Automatic steering systems for ships

Abstract

This technology provides a way to reduce the deterioration of stability caused by nominal values ​​of hull parameters. [Solution] One embodiment is a ship's automatic steering system 1 that controls the heading of a ship based on a reference heading and a heading, comprising: an identification calculation unit 121 that calculates identification values ​​of hull parameters of a hull model relating to the ship's hull motion based on the ship's course change response; an update calculation unit 122 that calculates updated values ​​of hull parameters based on the calculated identification values; a feedback controller 15 that has a state estimator and state feedback and causes the heading to follow the reference heading based on the nominal values ​​of the hull parameters; and a replacement determination unit 123 that determines whether or not to replace the nominal values ​​with the calculated updated values ​​and set them as new nominal values ​​based on an estimated heading error, which is a state quantity of the state estimator when the ship changes course, and a prediction deviation, which is the deviation between the calculated updated values ​​and the prediction error calculated based on the reference heading when the ship changes course.

Description

[Technical Field]

[0001] This invention relates to a technique for identifying hull parameters in an identification model. [Background technology]

[0002] A ship's automatic steering system is a device that controls the rudder to make the heading from a gyrocompass follow a set course. Its control system calculates the deviation and turning angular velocity from the input of the set course and the ship's heading, multiplies them by a control gain, and outputs the command rudder angle, which is the control variable, to the steering gear. The steering gear moves the rudder to induce a turning angular velocity in the hull and change the heading. More specifically, the command rudder angle is calculated by adding the output of the feedback controller, which multiplies the deviation by a control gain, and the output of the feedforward controller. Identified hull parameters are input to the reference signal generation unit, which calculates the reference heading based on the trajectory plan from the set course, the feedback controller, and the feedforward controller, respectively, and these hull parameters are used in the calculations and controls of the reference signal generation unit, feedback controller, and feedforward controller.

[0003] Hull parameters are the parameters that constitute the hull motion model controlled by a ship's automatic steering system. Since these hull parameters are unknown in most cases, they are obtained through parameter identification. For example, in ships such as cargo ships and tankers, the draft of the hull changes due to loading and unloading of cargo, which alters the hull characteristics. Therefore, if control gains based on the hull parameters in the unloaded state are used for a hull loaded with cargo, the closed-loop stability of the steering system may decrease, potentially leading to a yawing phenomenon. To avoid such a situation, ship's automatic steering systems identify the hull parameters. In other words, ship's automatic steering systems improve the controllability of the hull by appropriately identifying these hull parameters.

[0004] As a technology related to such parameter identification, there is a known automatic steering system for ships comprising an identification calculation unit that calculates identified values ​​for hull parameters and an update calculation unit that calculates updated values ​​based on the identified values, wherein the identification calculation unit comprises an identification model including a hull model relating to the hull motion of the ship, a parameter adjustment unit that calculates identified values ​​by adjusting the hull parameters from the result of comparing the model output data with output data which are measured values ​​relating to the hull, and an average ship speed calculation unit that calculates the average value of the ship's surge speed as the average ship speed corresponding to the identified values, and the update calculation unit comprises an acquisition unit that acquires the identified values ​​and the average ship speed, an identification value storage unit that stores the acquired identified values ​​in a time series, and an calculation processing unit that calculates the coefficients of an update function which is a linear function relating to ship speed by the least squares method based on the stored identified values, and calculates updated values ​​for hull parameters using the update function which uses the calculated coefficients and the acquired average ship speed (see Patent Document 1). [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2020-032902 [Non-patent literature]

[0006] [Non-Patent Document 1] Shigeo Sugimoto, On reducing the number of personnel on naval vessels through the introduction of AI technology, Naval Staff College Strategic Studies, Vol. 9, No. 1, pp. 159-173, 2019. [Non-Patent Document 2] Fuyuki Hane, A Comprehensive Identification Method for Hull Motion Parameters, Transactions of the Japan Society of Naval Architects and Ocean Engineers, Vol. 20, pp. 27-38, Dec 2014. [Non-Patent Document 3] Fuyuki Hane, Design of a Course-Keeping System Using Analytical Methods Based on Course-Keeping Control, Transactions of the Japan Society of Naval Architects and Ocean Engineers, Vol. 23, pp. 33-44, Jun 2016. [Non-Patent Document 4] Fuyuki Hane, Design of a reference signal considering initial heading conditions and steering gear constraints: Application to ship course control, Transactions of the Society of Instrument and Control Engineers, Vol. 44, No. 4, pp. 333-342, 2008. [Non-Patent Document 5] TIFossen and T. Perez, Marine systems simulator, basic libraries and system examples for guidance, navigation and control, 2008. [Overview of the Initiative] [Problems that the invention aims to solve]

