Motion state observation method for autonomous navigation of ship

Abstract

A motion state observation method for autonomous navigation of a ship. The method uses a determination method for system matrixes Ad, Bd, Ed and Hd, and a determination method for a related measurement covariance matrix R and noise covariance matrix Q, and uses a discrete Kalman filtering algorithm to perform motion state observation and prediction for autonomous navigation of a smart ship. The method fully uses ship position and heading measurement information of a smart ship during autonomous navigation, uses a ship motion model and the discrete Kalman filtering algorithm, and selects an effective state observation variable, thus calculating reasonable and effective observer parameters. The state observation method can effectively calculate information such as the actual position, heading, speed and angular velocity of a smart ship at each time instant during autonomous navigation, and predict motion information at a next time instant, thus helping to reduce the difficulty of smart ship motion controller design, ensuring effective implementation of navigation control strategies, reducing ineffective thrust output and rotation of propulsion systems, and improving the autonomous navigation capability of smart ships.

Description

A method for observing the motion state of a ship's autonomous navigation Technical Field

[0001] The present invention relates to an autonomous navigation motion state observation method based on discrete Kalman filtering, and in particular to an intelligent ship that is autonomously navigating in the ocean. The method uses measurement information obtained by a sensor system to effectively observe the motion state of the ship and estimate the actual position, heading, speed, angular velocity and other information of the intelligent ship during autonomous navigation. Background Art

[0002] Shipping, a high-volume, low-cost mode of transport that accounts for over 80% of international trade, plays a crucial role among the numerous modes of transport used in world trade. With the rapid development of economic globalization and the continuous rise of science and technology, the maritime industry has also developed rapidly, and the degree of automation of ships has also been increasing. Smart ships have become a focus of attention in the shipping sector.

[0003] When intelligent ships are moving in the ocean, disturbances in the surrounding marine environment will significantly affect the ship's motion and control. Furthermore, these disturbances are highly volatile and random, which can interfere with the study of the ship's motion and response characteristics. The main environmental factors affecting ship motion are wind, waves, and currents. The first-order wave force generated by waves causes the ship to experience high-frequency motion, a reciprocating motion with a mean of zero, which the ship should not resist. If the ship resists this high-frequency motion, it will cause unnecessary control, causing the ship's propulsion system to continuously respond to it. This will affect the ship's trajectory and navigation strategy during autonomous navigation, accelerate wear and aging of the propulsion system, and consume excessive energy.

[0004] During autonomous navigation, the introduction of effective state observation methods can filter out high-frequency motion signals from the acquired information, retain low-frequency signals, estimate the ship's actual low-frequency position, and thus calculate the Gaussian white noise in the measurement system. This allows the intelligent ship to respond only to low-frequency motion disturbances, effectively improving the system's control capabilities, ensuring the smoothness of the ship's motion trajectory during autonomous navigation, effectively implementing navigation strategies, reducing unnecessary thrust output and propeller rotation, and improving the intelligent ship's autonomous navigation capabilities.

[0005] Therefore, it is necessary to design an autonomous navigation motion state observation method based on discrete Kalman filtering, make full use of the measurement information of the ship position and heading during the autonomous navigation of the intelligent ship, combine the high-frequency and low-frequency motion models of the ship, the first-order Markov model, the observation model, and the discrete Kalman filtering algorithm, select effective state observation variables, perform discrete optimization processing on the relevant system matrix equations, design reasonable and effective observer parameters, and obtain the low-frequency motion information of the ship at the current moment through effective calculation. At the same time, the low-frequency motion information at the next moment is predicted and calculated and corrected at the next moment, which is beneficial to the speed and trajectory maintenance of the intelligent ship during autonomous navigation, so that the navigation control strategy can be effectively implemented and the unnecessary thrust output and rotation of the propulsion system can be reduced.

