A high-speed unmanned surface vehicle path tracking control method based on improved EMPC

Abstract

The application discloses a high-speed unmanned ship path tracking control method based on improved EMPC, and comprises the following steps: step 1, constructing the basic model and the coordinate system of the underactuated high-speed unmanned ship on water surface; step 2, in the guidance method, the LOS front-view garden radius is optimized by the dynamic line-of-sight method in combination with the current ship speed, and the desired heading angle is obtained on the basis of the optimized LOS front-view garden radius in combination with the lateral deviation rate; step 3, in the offline state of the EMPC controller, the optimal control law is solved for each state partition in the offline state by the heron optimization algorithm, and the linear control law on each state partition and the corresponding partition is obtained; step 4, in the online state of the EMPC controller, the linear control law on the corresponding partition obtained in step 3 is found by the reachable partition searching method, the unmanned ship is controlled to run and track the last desired path point, and the tracking is completed. The application improves the control precision, reduces the lateral deviation and improves the adaptive ability of the guidance parameters.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of high-speed unmanned ship control, and particularly relates to a high-speed unmanned ship path tracking control method based on improved EMPC. BACKGROUND

[0002] As an important guarantee for the safety, autonomy, precision and rapidity of unmanned ships in completing various tasks, the control target of path tracking technology is to design a controller to enable the unmanned ship to accurately track and remain on a desired path independent of time in a tracking space. During high-speed navigation of the underactuated unmanned ship, there are problems such as nonlinearity, model uncertainty, time-varying of external disturbances, and in actual engineering applications, there are problems such as parameter adaptability, controller robustness, tracking effect stability and navigation safety. Therefore, designing a path tracking controller with strong adaptability, high robustness, strong anti-interference ability and high real-time performance will play an important role in the task adaptability and debugging safety of the high-speed unmanned ship.

[0003] Currently, main control methods include backstepping, sliding mode control, dynamic surface control, disturbance observer (Disturbance Observers, DO), neural network, adaptive control and model predictive control. However, these methods have various shortcomings in actual application. In view of these shortcomings, researchers at home and abroad have proposed various improvement schemes. For example, Sun Z et al. proposed a practical proportional integral sliding mode method to solve the problem that the integral sliding mode will produce a non-eliminable term ψ c r c , which will affect the stability of the system. The method additionally considers the heading angle and position tracking error when designing the sliding surface, effectively solving the problem. Chen Zhijuan designed a controller combining model predictive control to solve the rudder angle saturation problem and used the training approximation characteristics of the radial basis neural network to approximate the external disturbance through the training of the ship's historical information, compensate the MPC and improve the robustness of the system. Yang Tiantian combined the LOS line-of-sight guidance algorithm with the predictive function control to propose a LOS+PFC-based path tracking control system, and used the simulated annealing algorithm to solve the optimal control sequence of the predictive control algorithm, further improving the accuracy and real-time performance of the path tracking control.

[0004] The closest prior art to the method herein is the method of using the display model predictive control to solve the real-time requirement problem in the actual navigation of the unmanned ship proposed by Chen Tianyuan in the thesis of "Unmanned Ship Path Control Method Based on Display Model Predictive Control". The method first simplifies the unmanned ship path control into heading control by using the line of sight guidance method (LOS), and then applies the display model predictive control algorithm to the unmanned ship heading control problem, which improves the calculation speed while ensuring the control accuracy, and to some extent solves the real-time problem in high-speed navigation.

[0005] The closest prior art to the method herein, although to some extent reduces the calculation time, but in the actual control, loses a certain control accuracy, so it is necessary to add a suitable algorithm to optimize the objective function on each state partition in offline calculation to obtain the corresponding optimal control law, so as to compensate for the loss of control accuracy. At the same time, the parameter adaptive ability of the traditional LOS guidance method is weak, and is easy to be disturbed by the environment to produce drift angle, so it is necessary to improve the straight line and curve guidance respectively, reduce the lateral deviation, and improve the adaptive ability when switching the path, so that the USV can reach the expected path faster. SUMMARY

[0006] The application provides a high-speed unmanned ship path tracking control method based on improved EMPC to solve the problem of loss of control accuracy, weak parameter adaptive ability of the guidance method, and easy to be disturbed by the environment to produce drift angle in the prior art.

