An unmanned ship autonomous guiding shipborne unmanned aerial vehicle recovery control method

Abstract

The application discloses a kind of unmanned ship autonomous guidance shipborne unmanned plane's recovery control method, comprising the following steps: according to the sea environment where current unmanned ship is located, the best heading and best speed of execution unmanned plane recovery are judged;Unmanned ship according to the best condition obtained in last step, unmanned plane calculates relative position information according to the inertial sensor and GPS carried by itself;Through the relative position information of boat-machine obtained in last step, unmanned plane is controlled to the vicinity of unmanned ship parking apron, realizes preliminary alignment, keeps unmanned plane and parking apron position relative vertical;After unmanned plane approaches the top of unmanned ship, the visual sensor of unmanned plane is used to accurately align unmanned ship parking apron standard mark;Unmanned ship and unmanned plane are kept relative static, whether landing condition is satisfied is judged, if satisfied, unmanned plane recovery is completed, if not satisfied, return to third step.The application ensures that positioning is realized under the condition of best speed and heading, improves the precision of unmanned plane landing on unmanned ship.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of coordinated control of a boat and a machine, and in particular to a control method for autonomous guidance of a ship-borne unmanned aerial vehicle by an unmanned boat. BACKGROUND

[0002] In recent years, with the vigorous development of the strategy of a maritime power, the sea-air unmanned system has been greatly developed. The current sea-air unmanned system can achieve a certain degree of unmanned operation, which has been highly valued by countries around the world in the civil and military fields. China is also vigorously promoting the construction of the air-sea unmanned system capability and enhancing the research and development of unmanned system equipment. With the unmanned trend of the air-sea system, the development of unmanned systems to achieve more functions and more scene applications, and the full play of the efficiency of unmanned systems is the mainstream of current unmanned system research.

[0003] With the development and testing of medium and large unmanned boats, unmanned boats carrying unmanned aerial vehicles to perform tasks have become a new use style. The unmanned boat plans the action according to the task, sails to the standby sea area, releases the boat-borne unmanned aerial vehicle and guides it to the task sea area for reconnaissance, and the boat and the machine share the detection information in real time through the measurement and control chain to the unmanned boat control system, which can effectively extend the perception range of the unmanned boat, improve the ability of a single boat to perform tasks, and solve the problem of the unmanned boat using its own information to guide the over-the-horizon anti-ship missile attack. The present application mainly proposes a control method for autonomous guidance of a boat-borne unmanned aerial vehicle by an unmanned boat.

[0004] In the existing technology, for the control method of unmanned aerial vehicle-unmanned boat coordinated landing, the method of unmanned aerial vehicle tracking unmanned boat control is generally considered in an ideal state, for example, the unmanned aerial vehicle dynamic landing control algorithm [Liu F, Yu Y, Guo X, Bi W, Gao J. Research on unmanned aerial vehicle autonomous landing technology [J]. Journal of Yanshan University, 2022, 46(04):353-361.], the information of the unmanned boat and the surrounding environment is obtained by using the on-board camera, and the captured image is processed to make the unmanned aerial vehicle move to the top of the unmanned boat. However, this landing technology considers an ideal water environment and uses the unmanned boat as a mobile platform, and does not involve information interaction between the boat and the machine, so it cannot further make full use of the boat and machine information for positioning. SUMMARY

[0005] The technical problem to be solved by the present application is to provide a control method for autonomous guidance of a ship-borne unmanned aerial vehicle by an unmanned boat, to ensure that the boat and the machine are positioned cooperatively under the best speed and heading conditions, to improve the precision of the unmanned aerial vehicle landing on the unmanned boat, and to reliably guarantee the unmanned aerial vehicle-unmanned boat coordinated landing control.

[0006] To solve the above technical problems, the present application provides a control method for autonomous guidance of a ship-borne unmanned aerial vehicle by an unmanned boat, comprising the following steps:

[0007] S1. Determine the optimal course and speed for drone recovery based on the current marine environment of the unmanned surface vessel;

[0008] S2. The unmanned surface vessel (USV) navigates according to the optimal heading and speed obtained from S1, while the unmanned aerial vehicle (UAV) calculates the relative position information of the USV and the UAV based on its own inertial sensors and GPS.