[0007] In the conventional automatic steering systems for ships described above, the hull parameter identification system estimates the identified values ​​of the hull parameters from the hull motion and estimates updated values ​​from the estimated identified values ​​(see Non-Patent Literature 2). The estimated updated values ​​become the nominal values ​​of the control system. In reality, the updated values ​​have a considerable amount of error. After calculating the setpoints and control gains of the control system from the nominal values ​​obtained by substituting the updated values, the control operation continues in real time. At this time, in the worst case, the automatic steering system for ships may generate a yawing phenomenon caused by the nominal values ​​with errors. Even if this does not occur, such nominal values ​​may worsen the ability to follow the target value and the damping ability.

[0008] This invention was made to solve the above-mentioned problems and aims to provide a technology that reduces the deterioration of stability caused by nominal values ​​of hull parameters. [Means for solving the problem]

[0009] One embodiment is an automatic steering system for ships that controls the heading of a ship based on a reference heading and a heading, comprising: an identification calculation unit that calculates identification values ​​of hull parameters of a hull model relating to the ship's hull motion based on the ship's course change response; an update calculation unit that calculates updated values ​​of the hull parameters based on the calculated identification values; a feedback controller having a state estimator and state feedback, which causes the heading to follow the reference heading based on the nominal values ​​of the hull parameters; and a replacement determination unit that determines whether or not to replace the nominal values ​​with the calculated updated values ​​to create new nominal values, based on an estimated heading error which is a state quantity of the state estimator when the ship changes course, and a prediction deviation which is the deviation between the calculated updated values ​​and the prediction error calculated based on the reference heading when the ship changes course. [Effects of the Invention]

[0010] According to the present invention, a technology can be provided to reduce the deterioration of stability caused by nominal values ​​of hull parameters. [Brief explanation of the drawing]

[0011] [Figure 1] This is a block diagram showing a control system including an automatic steering device for a ship according to an embodiment. [Figure 2] This figure shows the time series of the reference signal according to the embodiment. [Figure 3] This table shows the relationship between the turning conditions and the reference signal according to the embodiment. [Figure 4] This is a table showing the subscript notation according to the embodiment. [Figure 5] This figure shows the algorithmic structure of the prediction and judgment method according to the embodiment. [Figure 6] This figure shows the prediction error ratio according to the embodiment. [Figure 7] This table shows the parameters of the target ship model in the simulation. [Figure 8] This table shows the settings for actual values ​​in the simulation. [Figure 9]This table shows the settings for nominal values ​​in the simulation. [Figure 10] This figure shows the reference signal and reference angular velocity in the simulation. [Figure 11] This figure shows the azimuth error obtained from the simulation. [Figure 12] This table shows the initial values ​​of the nominal values ​​and the simulation results. [Figure 13] This figure shows the parameter identification history obtained through simulation. [Figure 14] This figure shows the simulation result No. 1. [Figure 15] This figure shows the simulation result No. 2. [Figure 16] This figure shows the simulation result No. 3. [Figure 17] This figure shows the simulation result No. 4. [Figure 18] This figure shows the percentage of prediction error obtained from the simulation. [Modes for carrying out the invention]

[0012] Embodiments of the present invention will be described below with reference to the drawings.

[0013] 1. Configuration of a ship's automatic steering system First, a control system including a ship's automatic steering system according to the present invention will be described. Figure 1 is a block diagram showing a control system including a ship's automatic steering system according to this embodiment.