[0007] The application provides a high-speed unmanned ship path tracking control method based on improved EMPC, comprising the following steps:

[0008] Step 1: constructing the dynamic model, kinematic model, fixed coordinate system and unmanned ship carrier coordinate system of the underactuated high-speed unmanned ship on the water surface;

[0009] Step 2: in the guidance method, the LOS forward vision radius is optimized by combining the current ship speed through the dynamic line of sight method, and the expected heading angle is obtained on the basis of the optimized LOS forward vision radius combined with the lateral deviation rate, and the deviation angle is obtained according to the actual heading angle;

[0010] Step 3: in the offline state of the EMPC controller of the control method, the white egret group optimization algorithm is used to solve the optimal control law of each state partition in the offline state, and the linear control law of each state partition and the corresponding partition is obtained;

[0011] Step 4: In the EMPC controller online state of the control method, the linear control law on the corresponding partition obtained in step 3 is searched by the reachable partition search method, and the unmanned ship is controlled to run with the searched linear control rate until the unmanned ship tracks the last expected path point, and the tracking control process is completed.

[0012] Further, in step 1, the disturbance model in the dynamic model of the underactuated high-speed surface unmanned ship includes a wind disturbance force model, a wave disturbance force model, and a flow disturbance force model, and the specific formula is:

[0013]

[0014] In the formula, [u, v, r] T represents the speed of the unmanned ship, which is denoted by m ii (1, 2, 3) represents the inertial water force of the unmanned ship, which is the force generated by the surrounding water flow during the acceleration process of the USV, and is specifically denoted by d 11 =-X u , d 22 =-Y v , d 33 =-N r is the water force damping coefficient; τ X is the longitudinal thrust, τ N is the rotation force, τ wX is the longitudinal disturbance force, τ wY is the transverse disturbance force, and τ wN is the rotation disturbance force.

[0015] Further, the wind disturbance force model is:

[0016]

[0017]

[0018]

[0019] In the formula, is the relative wind speed, γ R =tan -1 (u R / v R ) is the relative speed, C X , C Y is the thrust coefficient; C N is the moment coefficient; ρ w is the air density, with the unit of kg / m 3 ; A T , A L are the transverse and longitudinal projection areas, respectively; L is the total length of the unmanned ship, with the unit of m; VR is the wind speed, unit is m / s.

[0020] Further, the wave interference force model is:

[0021]

[0022] wherein, p is the density of seawater, x is the encounter angle, m is the number of wave spectrum frequency division, Xb, Yb, Nb are test coefficients, is the wave surface equation, is the wavelength.

[0023] Further, the flow interference force model is:

[0024]

[0025] wherein, F Hr = -C(v r )v r -D(v r )v r is the fluid force after the interference of the sea current, v r = [u+u c , v+v c , r] T is the projection of the relative speed of the water surface unmanned ship movement to the water flow; F H = -C(v)v-D(V)v is the force generated by the relative motion of the fluid.

[0026] Further, in step 2, the formula for optimizing the LOS front view circle radius by the current ship speed is as follows:

[0027] R K = e L +e -λU k L

[0028] wherein, R K is the LOS front view circle radius; e t is the vertical distance of the water surface unmanned ship from the desired route at the current time; U is the current speed;

[0029] The specific formula for obtaining the desired heading angle based on the optimized LOS front view circle radius combined with the lateral deviation rate is as follows:

[0030]

[0031] wherein,

[0032] Further, the discrimination condition in the white egret group optimization algorithm is:

[0033]

[0034] Advantages of the present application:

[0035] The method is aimed at the problem of large parameter calculation amount and slow solving efficiency of traditional model predictive control in the path tracking process of high-speed unmanned ship, adopts the method of explicit model predictive control, and introduces a heron optimization algorithm in offline calculation part to improve solving of the optimal control law of each state partition. BRIEF DESCRIPTION OF DRAWINGS

[0036] The features and advantages of the present application will be more clearly understood through reference to the following drawings, which are presented as exemplary and should not be construed as limiting the application, in which:

[0037] Figure 1 The figure is a line-of-sight guidance schematic diagram in the embodiment of the present application;

[0038] Figure 2 The figure is a heron optimization algorithm block diagram in the embodiment of the present application;

[0039] Figure 3 The figure is an optimal algorithm convergence curve comparison diagram in the embodiment of the present application.