[0006] Summary of the Invention

[0007] The purpose of the present invention is to provide a motion state observation method for an intelligent ship during autonomous navigation, so as to accurately observe the actual position, heading, speed, angular velocity and other information of the intelligent ship during autonomous navigation.

[0008] To achieve the above purpose, the present invention adopts the following technical solutions: a method for observing the motion state of a ship's autonomous navigation, using the system matrix A d 、B d 、E d 、H d The determination method of , the determination method of the related measurement variance matrix R, and the noise covariance matrix Q; the discrete Kalman filtering algorithm is used to observe and predict the autonomous navigation motion state of intelligent ships.

[0009] Furthermore, the system matrix A d 、B d 、E d 、H d Determination method: According to the wave intensity parameter σ, damping coefficient λ, dominant wave frequency ω0, mass matrix M, damping coefficient matrix D, time constant diagonal matrix T, thruster vector arrangement matrix B u To determine the system matrix A d 、B d 、E d and H d .

[0010] Furthermore, the method for determining the correlation measurement variance matrix R and the noise covariance matrix Q is as follows: the covariance is calculated based on the samples of the longitudinal position information, the transverse position information, and the heading information measured by the intelligent ship when it is not moving in the port. To estimate the matrix R; combine the high-frequency motion model of intelligent ships and the data during autonomous navigation to calculate the matrix Q1; the noise estimation matrix Q2 of position and heading measurement; the motion deviation estimation matrix Q3 caused by uncertain environmental forces; and form the matrix Q.

[0011] Furthermore, the discrete Kalman filter algorithm is used to observe and predict the autonomous navigation motion state of the intelligent ship. The discrete Kalman filter gain iteration formula is: K(h)=P - (h)*H d T (h)*(H d (h)*P - (h)*H d T (h)+R) -1 ; The iterative formula for motion state observation is x(h)=x - (h)+K(h)*(y(h)-H d *x - (h)), and the north position in the state observation variable x(h) is the 7th element, the east position is the 8th element, the heading angle is the 9th element, the longitudinal speed is the 10th element, the lateral speed is the 11th element, and the bow angular velocity is the 12th element; the iterative formula for updating the matrix is ​​p(h) = (I 15×15 -K(h)*H d )*P - (h)*(I 15×15 -K(h)*H d ) -1 +K(h)*R*K(h) T ; The iterative formula of the next moment state observation variable is x(h+1) - =A d *x(h)+B d *u; at the same time, the state observation variable x(h+1) - The north position is the 7th element, the east position is the 8th element, the heading angle is the 9th element, the surge velocity is the 10th element, the sway velocity is the 11th element, and the bow angular velocity is the 12th element. The iterative formula for updating the matrix at the next moment is predicted to be p(h+1) - =A d *p(h)*A d T +E d *Q*E d T .

[0012] Furthermore, the method for observing the autonomous navigation motion state of a ship comprises the following specific steps:

[0013] In the first step, during autonomous navigation, the intelligent ship only considers the three degrees of freedom on the water surface: surge, sway, and pitch. The position and heading in the northeast coordinate system are selected as the variable η, η = [NE ψ] T (1)

[0014] Where, N is the north position information; E is the east position information; ψ is the heading information;

[0015] Select the velocity in the hull coordinate system as vector vv=[uvr] T (2)

[0016] Where u is the velocity in the longitudinal direction; v is the velocity in the transverse direction; and r is the angular velocity in the yaw direction.

[0017] The conversion relationship between the two is as follows:

[0018] Where: R(ψ) is the transformation matrix.

[0019] In the second step, the slowly varying disturbance motion generated by the environmental forces during autonomous navigation is estimated using a first-order Gauss-Markov model, namely:

[0020] Where b∈R 3 is the deviation between force and torque; T∈R 3×3 is the diagonal matrix of time constants; E b ∈R 3×3 is a diagonal matrix; it indicates the amplitude of the environmental disturbance; ω b ∈R 3 is the Gaussian white noise matrix;

[0021] The third step is to establish a nonlinear low-frequency motion model of intelligent ships.