[0014] As shown in Figure 1, the control system in this embodiment includes a ship's automatic steering system 1 and a ship 2 that is the object of its control. The ship 2 comprises a steering gear 21, a hull 22, and sensors 23. The ship's automatic steering system 1 comprises a reference signal generation unit 11, an identification calculation unit 121, an update calculation unit 122, a replacement determination unit 123, a subtractor 14, a feedback controller 15, a feedforward controller 16, and an adder 17.

[0015] The reference signal generation unit 11 calculates a reference azimuth ψ C , R based on the input set course ψ R from the trajectory plan. The identification calculation unit 121 identifies the hull parameters, which are the parameters constituting the hull model described later, and outputs them as the identified values of the hull parameters. The update calculation unit 122 calculates an updated value of the hull parameters based on a plurality of identified values identified by the identification calculation unit 121. The replacement determination unit 123 determines whether to replace the updated value of the hull parameters calculated by the update calculation unit 122 with the nominal value of the hull parameters. The new nominal value is output to the reference signal generation unit 11, the feedback controller 15, and the feedforward controller 16. The reference signal generation unit 11, the feedback controller 15, and the feedforward controller 16 perform respective calculations and controls using the nominal value of the hull parameters.

[0016] The subtractor 14 outputs a deviation e between the reference azimuth ψ R output from the reference signal generation unit 11 and the bow azimuth ψ of the hull 22. The feedback controller 15 multiplies the deviation e output from the subtractor 14 by a control gain to output a feedback rudder angle δ FB . The feedforward controller 16 outputs a feedforward rudder angle δ R based on the reference azimuth ψ FF output by the reference signal generation unit 11. The adder 17 adds the feedback rudder angle δ FB output by the feedback controller 15 and the feedforward rudder angle δ FF output by the feedforward controller 16, and outputs a command rudder angle δ C to the steering machine 21.

[0017] Furthermore, the sensors 23 of the vessel 2 include a speed log for detecting the surge velocity (hereinafter referred to as ship speed) u of the hull 22, a gyrocompass for detecting the heading ψ of the hull 22, and a GNSS sensor for detecting the ship's position (x,y) from a satellite positioning system (GNSS) such as GPS. The identification calculation unit 121 is input with the ship speed u, heading ψ, and ship's position (x,y) detected by the sensors 23, and the update calculation unit 122 is input with the ship speed u and ship's position (x,y) detected by the sensors 23. Here, the ship's position (x,y) is the ship's position in a local horizontal coordinate system where the X axis is north and the Y axis is east.

[0018] Thus, the control system according to this embodiment employs a closed-loop control system with a track-following method in order to make the ship's heading follow the reference heading. In the ship's automatic steering system 1, a course change control system, a holding control system, and a parameter identification system are configured to control the azimuth motion of the hull 22, using nominal values ​​of hull parameters.

[0019] 1.1 Vehicle Steering Control System This section describes the course control system in the automatic steering system for ships according to this embodiment. Figure 2 is a diagram showing the time series of the reference signal according to this embodiment. Figure 3 is a table showing the relationship between the turning conditions and the reference signal according to this embodiment.

[0020] The course control system uses a reference signal generation unit 11 and a feedforward controller 16 to improve the tracking performance of the course change response. The reference heading is input from the reference signal generation unit 11, and the feedforward rudder angle is input from the feedforward controller 16, respectively, to the closed-loop control system, so that the ship's heading appropriately follows the reference heading. In this embodiment, as shown in Figure 2, the reference heading ψ R and reference rudder angle δ R These together are called the reference signal. The rotation conditions for the reference signal are as shown in Figure 3. In addition to those shown in Figure 3, the rotation conditions also include ψ · R (0), ψ ·· R (0) or δFF Initial values ​​such as (0) are also taken into consideration.

[0021] 1.2 Holding Control System This embodiment describes the holding control system in a ship's automatic steering system.

[0022] The holding control system uses a feedback controller 15 to ensure closed-loop stability and disturbance rejection. The feedback controller 15 consists of a state estimator and state feedback. The damping coefficient of the characteristic polynomial of the closed-loop system is used as a measure of closed-loop stability.

[0023] 1.3 Parameter Identification System This section describes the parameter identification system in the automatic steering system for ships according to this embodiment. Figure 4 is a table showing the subscript notation according to this embodiment.