[0040] Figure 4 The figure is a reachable partition algorithm flowchart in the embodiment of the present application;

[0041] Figure 5 The figure is a control system block diagram in the embodiment of the present application. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be described clearly and completely below in combination with the drawings of the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0043] The embodiment of the present application provides a high-speed unmanned ship path tracking control method based on improved EMPC, including the following steps:

[0044] Step 1: Construct the kinematics and dynamics mathematical model of the underactuated high-speed surface unmanned surface vehicle (USV). The Fossen ship dynamics mathematical model is adopted, and the ship six-degree-of-freedom model is simplified into the unmanned surface vehicle (USV) motion model of three degrees of freedom of the USV surge, sway and yaw.

[0045] 1) Kinematics model of high-speed surface unmanned surface vehicle:

[0046]

[0047] 2) Dynamics model of high-speed surface unmanned surface vehicle:

[0048]

[0049] In the formula, [x, y, ψ] T represents the position vector of the unmanned surface vehicle, [u, v, r] T represents the velocity of the unmanned surface vehicle, where m ii (1, 2, 3) represents the inertia of the USV in the acceleration process, which is the force generated by the surrounding water flow inertia, and is specifically represented as d 11 =-X u , d 22 =-Y v , d 33 =-N r is the water power damping coefficient; τ x is the longitudinal thrust, τ N is the rotation force, τ wX is the longitudinal disturbance force, τ wY is the lateral disturbance force, and τ wN is the rotation disturbance force. In the above formula, X u , Y v , Y r , N v , N r , I z is the water power coefficient.

[0050] The unmanned surface vehicle in the embodiment can only generate longitudinal thrust τ x and rotation force τ N , in order to make the model more accurate, an external disturbance model is additionally added, which is defined as follows:

[0051]

[0052] In the formula, is the wind disturbance force model, For the wave disturbance force model, For the flow disturbance force model.

[0053] The wind speed generally contains a slowly changing component (average wind speed) and a high frequency component (gust). The resultant force and moment acting on the surface unmanned vehicle are usually expressed in terms of the relative wind speed V R and the relative velocity γ R , which depend on:

[0054]

[0055] γ R = tan -1 (u R / v R ) (5)

[0056] V R is the resultant velocity in the x, y direction. The wind speed is decomposed along the body-fixed coordinate system, taking into account the relative wind speed of the surface unmanned vehicle in the sailing state:

[0057] u R = V w cos (ψ R )-u+u c

[0058] v R = V w sin (ψ R )-v+v c (6)

[0059] where ψ R = ψ-ψ w is the relative angle between the wind direction and the heading of the vehicle. For most surface unmanned vehicles, the gust cannot be compensated by the control system. However, by calculating the slowly changing wind force from the wind speed and direction, the slowly changing wind force is fed forward to the controller, and the wind force and moment acting on the vehicle are as follows:

[0060]

[0061]

[0062]

[0063] where C X , C Y are the thrust coefficients; C N is the moment coefficient; p w is the air density, with units of kg / m 3 ; A T , A L are the lateral and longitudinal projected areas, respectively; L is the total length of the unmanned vehicle, with units of m; VR V is the wind speed, unit is m / s.

[0064] For the sea waves, we usually use linear superposition to describe it. In order to predict the disturbance force of the surface unmanned vehicle in irregular sea waves, we usually use the sea wave spectrum to describe it in the theory of surface unmanned vehicle seakeeping. The commonly used sea wave spectrum includes PM spectrum, single parameter spectrum and double parameter spectrum. In this paper, we use the ITTC double parameter sea wave spectrum recommended by the International Ship Model Basin Conference to predict the irregular sea waves, and the formula is as follows:

[0065]

[0066] In the formula, ζ W / 3 is the significant wave height, ω is the wave frequency, and T1 is the average wave period.

[0067] According to the sea wave spectrum, the wave surface equation of the sea wave can be obtained:

[0068]

[0069] The wavelength is:

[0070]

[0071] Finally, according to the wave force formula on the ship body, the wave force calculation formula on the surface unmanned vehicle is as follows:

[0072]

[0073] In the formula, ρ is the density of seawater, χ is the encounter angle, m is the number of frequency division of the sea wave spectrum, Xb, Yb, Nb are test coefficients, which can be estimated according to the regression formula obtained by J.W English et al.:

[0074]

[0075] Let the flow velocity be V c , and the flow direction be β. The flow velocity is decomposed along the fixed coordinate system on the sea surface:

[0076]

[0077] According to the coordinate transformation formula, we can get:

[0078]

[0079] Substitute formula 3-22 into the above formula:

[0080]

[0081] Let the components of the water surface unmanned vehicle motion velocity in the body coordinate system be u, v, and the projection of the water surface unmanned vehicle relative velocity to the water flow be:

[0082] v r = [u + u c , v + v c , r] T (16)

[0083] For the water surface unmanned vehicle dynamics model, the force generated by the relative motion of the fluid is:

[0084] F H = -C(v)v-D(v)v (17)

[0085] Using the relative velocity to replace the actual velocity, the fluid force after the sea current interference can be obtained:

[0086] F Hr = -C(v r )v r -D(v r )v r (18)

[0087] After difference, the sea current force can be obtained:

[0088]

[0089] Step 2: In the guidance method, the LOS front view radius is optimized by dynamic line of sight method combined with the current ship speed, and the desired heading angle is obtained on the basis of the optimized LOS front view radius combined with the lateral deviation rate, and the deviation angle is obtained according to the actual heading angle. Assume that the unmanned vehicle is located at point p t (x t , y t ) at the current time, and the expected center of the tracking point is point p k (x k , y k ), as shown in Figure 1 .

[0090] According to the current position of the unmanned vehicle and the expected center position, the target angle is obtained:

[0091]

[0092] According to the circle tracking direction, the expected path point is obtained:

[0093]

[0094]

[0095] Then the desired heading angle is obtained:

[0096]

[0097] In the complex sea conditions, the unmanned surface vehicle (USV) will produce a drift angle when tracking a circular path, which will interfere with the heading. To eliminate the influence of the drift angle, an S-plane control radian correction guidance law is designed according to the lateral deviation and the lateral deviation variation law to guide the USV to track a circle.

[0098] The lateral deviation of the USV from the desired path is:

[0099]

[0100] The lateral deviation variation rate is:

[0101]

[0102] The desired heading angle of the USV is:

[0103]

[0104] According to the above formula, the desired heading of the USV after correction is:

[0105]

[0106] In the formula,

[0107]

[0108] Therefore, according to the actual heading angle and the difference between the desired heading angle, the deviation angle

[0109]

[0110] Path segment update. When selecting the path P n+1 , it is determined whether the USV is in the circle with P n as the center and R as the radius. If it is satisfied, the next path P n+1 is tracked. Let the current position of the USV be (x n (t), y n (t)) satisfy:

[0111]

[0112] Select (x n+1 (t), y n+1 (t)) as the end point of the next path. According to the vertical distance of the USV from the desired path and the current speed of the USV, the front-view circle radius of the USV guidance law is dynamically adjusted:

[0113] R K = e L + e -λU k L (23)

[0114] where e L is the vertical distance between the current position of the USV and the desired path.

[0115] The linearization of the predictive control model is performed to track the desired path by the given desired state, the desired heading angle after the LOS algorithm processing, and the actual heading angle of the current state The reference system equation, i.e., the reference trajectory without considering the disturbance, is shown in equation (32):

[0116]

[0117] The first-order Taylor expansion of the function at any reference point (xR, uR) is shown in equation (33)

[0118]

[0119] Subtracting equation (33) from equation (32) gives

[0120]

[0121] where A and B are Jacobian matrices.

[0122] The new predictive model, i.e., equation (34), is discretized. There are many methods for discretization, such as the Runge-Kutta method, and the forward Euler method is used here to obtain equation (35):

[0123]

[0124] where T is the sampling time, and I is the identity matrix. Combining the equations gives

[0125]

[0126]

[0127] where P is the identity matrix.

[0128] The above equations are transformed and simplified into the state space model in the form of control increment:

[0129]

[0130]

[0131] wherein

[0132] Step 3: In the EMPC controller offline state of the control method, the optimal control law of each state partition in the offline state is solved by the heron optimization algorithm under the setting of constraint conditions, and the linear control law on each state partition and the corresponding partition is obtained.

[0133] Constraint condition setting: Due to the influence of mechanical performance on the propeller and rudder of the unmanned ship, the motion performance and speed are limited, and saturation phenomenon is easy to occur during high-speed navigation. Therefore, the control limit, control increment, and output constraint are set at time k and in the prediction time domain.