[0009] S3. Obtain the relative position information of the unmanned surface vessel and the drone through S2, control the drone to the vicinity of the unmanned surface vessel landing pad, achieve initial alignment, and keep the drone and the landing pad position relatively perpendicular;

[0010] S4. After the drone approaches the unmanned surface vessel (USV) from above, it uses the drone's visual sensors to precisely align with the standard markings on the USV's landing pad.

[0011] S5. Control the unmanned surface vessel and the drone to keep them relatively stationary, monitor the alignment error in real time, and determine whether the landing conditions are met. If they are met, complete the drone recovery; otherwise, return to S3.

[0012] Preferably, in step S1, determining the optimal course and speed for drone recovery based on the current marine environment of the unmanned surface vessel specifically includes the following steps:

[0013] S11. Assign weights to the pitch and roll motions that affect the stability of the unmanned surface vessel (USV). The degree of influence z of the determined pitch and roll motions on the USV's navigation is calculated by the following formula: z = w1z1 + w2z2. Where z1 and z2 represent the pitch and roll of the USV, respectively, and w1 and w2 represent the weights of the pitch and roll motions, respectively.

[0014] S12. Collect data on pitch and roll motions and assign weights based on S11; collect n sets of pitch and roll motion data within the sampling time T and obtain the comprehensive influence factor z based on the weights assigned in S11. i i = 1, 2, ..., n, and at the same time, record each z i Corresponding heading x i and speed y i ;

[0015] S13. Fit the relationship between the comprehensive influence factor obtained in S12 and the heading and speed using the least squares method; fit the z-value of S12 using the least squares method. i With heading x i and speed y i Relationship with time; fitting the acquired heading x i and speed y i The n×n rectangular point domain (x,y) is formed, where x=[x1,x2,...,x n ] T ,y=[y1,y2,...,yn ] T the function value z on the point x i ; fixed speed set y, the heading set x to construct n least square fitting polynomial

[0016]

[0017] Wherein each Constructed by the following recursive formula:

[0018]

[0019] Let Obtain

[0020]

[0021] According to the least square principle

[0022]

[0023] Then construct the least square fitting polynomial of the speed y

[0024]

[0025] Wherein each ψ l (y) (l = 0, 1, K, q-1) is constructed by the following recursive formula:

[0026]

[0027] According to the least square principle

[0028]

[0029] Finally, the fitting polynomial of the longitudinal and lateral shaking influence value and the unmanned ship heading and speed is

[0030]

[0031] S14, according to the fitting result of S13, the minimum heading and speed of longitudinal and lateral shaking are obtained; the minimum point of the function f(x, y) of S13 fitting is the best heading and speed.

[0032] Preferably, in step S2, the unmanned ship sails according to the best heading and speed obtained in S1, and the unmanned aerial vehicle calculates the relative position information of the unmanned ship and the unmanned aerial vehicle according to the inertial sensor and GPS carried by itself, specifically including the following steps:

[0033] S21, the unmanned aerial vehicle obtains the current attitude information of the body through the inertial measurement unit;

[0034] The k-1 moment angular velocity w(k-1) obtained by the inertial measurement unit is integrated to obtain the k moment attitude quaternion data Specifically as follows:

[0035]

[0036] Wherein, T represents the sampling time, and the attitude matrix E(q(k-1)) is expressed as follows:

[0037]

[0038] Wherein, I3 is a three-dimensional unit matrix, The matrix S(q v (k-1)) is expressed as follows:

[0039]

[0040] S22, calculating the rotation matrix of the navigation coordinate system and the airborne coordinate system;

[0041] The rotation matrix is obtained from the k moment attitude data obtained by S21, and is specifically as follows:

[0042]

[0043] Wherein:

[0044]

[0045]

[0046]

[0047] S23, calculating the position, attitude and speed of the unmanned aerial vehicle in the navigation coordinate system;

[0048] The position information p(k) and the speed information v(k) of the unmanned aerial vehicle in the navigation coordinate system can be obtained by GPS, and the angular velocity information can be obtained by the rotation matrix calculated by S22, and is specifically as follows:

[0049] w n (k)=R(q(k))w(k);

[0050] Wherein, w n (k) is the angular velocity of the unmanned aerial vehicle in the navigation coordinate system.