[0022] Where M is the ship mass matrix including the additional mass; D is the ship damping coefficient matrix; τ is the resultant torque output by the propulsion system in the three degrees of freedom direction,

[0023] Where M is represented by:

[0024] Where m is the mass of the smart ship; x G is the position vector in the x direction in the center of gravity hull coordinate system; I z is the moment of inertia of the smart ship; and are the additional mass of the hydrodynamic force; The additional mass generated by the mutual coupling of longitudinal surge and bow motion;

[0025] D is represented by:

[0026] Where, X u 、Y v 、Y r 、N v and N r are the coefficients of linear damping of the intelligent ship in three degrees of freedom,

[0027] τ is expressed as: τ = B u T F =[F x F y M z ] T (9)

[0028] Where B u is the thruster vector arrangement matrix; T F Output thrust for the thruster;

[0029] The fourth step is to calculate the high-frequency motion caused by the first-order waves during the autonomous navigation of the intelligent ship. The high-frequency motion in the surge, sway and heading directions is expressed as a separate motion in a single degree of freedom:

[0030] Expand the above equation into:

[0031] Where x ω1 and x ω2 is the motion generated by the first-order and second-order wave forces acting on the smart ship; ω0 is the dominant wave frequency in the wave PM spectrum; λ is the relative damping coefficient; σ is a constant value related to the wave intensity; η ω It is the low-frequency motion caused by the second-order wave force;

[0032] Step 5: Based on the above equation, select the following intelligent ship motion equation:

[0033] Where, E v Diagonal matrix; ω v is Gaussian white noise; at the same time, select x=[ξ T ,η,v,b T ] T ∈R 15 is the state observation variable;

[0034] Step 6: Based on the position and heading information provided by the sensor system, design the following initial state linear observation equation: y=H dx+υ (16)

[0035] Where: A d is a 15×15 order state transfer matrix; B d is the 15×p-order control gain input matrix; u is the p-order control force input vector; E d is a 15×15 order noise input matrix; ω is a 15-order Gaussian white noise; y is a 3D measurement vector; H d is a 15×1 measurement matrix; v is a 15-order measurement Gaussian white noise; ω and v are independent white noises with a mean of 0;

[0036] Thus, the discrete system matrix A can be obtained d 、B d 、E d and H d for: H d =[0 3×3 I 3×3 I 3×3 0 3×3 0 3×3 ] (20)

[0037] A in the system matrix ω and E ω The settings are as follows:

[0038] Step 7: Set the measurement variance matrix R = R 3×3 , the matrix R can use the following matrix:

[0039] Where, The covariance of the measurement noise of the surge position information, sway position information, and heading information can be calculated by the covariance of the data samples measured when the smart ship is not moving in the port;

[0040] Step 8: Set the state noise covariance matrix Q = R 9×9 , the diagonal line is selected as the positive adjustable parameter, that is:

[0041] Matrix Q1 represents the noise of linear wave motion, which is calculated and set by combining the high-frequency motion model of intelligent ships and data from autonomous navigation. Matrix Q2 represents the noise of position and heading measurements during low-frequency motion. Matrix Q3 represents the motion deviation caused by uncertain forces caused by various environmental forces.

[0042] Step 9: Select the estimated position and heading of the starting point of autonomous navigation of the intelligent ship and set the initial value x - (h)[0]; Based on the true initial value x[0] and the estimated initial value x - (h)[0] variance calculation updates the initial value P of the matrix - (h)[0], or directly use the unit matrix to converge quickly, that is: P - (h)[0]=I 15×15 ;