[0024] The parameter identification system provides adaptive functionality to the control system and outputs hull parameters, which are parameters of the hull model, and wave parameters, which are parameters of the wave model, as nominal values. In the control system, these nominal values ​​are used to calculate setpoints and control gains. Hull parameters are calculated based on time-series data output to sensors 23 during turning conditions and course changes. The parameter identification system uses an identification calculation unit 121, an update calculation unit 122, and a replacement determination unit 123.

[0025] The identification calculation unit 121 estimates the identification values ​​of the hull parameters from the course change motion. The update calculation unit 122 estimates the updated values ​​of the hull parameters using robust regression based on the identification values ​​of the multiple hull parameters estimated by the identification calculation unit 121. The replacement determination unit 123 determines whether or not to replace the current nominal values ​​in the control system with the updated values ​​of the hull parameters estimated by the update calculation unit 122 as new nominal values.

[0026] The substitution determination unit 123 determines whether or not to substitute the nominal value based on the prediction determination method described below. The subscript notation used in the following explanation is as shown in Figure 4.

[0027] 2. Principles of Predictive Judgment Methods I will explain the principles of the predictive decision method.

[0028] The principle of the prediction and judgment method is based on the following:

[0029] • The updated values ​​are estimated by identifying the actual hull parameters. • The updated values ​​are set in the hull parameters, and the conditions for the actual course-changing response are set in the control system, making the course-changing response predictable. By comparing the actual response (the actual change in direction) with the predicted response (the predicted change in direction), the frequency of adaptability of the updated value can be evaluated. Based on this frequency, it is possible to decide whether or not to replace the current nominal value with the updated value as the new nominal value.

[0030] 2.1 Transmission characteristics of needle change response The closed-loop transfer function during course change response is derived. In this case, the hull parameters of the controlled vessel become actual values, while the hull parameters of the control system become nominal values.

[0031] The transfer function of the orbit-following control system is determined based on the configuration shown in Figure 1.

[0032]

number

[0033] From the above equation, closed-loop stability and disturbance rejection are given by C FB (s) is dependent. The reference direction does not affect their performance, but it does affect the course change response.

[0034] The reference signal is a function of time, and the reference direction is ψ. R and reference rudder angle δ R It consists of the following: Feedforward rudder angle δ FF The reference rudder angle is δ. R This corresponds to the first-order lag output. As shown in Figure 2, the time range including the approach time and the change of direction time is used in the identification calculation. In the prediction of the change of direction response, the change of direction time 0 ≤ t ≤ t t This is used.

[0035] Feedforward rudder angle Δ FF (s)

[0036]

number

[0037]

number

[0038]

number

[0039] 2.2 Error characteristics of needle change response In this embodiment, the parameter error model is derived from equation (1).

[0040] 2.2.1 Hull Model The hull model of the azimuth motion that is the target of control

[0041]

number

[0042] 2.2.2 Control System The control system is modified by omitting the reference signal generation unit 11 and using a feedforward controller C FF (s) and feedback controller C FB It consists of (s). At this time,

[0043]

number

[0044]

number

[0045] 2.2.3 Closed-loop transfer characteristics Determine the closed-loop transmission characteristics during a change of direction. In equation (1), remove the disturbance term and consider the reference direction Ψ. R We determine the transfer characteristics for (s). Substituting equations (5), (6), and (7) into equation (1),

[0046]

number

[0047]

number

[0048] The above equation consists of the nominal and actual values ​​of the hull parameters. Parameter uncertainty is used to establish the relationship between the two.

[0049] If we define the nominal value from the actual value, and conversely express the actual value using the nominal value,

[0050]

number

[0051] Substituting the actual values ​​from the above equation into equation (10), the transfer function is expressed in terms of nominal values.

[0052]

number

[0053]

number

[0054] 2.2.4 Transmission characteristics of azimuth error The azimuth error is obtained by inputting the reference azimuth into the above transfer function. The azimuth error characteristics are shown from the numerator of the transfer function. The effect of the denominator is ignored. The effect of parameter uncertainty is given by equation (14) as follows.

