; N is the number of sampling points. when At this time, the current sea state level is determined to be S = 1, representing calm sea state; when At that time, the current sea state level is determined to be S = 2, representing a moderate sea state; when At that time, the current sea state level is determined to be S = 3, representing severe sea state; in, and These are the preset first standard deviation threshold and the second standard deviation threshold, respectively. .

4. The multi-mode automatic steering control method adapted to a full-radius thruster according to claim 3, characterized in that, In S2, the calculation of the adaptive PID control parameters for the current control cycle through gain scheduling includes: ； ； ； in, , , These are the preset baseline PID control parameters; Sea state correction factor; For speed correction factor; , , These are the adaptive PID control parameters for the current control cycle.

5. The multi-mode automatic steering control method adapted to a full-radius thruster according to claim 4, characterized in that, The sea state correction factor The values ​​are 1.0, 0.7, and 0.4 for sea state grades S = 1, S = 2, and S = 3, respectively. The speed correction factor for: ； in, The preset reference speed; This is the current speed.

6. The multi-mode automatic steering control method adapted to a full-radius thruster according to claim 3, characterized in that, In S2, the pipe width W(S) is determined according to the sea state class S: When S = 1, the pipe width W(S) = ; When S = 2, the pipe width W(S) = ; When S = 3, the pipe width W(S) = ; in, , , This is a preset pipe width threshold corresponding to the sea state level, and < < .

7. The multi-mode automatic steering control method adapted to a full-radius thruster according to claim 1, characterized in that, In S4, the step of solving the predicted trajectory of the bow angular velocity within a preset step size and integrating the predicted trajectory of the bow angular velocity within the preset step size includes: Assuming the equivalent rudder angle remains constant, the Nomoto first-order model is discretized and solved to obtain a prediction step size of T. p Predicted bow roll rate at time : ； in, T is the current bow roll rate; p The prediction step size; T is the ship maneuvering time constant; The reversibility index; δ eq (k) represents the current equivalent rudder angle; Integrate the predicted bow roll rate and add it to the current heading angle. By superimposing the values, the predicted heading angle is obtained. : ； The predicted heading deviation for: ； in, For the target course; The advance compensation control quantity for: ； in, To predict the compensation gain coefficient; The final equivalent rudder angle for: ； in, The equivalent rudder angle command generated for S3.

8. The multi-mode automatic steering control method adapted to a full-radius thruster according to claim 1, characterized in that, In S5, the calculation of the classification coefficient based on the heading deviation in S3 and the preset deviation threshold includes: when hour, = 1; when hour, ; when hour, = 0. Where λ(eψ) is the grading coefficient; For heading deviation; To preset a large deviation threshold, To preset a small deviation threshold, and < .

9. The multi-mode automatic steering control method adapted to a full-radius thruster according to claim 8, characterized in that, In S5, the step of decomposing the required yaw torque into the thruster's rotation angle adjustment and propeller speed adjustment relative to the current state includes: The required bow roll torque for: ； in, This is the final equivalent rudder angle obtained from S4; These are the equivalent conversion factors; The rotation angle adjustment amount for: ； in, This is the current rotation angle; This represents the maximum change in rotation angle in a single step. This refers to the gyration gain under small deviations. It is a symbolic function; The propeller speed adjustment amount for: ； in, The current rotational speed; For speed maintenance compensation; This is a steering speed compensation item; Δn speed = K v ·(In ref - V(k); Δn turn = K n ·|α cmd (k) - α(k)|·V(k); Among them, K v V is the speed compensation coefficient; ref V(k) represents the target speed; V(k) represents the current speed; K represents the target speed. n This is the steering compensation coefficient; when|e ψ | < θ deadzone When the heading deviation is determined to have reached a steady state, turn angle zeroing control is activated: a cmd (k) = α(k)·(1 - K return ·Δt); Where, θ deadzone K is the preset dead zone threshold; return Δt is the rotation angle attenuation coefficient; Δt is the control period.

10. The multi-mode automatic steering control method adapted to a full-radius thruster according to claim 1, characterized in that, In S6, the steering mode switching process includes: When switching between automatic steering mode and servo steering mode, the control quantity is adjusted according to the following exponential decay function. Perform a gradual transition: ； in, This is the control quantity used to switch the start time. The control variable for the target mode; t is the preset transition time constant; t is the time elapsed after the switch. when When the handover is complete, it is determined that the handover is finished; among them, The preset threshold for completing the switchover; When switching to emergency over-control steering mode, the ongoing gradual transition is immediately interrupted, and the control quantity of emergency over-control steering mode is directly output. When switching from emergency over-control steering mode to automatic steering mode or follow-up steering mode, the switch should be performed only after the operator confirms the cancellation.

Images