[0043] The tenth step is to use the discrete Kalman filter algorithm of the following formulas (25) to (29) based on the initial value and the position and heading information y(h) provided by the sensor system to obtain the position, heading, speed, and angular velocity information of the ship at the current moment during autonomous navigation, namely, the 7th, 8th, 9th, 10th, 11th, and 12th elements in x(h); and predict the position, heading, speed, and angular velocity information of the ship at the next moment, namely, x(h+1) - The 7th, 8th, 9th, 10th, 11th, and 12th elements in K(h); K(h)=P - (h)*H d T (h)*(H d (h)*P - (h)*H d T (h)+R) -1 (25) x(h)=x - (h)+K(h)*(y(h)-H d *x - (h)) (26) p(h)=(I 15×15 -K(h)*H d )*P - (h)*(I 15×15 -K(h)*H d ) -1 +K(h)*R*K(h) T (27) x(h+1) - =A d *x(h)+B d *u (28) p(h+1) - =A d *p(h)*A d T +E d *Q*E d T (29)

[0044] In the eleventh step, the obtained real position, heading, speed, and angular velocity are input into the controller as information for actual movement of the intelligent ship's autonomous navigation for control calculation.

[0045] The beneficial effects of the present invention are:

[0046] The present invention provides a method for observing the motion state of a ship during autonomous navigation. This method fully utilizes the measured information of the ship's position and heading during the autonomous navigation process of an intelligent ship. By combining a ship motion model with a discrete Kalman filter algorithm, it selects effective state observation variables and calculates reasonable and effective observer parameters. This state observation method can effectively calculate the actual position, heading, velocity, angular velocity, and other information of the intelligent ship at each moment during the autonomous navigation process, and predict the motion information at the next moment. This method helps reduce the design difficulty of the intelligent ship's motion controller, enables the effective implementation of navigation control strategies, reduces ineffective thrust output and rotation of the propulsion system, and improves the autonomous navigation capability of the intelligent ship.

[0047] The method of the present invention provides an autonomous navigation motion state observation method based on discrete Kalman filtering, which can effectively improve the control capability of the system, ensure the smoothness of the ship's motion trajectory during autonomous navigation, and effectively implement the navigation strategy. It has important theoretical significance and engineering value for the autonomous navigation of intelligent ships. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] FIG1 is a schematic diagram of the principle of the autonomous navigation motion state observation method based on discrete Kalman filtering of the present invention. DETAILED DESCRIPTION

[0049] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0050] As shown in FIG1 , the present invention provides an autonomous navigation motion state observation method based on discrete Kalman filtering, and the steps are as follows:

[0051] In the first step, during autonomous navigation, the intelligent ship only considers the motion of the three degrees of freedom (surge, sway, and pitch) on the water surface. The position and heading in the northeast coordinate system are selected as the variable η, η = [NE ψ] T (1)

[0052] Where N is the north position information; E is the east position information; ψ is the heading information.

[0053] Select the velocity in the hull coordinate system as vector vv=[uvr] T (2)

[0054] Where u is the velocity in the longitudinal direction; v is the velocity in the transverse direction; and r is the angular velocity in the yaw direction.

[0055] The conversion relationship between the two is as follows:

[0056] Where: R(ψ) is the transformation matrix.

[0057] In the second step, the slowly varying disturbance motion generated by the environmental forces during autonomous navigation is estimated using a first-order Gauss-Markov model, namely:

[0058] Where b∈R 3 is the deviation between force and torque; T∈R 3×3 is the diagonal matrix of time constants; E b ∈R 3×3 is a diagonal matrix; it indicates the amplitude of the environmental disturbance; ω b ∈R 3 is the Gaussian white noise matrix.

[0059] The third step is to establish a nonlinear low-frequency motion model of the intelligent ship.

[0060] Where M is the ship mass matrix including the additional mass; D is the ship damping coefficient matrix; τ is the resultant force (moment) output by the propulsion system in the three degrees of freedom directions.

[0061] Where M is represented by:

[0062] Where m is the mass of the smart ship; x G is the position vector in the x direction in the center of gravity hull coordinate system; I z is the moment of inertia of the smart ship; and are the additional mass of the hydrodynamic force; It is the additional mass generated by the mutual coupling of longitudinal surge and bow motion.