[0055] α is s0 It acts on the term, and β is s 1 and s 0 It acts on the term. ·ψ e The sign of is influenced by α and β, and for example, it becomes as follows:

[0056]

number

[0057] 2.3 Overview of the Prediction and Judgment Method As explained in "2.2 Error Characteristics of Course Change Response," the error model for course change response can be constructed from the actual and nominal values ​​of the hull parameters. The heading error during a course change can be obtained by providing a reference heading. Therefore, the heading error can be predicted by replacing the actual value with the updated value.

[0058] The updated values ​​are output from time-series data of the actual direction change response through identification and estimation calculations. Therefore, if the parameter errors included in the updated values ​​are small, the predicted direction error (prediction error) will closely match the actual direction error (actual error). Conversely, if the parameter errors are large, the prediction error will not match the actual error.

[0059] Therefore, the prediction method involves the following when the course changes.

[0060] (1) Set the time-series data of the actual error, the updated value, the nominal value of the control system, and the conditions for changing the reference direction. (2) Time series data of the prediction error is calculated using the nominal values ​​obtained by substituting the updated values ​​with the hull parameters of the controlled vessel, and the course change conditions. (3) Compare the time-series data of the actual error and the predicted error to determine whether or not to replace the updated value with the nominal value. Specifically, if the difference between the predicted error and the actual error is within a predetermined threshold, replace the updated value with the nominal value; otherwise, do not replace it.

[0061] 3. Operation of the substitution determination unit The operation of the replacement determination unit of the automatic steering system for ships according to this embodiment will now be described. As shown in Figure 5, the operation of the replacement determination unit in the prediction determination method can be broadly divided into three operational elements: (a) acquisition of actual error data, (b) generation of prediction error data, and (c) replacement determination.

[0062] 3.1 Acquisition of actual error data

[0063] As an operating element (a), the substitution determination unit 123 determines the actual heading error ψ from the needle change response. e a (t) is the time-series data of the control system C(s) and the estimated azimuth error (actual error) ψ^ e a Extract (t). Here, the subscript a represents the actual value and t is time. The control system C(s) is a feedforward controller C FF (s) and feedback controller C FB It consists of (s). ψ^ e a (t) is a feedback controller C FB This is the state variable of the state estimator of (s).

[0064] Furthermore, the substitution determination unit 123 extracts the course change conditions for generating the reference heading, the nominal and updated values ​​of the hull parameters, and the set values ​​of the control gain. These are reproduction parameters necessary to simulate the course change response.

[0065] 3.1.1 Selection of azimuth error signal The substitution determination unit 123 uses the actual azimuth error ψ as the azimuth error signal. e a (t) is not the estimated azimuth error ψ^ e a Select (t). Azimuth deviation ψ e As shown in Figure 1,

[0066]

number

[0067] 3.2 Generation of prediction error data As an operating element (b), the substitution determination unit 123 simulates the predicted direction change response and estimates the direction error (prediction error) ψ^ e Generate time series data for (t).

[0068] Based on the reproduction parameters of the operating element (a), the updated values ​​are replaced with the hull parameters of the controlled vessel, the control system for course changes is set and the reference heading is input, and by numerical calculation of equation (1), ψ^ e (t) is obtained. Here, the reference direction is ψ^ in the operating element (a). e a It is either extracted in sync with (t) or generated based on the needle change conditions.

[0069] For simulation calculations of the course change response, please refer to Non-Patent Document 4 for reference direction and Non-Patent Document 3 for closed-loop control systems.

[0070] 3.3 Replacement judgment As an operational element (c), the replacement determination unit 123 determines whether or not to replace the current nominal value with the updated value as the new nominal value. Here, the replacement determination unit 123 calculates the predicted error ratio from the actual error and the predicted error, and makes a determination based on a comparison of this predicted error ratio with a preset threshold.

[0071] Prediction error ψ^ e Actual error of (t) ψ^ e a Prediction deviation Δψ^ for (t) e If (t) is determined from Figure 6,

[0072]

number

[0073] Δψ^ e (t) is ψ^ e a Since it is affected by the amplitude of (t), we make it dimensionless. Therefore, the prediction error ratio (called the adaptive frequency) is

[0074]

number

[0075] The substitution determination unit 123 determines whether the absolute value of the adaptability frequency is less than or equal to the threshold (OK) or whether the absolute value of the adaptability frequency is greater than the threshold (NG), as shown in the following equation.