[0134] Optimal solution of the objective function: Considering the speed of the target and the energy loss of the system input control quantity, the state quantity deviation of the system, the control quantity, and the control increment are used to construct the objective function, which is as follows:

[0135] J = (Y-Yref) T Q(Y-Yref)+ΔU T RΔU (32)

[0136] where Yref is the expected value, ΔU is the control increment, and Q and R are weight matrices (which are continuously adjusted according to the needs of control).

[0137] The heron optimization algorithm consists of three main parts: waiting strategy, aggressive strategy, and discrimination condition. The algorithm block diagram is shown in Figure 2 .

[0138] Each heron group can be composed of n heron squads, and each heron squad contains three herons, among which heron A implements the waiting strategy, and herons B and C use the random walk and surround mechanism in the aggressive strategy, respectively.

[0139] 1) Waiting strategy (heron A). The observation equation of the ith heron A can be described as The real fitness y i obtained by each iteration can be used to obtain the pseudo-gradient g i of the weight in the observation equation, so the updated position of heron A is represented as follows:

[0140] X a，i = X i +exp(-t / (0.1*t max ))*0.1*hop*g i (41)

[0141] where t is the current iteration number, t maxN is the maximum number of iterations, and hop is the feasible range of the independent variable.

[0142] 2) Aggressive strategy (Heron B, Heron C). Heron B is a random walk, which updates the position as follows:

[0143] X b，i = X i + tan(r b，i )*hop / (1+t) (42 )

[0144] r b，i is a random number between (-π / 2, π / 2).

[0145] Heron C is a containment strategy, which updates the position as follows:

[0146] D h = X ibest -X i (43)

[0147] D g = x gbest -X i (44 )

[0148] X c，i = (1-r i -r g )*X i +r h *D h +r g *D g (45 )

[0149] where X ibest and X gbest are the optimal value of the Heron squadron and the optimal value of the Heron population, respectively, and r h and r g are random numbers between [0, 1].

[0150] 3) Discrimination condition. After each Heron of the Heron squadron calculates the updated position, they collectively determine the updated position of the Heron squadron, which is as follows:

[0151] X Si＝ = [X a，i X b，i X c，i ] (46)

[0152] y s，i = [y a，i y b，i y c，i ] (47)

[0153] c i =argmin(y s，i (48)

[0154]

[0155] The egret team compares the updated positions and fitness of the three egrets with the fitness of the previous iteration. If the updated position of one egret is better than the position of the previous iteration, the update is adopted; if the updated positions of each egret are worse than the previous one, there is a 33% probability that the optimal updated position is adopted.

[0156] Figure 3 This image compares the convergence curves of the egret flocking optimization algorithm and the particle swarm optimization algorithm. The left half of the image represents the search space of the objective function, and the right half represents the convergence curves of the two optimization algorithms. By comparing the curves, it is clear that the egret flocking optimization algorithm converges faster and has higher solution accuracy than the particle swarm optimization algorithm.

[0157] The optimal control increment sequence is obtained by using the egret flock optimization algorithm, resulting in equation (50):

[0158] ΔU * =[Δu(t|t) * Δu(t+1|t) * … Δu(t+Nc-1|t) * (50)

[0159] Step 4: In the online state of the EMPC controller of the control method, based on the actual state quantities output by the high-speed unmanned surface vessel (USV) sensors, an online search method is used to determine the state partition, find the corresponding optimal control law, and apply the control components to the USV.

[0160] The flowchart of the reachable partition search algorithm is as follows: Figure 4 As shown, the specific steps are as follows:

[0161] 1) For the initial state x0, locate the corresponding sub-region n using a sequential search method;

[0162] 2) Obtain the corresponding control quantity u based on the sub-region n;

[0163] 3) Apply the control variable to the system and detect the system state variable at the next moment;

[0164] 4) Search for the state partition where the state variable is located based on the system's state variables and reachable partitions.

[0165] 5) Repeat steps (2), (3), and (4) until the table lookup is complete.

[0166] The optimal control component obtained by the table lookup is applied to the high-speed unmanned ship.