[0063] D is represented by:

[0064] Where, X u 、Y v 、Y r 、N v and N r are the coefficients of linear damping of the intelligent ship in three degrees of freedom.

[0065] τ is expressed as:

[0066] τ=B u T F=[F x F y M z ] T (9)

[0067] Where B u is the thruster vector arrangement matrix; T F Output thrust for the thrusters.

[0068] The fourth step is to calculate the high-frequency motion caused by the first-order waves during the autonomous navigation of the intelligent ship. The high-frequency motion in the surge, sway and heading directions is expressed as a separate motion in a single degree of freedom:

[0069] Expand the above equation into:

[0070] Where x ω1 and x ω2 is the motion generated by the first-order and second-order wave forces acting on the smart ship; ω0 is the dominant wave frequency in the wave PM spectrum; λ is the relative damping coefficient; σ is a constant value related to the wave intensity; η ω It is the low-frequency motion caused by the second-order wave force.

[0071] Step 5. Based on the above equation, select the following intelligent ship motion equation.

[0072] Where, E v Diagonal matrix; ω v is Gaussian white noise.

[0073] At the same time, select x=[ξ T ,η,v,b T ] T ∈R 15 is the state observation variable.

[0074] Step 6: Based on the position and heading information provided by the sensor system, design the following initial state linear observation equation: y=H d x+υ (16)

[0075] Where: A d is a 15×15 order state transfer matrix; B d is the 15×p-order control gain input matrix; u is the p-order control force input vector; E d is a 15×15 order noise input matrix; ω is a 15-order Gaussian white noise; y is a 3D measurement vector; H dis a 15×1 measurement matrix; v is a 15-order measurement Gaussian white noise; ω and v are independent white noises with a mean of 0.

[0076] Thus, the discrete system matrix A can be obtained d 、B d 、E d and H d for: H d =[0 3×3 I 3×3 I 3×3 0 3×3 0 3×3 ] (20)

[0077] A in the system matrix ω and E ω The settings are as follows:

[0078] Step 7: Set the measurement variance matrix R = R 3×3 , the matrix R can use the following matrix:

[0079] Where, The covariance representing the measurement noise of the surge position information, sway position information, and heading information can be calculated by the covariance of data samples measured when the smart ship is not moving in the port.

[0080] Step 8: Set the state noise covariance matrix Q = R 9×9 , the diagonal line is selected as the positive adjustable parameter, that is:

[0081] Matrix Q1 represents the noise of linear wave motion, which is calculated and set by combining the high-frequency motion model of intelligent ships and data from autonomous navigation; matrix Q2 represents the noise of position and heading measurements during low-frequency motion; and matrix Q3 represents the motion deviation caused by the uncertainty forces caused by various environmental forces.

[0082] Step 9: Select the estimated position and heading of the starting point of autonomous navigation of the intelligent ship and set the initial value x - (h)[0]; Based on the true initial value x[0] and the estimated initial value x - (h)[0] variance calculation updates the initial value P of the matrix - (h)[0], or directly use the unit matrix to converge quickly, that is: P - (h)[0]=I 15×15 .

[0083] The tenth step is to use the discrete Kalman filter algorithm of formula (25) to formula (29) according to the initial value and the position and heading information y(h) provided by the sensor system, as shown in Figure 1. The position, heading, speed, and angular velocity information of the ship at the current moment during autonomous navigation are obtained, namely, the 7th, 8th, 9th, 10th, 11th, and 12th elements in x(h); the position, heading, speed, and angular velocity information of the ship at the next moment are predicted, namely, x(h+1) - The 7th, 8th, 9th, 10th, 11th, and 12th elements in K(h). - (h)*H d T (h)*(H d (h)*P - (h)*H d T (h)+R) -1 (25) x(h)=x - (h)+K(h)*(y(h)-H d *x - (h)) (26) p(h)=(I 15×15 -K(h)*H d )*P - (h)*(I 15×15 -K(h)*H d ) -1 +K(h)*R*K(h) T (27) x(h+1) - =A d *x(h)+B d *u (28) p(h+1) - =A d *p(h)*A d T +E d *Q*E d T (29)

[0084] In the eleventh step, the obtained real position, heading, speed, and angular velocity are input into the controller as information for actual movement of the intelligent ship's autonomous navigation for control calculation.