[0076]

number

[0077]

number

[0078] If the determination result is OK, the replacement determination unit 123 sets Par^ to the updated value of the hull parameter set, Par n If we consider the nominal value of the hull parameter set, then the nominal value of the hull parameter set

[0079]

number

[0080] In this way, when the absolute value of the adaptability frequency is less than or equal to the threshold value, that is, within the preset range, the replacement determination unit 123 replaces the current nominal value with the updated value and sets the updated value as the new nominal value. Otherwise, the current nominal value is retained.

[0081]

Number

[0082] 4 Verification Verify the above prediction judgment method by simulation.

[0083] In the simulation, the actual value of the target ship model is fixed, and the nominal value of the hull model of the control system is changed. The parameter of the control system is K p = 1, ζ n = 1 / √2.

[0084] The target hull model is a mariner class with the non-linear characteristics described in Non-Patent Document 5, and the ship length L = 160.93 m. The rudder angle offsets are respectively δ ro = δ vo = -1 deg. The parameters of the hull model of this ship are summarized in Fig. 7. In Fig. 7, U0 is the initial ship speed (speed), K v is the cross-flow gain used in course control, and the numerical value is the value obtained by repeating the tacking and identification calculation and the update algorithm 5 times using the identification algorithm (see Non-Patent Document 2). The initial ship speed is U0 = 15 kn, and the tacking amount is 15 deg.

[0085] When the actual values of the parameters of the target hull model are set as shown in Fig. 8, ω a = 0.0488 rad / s, and the natural period per a = 128.8 s.

[0086] 4.1 Principle of the prediction judgment method Verify the validity of the principle described in "2 Principle of the prediction judgment method".

[0087] 4.1.1 Parameters of the control system The nominal values of the hull model of the control system are set as follows using the coefficients α and β of parameter uncertainty in Fig. 9. However, they are changed one by one.

[0088]

Equation

[0089] 4.1.2 Results of numerical calculation The reference signal is defined as shown in Fig. 10. The reference angular velocity is approximated by a trapezoid (upper base 0 deg / s), and the reference azimuth is the integral of the reference angular velocity (the amount of course change is 12 deg).

[0090] The azimuth error is obtained from Equation (10) using the reference azimuth as shown in Fig. 11. Therefore, the validity of this principle could be confirmed from the following results.

[0091] · In the transfer characteristics of the azimuth error, the sign became the same as that explained in "2.2.4 Transfer characteristics of azimuth error". · In terms of the azimuth error, compared with Fig. 18 which is the result of detailed numerical calculation, the peak of the azimuth error became qualitatively the same.

[0092] 4.2 Operation of the replacement determination unit Verify the effectiveness of the operation described in "3 Operation of the replacement determination unit" through simulation.

[0093] 4.2.1 Conditions · As the control object of the algorithm configuration of the prediction judgment method shown in Fig. 5, the above mariner class is adopted for the hull motion model of the actual system, and the following equation is adopted for the hull motion model of the simulation system.

[0094]

Equation

[0095] In the control system C(s) in Figure 5, an estimator and state feedback configuration (see Non-Patent Document 3) is used to obtain an estimated azimuth error ψ^ from the estimator. e a and ψ^ e Take it out. • In Figure 5, the reference direction ΨR(s) is generated from nominal values ​​using Non-Patent Document 4.

[0096] 4.2.2 Method (1) In the operating elements (a) and (b), respectively ψ^ e a and ψ^ e This is determined by numerical calculation from the configuration shown in Figure 1. Note that the hull motion models differ between the two. (2) In the operating element (c), ψ^ e a and ψ^ e Based on the input and "3.3 Replacement Judgment," a decision is made to replace the updated value with the nominal value. The set value ψ is then determined from the result. set ′ to decide. (3) Numerical calculations are performed twice, including the calculations for parameter identification and updating.