[0167] u(t) * = Δu(t|t) * + u(t-1) (51)

[0168] The whole control system block diagram is shown in Figure 5 The improved prediction control method is used in this paper, which is LOS-ESOA-EMPC control method. The dynamic model of the underactuated unmanned ship is established as the state space model of the prediction control. The state space model is linearized and discretized to be the error form of the control increment equation as the system model. A desired path curve is set, which is evenly divided into multiple desired points. The position information of the desired points is obtained and the desired heading angle is calculated by using the improved LOS guidance algorithm according to the current state of the unmanned ship. The error between the desired heading angle and the actual heading angle is used as the input of the EMPC controller. In the offline calculation, the white egret optimization algorithm (ESOA) is used to optimize the objective function under the setting of the initial state and the constraint condition, so as to obtain the linear control law of each state sub-region and the corresponding sub-region. In the online calculation, the actual state quantity output by the sensor of the high-speed unmanned ship is used to judge the state sub-region, to find the corresponding control law, and to apply the control component to the unmanned ship. According to the planned desired path, steps 1-3 are repeated. At the next time, the actual value measured by the sensor of the high-speed unmanned ship is fed back to the EMPC controller, and the online searching is continuously carried out until the high-speed unmanned ship tracks to the last desired point.

[0169] Although the embodiments of the present application are described in conjunction with the drawings, various modifications and changes can be made by those skilled in the art without departing from the spirit and scope of the present application, and such modifications and changes fall within the scope defined by the appended claims.

Claims

1. A path tracking control method for high-speed unmanned surface vessels based on improved EMPC, characterized in that, Includes the following steps: Step 1: Construct the dynamic model, kinematic model, fixed coordinate system, and unmanned surface vessel carrier coordinate system of the underactuated high-speed surface vessel; Step 2: In the guidance method, the LOS forward radius is optimized by dynamic line-of-sight method in combination with the current ship speed. Based on the optimized LOS forward radius, the desired heading angle is obtained by combining the lateral deviation rate, and the yaw angle is obtained according to the actual heading angle. Step 3: In the offline state of the EMPC controller in the control method, the optimal control law is solved for each state partition in the offline state using the egret flocking optimization algorithm, obtaining each state partition and the corresponding linear control law on each partition. The discrimination condition of the egret flocking optimization algorithm is: ； Step 4: In the online state of the EMPC controller of the control method, the linear control law on the corresponding partition obtained in Step 3 is found by the reachable partition search method. The unmanned surface vessel is controlled by the queried linear control law until the unmanned surface vessel tracks the last desired path point, thus completing the tracking control process.

2. The high-speed unmanned surface vessel path tracking control method based on improved EMPC as described in claim 1, characterized in that, In step 1, the disturbance models in the dynamic model of the underactuated high-speed unmanned surface vessel include: wind disturbance force model, wave disturbance force model, and current disturbance force model, with specific formulas as follows: ； In the formula, The speed of the unmanned surface vessel is expressed here using... The inertial hydrodynamic force of an unmanned surface vessel (USV) is represented as the force generated by the inertia of the surrounding water flow during the USV's acceleration, specifically expressed as... , , ; , , This is the hydrodynamic damping coefficient; For longitudinal thrust, For rotational force, For longitudinal interference force, For lateral interference force, It is a gyratory disturbance force.

3. The high-speed unmanned surface vessel path tracking control method based on improved EMPC as described in claim 2, characterized in that, The wind interference force model is as follows: ； In the formula, Relative wind speed, For relative velocity, , This is the thrust coefficient; This is the torque coefficient; air density; , These represent the horizontal and vertical projected areas, respectively; L is the total length of the unmanned surface vessel. This refers to wind speed.

4. The high-speed unmanned surface vessel path tracking control method based on improved EMPC as described in claim 2, characterized in that, The wave interference force model is as follows: ； In the formula, The density of seawater, Let m be the encounter angle, and m be the number of frequency divisions in the wave spectrum. For experimental coefficients, The equation for the wave surface of an ocean wave is... λ is the wavelength.

5. The high-speed unmanned surface vessel path tracking control method based on improved EMPC as described in claim 2, characterized in that, The flow interference force model is as follows: ； In the formula, The fluid forces resulting from ocean current disturbance. The projection of the relative velocity of the unmanned surface vessel to the water flow; Forces generated by the relative motion of fluids.

6. The high-speed unmanned surface vessel path tracking control method based on improved EMPC as described in claim 1, characterized in that, In step 2, the formula for optimizing the LOS forward radius based on the current ship speed is as follows: ； In the formula, Let LOS be the radius of the forward-looking circle; U represents the vertical distance of the unmanned surface vessel from the desired route at the current moment; U represents the current speed. The specific formula for obtaining the desired heading angle based on the optimized LOS forward radius and the lateral deviation rate is as follows: ； In the formula, , , , .

Images