[0085] Specific application examples of the present invention:

[0086] In the first step, according to the ship parameters of the smart ship, the ship mass matrix M containing the additional mass and the ship damping coefficient matrix D are obtained through the hydrodynamic software (CFD).

[0087] The second step is to calculate the covariance based on the samples of the longitudinal position information, transverse position information, and heading information measured by the smart ship when it is not moving in the port. Estimation matrix R;

[0088] The third step is to combine the high-frequency motion model of intelligent ships and the data during autonomous navigation to calculate the matrix Q1; the noise estimation matrix Q2 of position and heading measurement; the motion deviation estimation matrix Q3 caused by uncertain environmental forces; and the matrix Q.

[0089] The fourth step is to use the wave intensity parameter σ, damping coefficient λ, dominant wave frequency ω0, mass matrix M, damping coefficient matrix D, time constant diagonal matrix T, and thruster vector arrangement matrix B. u , determine the system matrix A according to formula (17) to formula (22) d 、B d 、E d and H d .

[0090] The fifth step is to set the initial value x according to the position of the starting point of autonomous navigation of the intelligent ship, the estimated heading value and the environmental information. - (h)[0]; Based on the true initial value x[0] and the estimated initial value x - (h)[0] variance calculation updates the initial value P of the matrix - (h)[0], or select the unit matrix P - (h)[0]=I 15×15 .

[0091] The sixth step is to measure the variance matrix R and the matrix H d , update matrix P - (h)[0] is substituted into formula (25) to calculate the discrete Kalman gain K(h).

[0092] Step 7: Substitute the discrete Kalman gain K(h) and matrix H d , initial value x - (h)[0], the position and heading information y(h) provided by the sensor system are substituted into formula (26) to calculate the current moment x(h), and the 7th element of the north position, the 8th element of the east position, the 9th element of the heading angle, the 10th element of the longitudinal velocity, the 11th element of the transverse velocity, and the 12th element of the bow angular velocity in x(h) are input into the controller as the information of the actual movement of the autonomous navigation of the intelligent ship for control calculation.

[0093] Step 8: Substitute the discrete Kalman gain K(h) and matrix H d , update matrix P -(h)[0] and the measurement variance matrix R are substituted into formula (27) to calculate the update matrix p(h) at the current moment.

[0094] The ninth step is to arrange the current x(h) and propeller vector into matrix B u , the output thrust u of the propulsion system, matrix A d Substitute into formula (28) to calculate the matrix x(h+1) - .

[0095] The tenth step is to update the current moment matrix p(h) and matrix A d , matrix E d , matrix Q is introduced into formula (29) to calculate the matrix p(h+1) - .

[0096] Step 11: At the next moment, the matrix p(h+1) - As the matrix P - (h), the matrix x(h+1) - As the matrix x - (h), combined with the input sensor information y(h), according to formulas (25) to (29), the actual position, heading, speed, and angular velocity information of the intelligent ship during autonomous navigation can be observed, thereby realizing the state observation of the intelligent ship during autonomous navigation.

Claims

1. A method for observing the motion state of an autonomous ship, characterized by: Using system matrix A d 、B d 、E d 、H d The determination method of , the determination method of the related measurement variance matrix R, and the noise covariance matrix Q; the discrete Kalman filtering algorithm is used to observe and predict the autonomous navigation motion state of intelligent ships.