[0097] 4.2.3 Results The initial values ​​of the nominal values ​​and the simulation results are summarized in Figure 12. Figure 12 shows the combinations of α and β and the corresponding simulation results No. 1 to No. 4 for each combination. Figures 14 to 17 show the simulation results No. 1 to No. 4, respectively.

[0098] 4.2.3.1 Identification History Figure 13 shows the parameter identification history for the nominal value. As shown in Figure 13, the nominal value almost reaches the value in Figure 7 after one identification update. The following simulation results use the value from the first identification update.

[0099] 4.2.3.2 Response Results The display method for Figures 14 to 17 is as follows.

[0100] • The graph in the upper section shows the reference direction ψ R The heading ψ is shown, and the graph below shows the heading error ψ. e Its estimated value ψ^ e This is shown. The graph on the left shows the actual system's shift response, the graph in the center shows the simulated system's shift response, and the graph on the right shows the actual system's shift response with updated values ​​replaced by nominal values.

[0101] The results from Figures 14 to 17 can be summarized as follows:

[0102] (1) In all figures, the pitch change responses of the actual system and the simulated system show almost the same trend. Therefore, this updated value is valid. From the bottom, ψ e =ψ^ e This can be confirmed. (2) Regarding the graph on the right, some transient phenomena occur in Figure 16, but these are reduced in the other figures. The initial conditions in Figure 16 can be said to be the most severe. (3) It is acceptable to replace all updated values ​​for the initial parameter values ​​in Figure 18 with nominal values ​​for evaluation.

[0103] 4.2.3.3 Prediction and Judgment The prediction error ratio Δψ^ explained in "3.3 Substitution Judgment" e ′ is shown in Figure 18. In each of the graphs shown in Figure 18, Δψ^ e The ' markings are on the left side. In Figure 18, the maximum peak error occurs at No. 3, which is slightly less than 0.25.

[0104] Δψ^ e The peak value of ' tends to be proportional to the peak value of the actual system when the nominal value of the above updated value is replaced. Therefore, the set value of the adaptability frequency

[0105]

number

[0106] 5. Conclusion This specification examines a method for predicting the course change response based on updated values ​​and replacing the updated values ​​with nominal values ​​from the adaptive frequency, and verifies its effectiveness through simulation.

[0107] The results can be summarized as follows:

[0108] • The principle of the predictive judgment method was demonstrated. - The transmission characteristics and error characteristics of the needle change response were clarified. - The proposed method for this prediction and decision-making method is presented. The algorithm for the prediction and judgment method was constructed from three operating elements (a), (b), and (c). -(a): Acquisition of actual error data, (b): Generation of predicted error data, and (c): Substitution determination were clarified. • Through simulation verification, the validity of the principle and the effectiveness of the algorithm were confirmed.

[0109] The embodiments of the present invention are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be carried out in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims and their equivalents. [Explanation of Symbols]

[0110] 1. Automatic steering system for ships 15. Feedback Controller 121 Identification Calculation Unit 122 Update calculation section 123 Replacement judgment part 22 hull

Claims

1. An automatic steering system for ships that controls the heading of a ship based on a reference bearing and a heading, An identification calculation unit that calculates identification values ​​of hull parameters of a hull model relating to the hull motion of the ship based on the ship's course change response, An update calculation unit that calculates updated values ​​for the hull parameters based on the calculated identification values, A feedback controller having a state estimator and state feedback, which causes the ship's heading to follow the reference heading based on the nominal values ​​of the hull parameters, A replacement determination unit determines whether or not to replace the nominal value with the calculated updated value to create a new nominal value, based on the estimated heading error, which is a state quantity of the state estimator when the ship changes course, and the predicted deviation, which is the difference between the calculated updated value and the predicted error calculated based on the reference heading when the ship changes course. A ship's automatic steering system equipped with the following features.

2. The automatic steering device for ships according to claim 1, characterized in that, if the ratio of the predicted deviation is within a preset range, the replacement determination unit replaces the nominal value with the calculated updated value to obtain a new nominal value.

3. The automatic steering device for ships according to claim 1, further characterized in that the substitution determination unit calculates the prediction error based on the hull model and the transfer function of the feedforward controller.

Images