2. The method for observing the motion state of an autonomous ship according to claim 1, characterized in that: The system matrix A d 、B d 、E d 、H d Determination method: According to the wave intensity parameter σ, damping coefficient λ, dominant wave frequency ω0, mass matrix M, damping coefficient matrix D, time constant diagonal matrix T, thruster vector arrangement matrix B u To determine the system matrix A d 、B d 、E d and H d .

3. The method for observing the motion state of an autonomous ship according to claim 1, characterized in that: The method for determining the correlation measurement variance matrix R and the noise covariance matrix Q is as follows: the covariance is calculated based on the samples of the longitudinal position information, lateral position information and heading information measured by the intelligent ship when it is not moving in the port. To estimate the matrix R; combine the high-frequency motion model of intelligent ships and the data during autonomous navigation to calculate the matrix Q1; the noise estimation matrix Q2 of position and heading measurement; the motion deviation estimation matrix Q3 caused by uncertain environmental forces; Form the matrix Q.

4. The method for observing the motion state of an autonomous ship according to claim 1, characterized in that: The discrete Kalman filter algorithm is used to observe and predict the autonomous navigation motion state of the intelligent ship. The discrete Kalman filter gain iteration formula is: K(h)=P - (h)*H d T (h)*(H d (h)*P - (h)*H d T (h)+R) -1 ; The iterative formula for motion state observation is x(h)=x - (h)+K(h)*(y(h)-H d *x - (h)), and the north position in the state observation variable x(h) is the 7th element, the east position is the 8th element, the heading angle is the 9th element, the longitudinal speed is the 10th element, the lateral speed is the 11th element, and the bow angular velocity is the 12th element; the iterative formula for updating the matrix is ​​p(h) = (I 15×15 -K(h)*H d )*P - (h)*(I 15×15 -K(h)*H d ) -1 +K(h)*R*K(h) T ; The iterative formula of the next moment state observation variable is x(h+1) - =A d *x(h)+B d *u; at the same time, the state observation variable x(h+1) - The north position is the 7th element, the east position is the 8th element, the heading angle is the 9th element, the surge velocity is the 10th element, the sway velocity is the 11th element, and the bow angular velocity is the 12th element. The iterative formula for updating the matrix at the next moment is predicted to be p(h+1) - =A d *p(h)*A d T +E d *Q*E d T .

5. The method for observing the motion state of an autonomous ship according to claim 1, characterized in that: The method for observing the autonomous navigation motion state of a ship comprises the following specific steps: In the first step, during autonomous navigation, the intelligent ship only considers the three degrees of freedom on the water surface: sway, swing, and pitch. The position and heading in the northeast coordinate system are selected as variables η. η=[NE ψ] T (1) Where, N is the north position information; E is the east position information; ψ is the heading information; Select the velocity in the hull coordinate system as vector v v=[uvr] T (2) Where u is the velocity in the longitudinal direction; v is the velocity in the transverse direction; and r is the angular velocity in the yaw direction. The conversion relationship between the two is as follows: Where: R(ψ) is the transformation matrix. In the second step, the slowly varying disturbance motion generated by the environmental forces during autonomous navigation is estimated using a first-order Gauss-Markov model, namely: Where b∈R 3 is the deviation between force and torque; T∈R 3×3 is the diagonal matrix of time constants; E b ∈R 3×3 is a diagonal matrix; it indicates the amplitude of the environmental disturbance; ω b ∈R 3 is the Gaussian white noise matrix; The third step is to establish a nonlinear low-frequency motion model of intelligent ships. Where M is the ship mass matrix including the additional mass; D is the ship damping coefficient matrix; τ is the resultant torque output by the propulsion system in the three degrees of freedom direction, Where M is represented by: Where m is the mass of the smart ship; x G is the position vector in the x direction in the center of gravity hull coordinate system; I z is the moment of inertia of the smart ship; and are the additional mass of the hydrodynamic force; The additional mass generated by the mutual coupling of longitudinal surge and bow motion; D is represented by: Where, X u 、Y v 、Y r 、N v and N r are the coefficients of linear damping of the intelligent ship in three degrees of freedom, τ is expressed as: τ=B u T F =[F x F y M z ] T (9) Where B u is the thruster vector arrangement matrix; T F Output thrust for the propeller; The fourth step is to calculate the high-frequency motion caused by the first-order waves during the autonomous navigation of the intelligent ship. The high-frequency motion in the surge, sway and heading directions is expressed as a separate motion in a single degree of freedom: Expand the above equation into: Where x ω1 and x ω2 is the motion generated by the first-order and second-order wave forces acting on the smart ship; ω0 is the dominant wave frequency in the wave PM spectrum; λ is the relative damping coefficient; σ is a constant value related to the wave intensity; η ω It is the low-frequency motion caused by the second-order wave force; Step 5: Based on the above equation, select the following intelligent ship motion equation: Where, E v Diagonal matrix; ω v is Gaussian white noise; at the same time, select x=[ξ T ,η,v,b T ] T ∈R 15 is the state observation variable; Step 6: Based on the position and heading information provided by the sensor system, design the following initial state linear observation equation: y=H d x+υ (16) Where: A d is a 15×15 order state transfer matrix; B d is the 15×p-order control gain input matrix; u is the p-order control force input vector; E d is a 15×15 order noise input matrix; ω is a 15-order Gaussian white noise; y is a 3D measurement vector; H d is a 15×1-order measurement matrix; v is a 15-order measurement Gaussian white noise; ω and v are independent white noises with a mean of 0; Thus, the discrete system matrix A can be obtained d 、B d 、E d and H d for: H d =[0 3×3 I 3×3 I 3×3 0 3×3 0 3×3 ] (20) A in the system matrix ω and E ω The settings are as follows: Step 7: Set the measurement variance matrix R = R 3×3 , the matrix R can use the following matrix: Where, The covariance of the measurement noise of the surge position information, sway position information, and heading information can be calculated by the covariance of the data samples measured when the smart ship is not moving in the port; Step 8: Set the state noise covariance matrix Q = R 9×9 , the diagonal line is selected as the positive adjustable parameter, that is: Matrix Q1 represents the noise of linear wave motion, which is calculated and set by combining the high-frequency motion model of intelligent ships and data from autonomous navigation. Matrix Q2 represents the noise of position and heading measurements during low-frequency motion. Matrix Q3 represents the motion deviation caused by uncertain forces caused by various environmental forces. Step 9: Select the estimated position and heading of the starting point of autonomous navigation of the intelligent ship and set the initial value x - (h)[0]; Based on the true initial value x[0] and the estimated initial value x - (h)[0] variance calculation updates the initial value P of the matrix - (h)[0], or directly use the unit matrix to converge quickly, that is: P - (h)[0]=I 15×15 ; The tenth step is to use the discrete Kalman filter algorithm of formula (25) to formula (29) below according to the initial value and the position and heading information y(h) provided by the sensor system to obtain the ship's current moment in autonomous navigation: The ship's position, heading, speed, and angular velocity information are: the 7th, 8th, 9th, 10th, 11th, and 12th elements in x(h); the ship's position, heading, speed, and angular velocity information at the next moment are predicted, namely: x(h+1) - The 7th, 8th, 9th, 10th, 11th, and 12th elements in ; K(h)=P - (h)*H d T (h)*(H d (h)*P - (h)*H d T (h)+R) -1 (25) x(h)=x - (h)+K(h)*(y(h)-H d *x - (h)) (26) p(h)=(I 15×15 -K(h)*H d )*P - (h)*(I 15×15 -K(h)*H d ) -1 +K(h)*R*K(h) T (27) x(h+1) - =A d *x(h)+B d *u (28) p(h+1) - =A d *p(h)*A d T +E d *Q*E d T (29) In the eleventh step, the obtained real position, heading, speed, and angular velocity are input into the controller as information for actual movement of the intelligent ship's autonomous navigation for control calculation.

Images