[0051] Preferably, in step S3, the relative position information between the unmanned aerial vehicle and the unmanned surface vehicle is obtained by S2, the unmanned aerial vehicle is controlled to the vicinity of the unmanned surface vehicle parking apron, the preliminary alignment is realized, and the unmanned aerial vehicle and the parking apron position are kept relatively vertical, and specifically includes the following steps:

[0052] S31, by the real-time attitude and position data of the current unmanned aerial vehicle, using Kalman filtering to predict the pose information of the unmanned aerial vehicle in the navigation coordinate system;

[0053] The pose information of the unmanned aerial vehicle at k+1 moment is predicted by Kalman filtering, including quaternion, velocity and position as follows:

[0054] X(k+1)=[q(k+1),v(k+1),p(k+1)] T

[0055] Wherein, q(k+1) is the quaternion attitude value at k+1 moment, v(k+1) is the velocity value at k+1 moment, p(k+1) is the three-dimensional space position at k+1 moment

[0056] The state prediction equation is

[0057] X(k+1)=FX(k);

[0058] Wherein, F is the state transition matrix, as follows

[0059]

[0060] The covariance P(k+1) propagation matrix is:

[0061] P(k+1)=FP(k)F T +Q;

[0062] Wherein, Q is the state noise covariance matrix; the measurement equation can be obtained by using the data of GPS:

[0063] z(k+1)=HX(k+1)+R;

[0064] Wherein, z(k+1)=[v(k+1),p(k+1)] T ,R is the measurement noise covariance, and H is the measurement matrix, as follows

[0065]

[0066] Then the update prediction equation is as follows:

[0067]

[0068] S32, the relative position information obtained by S2 is used to set the trajectory of the unmanned aerial vehicle approaching the unmanned ship;

[0069] The position point of the unmanned ship in the navigation coordinate system at the current moment is P t (k) obtained by S31, the unmanned aerial vehicle track is planned, and the specific method is to set n track points (x i ,yi ) satisfies:

[0070]

[0071] S33, the unmanned aerial vehicle moves to above the unmanned ship platform according to the preset track.

[0072] Preferably, in step S4, after approaching above the unmanned ship, the unmanned aerial vehicle precisely aligns the unmanned ship landing pad standard mark by using the visual sensor of the unmanned aerial vehicle, which specifically includes the following steps:

[0073] S41, correcting the distortion of the visual sensor carried by the unmanned aerial vehicle to reduce the distortion of the collected image caused by the defects of the lens of the camera lens;

[0074] Considering that the spatial three-dimensional coordinates of a point P in the world coordinate system are (X w ,Y w ,Z w ), the point P (X C ,Y C ,Z C ) in the camera coordinate system is converted by using a rotation matrix R 3×3 and a translation matrix T 3×1 , and the conversion relationship is as shown in the following formula:

[0075]

[0076] The camera intrinsic matrix K is obtained through camera calibration, and the pixel coordinate system of P is converted into an image as follows:

[0077]

[0078] Wherein, Z C is a scale ratio, which is given by the camera manufacturer; [uv1] T is the pixel coordinate of the image, f x , f y is the focal length, and u0, v0 is the distortion factor;

[0079] The spatial three-dimensional coordinates of the vertex of the mark code are P i =[X i ,Y i ,Z i ] T , the pixel point projected on the image is p i =[u i ,v i ] T , and the conversion relationship between the spatial point and the pixel point is obtained:

[0080]

[0081] wherein z i is the depth information, converted into a matrix form as:

[0082] z i p i = K [RP i + T] ;

[0083] The equation is solved iteratively to obtain the optimal pose R, T of the camera, i.e. distortion correction is completed;

[0084]

[0085] S42, keeping the angle of the UAV camera unchanged, the image taken by the UAV is preprocessed to obtain the pixel coordinates of the identifier image position;

[0086] The "H" identifier is selected as the parking identifier on the apron, which is surrounded by a black circle on a white background, has distinctive features, and the identifier is easy to be recognized by the camera. During the recognition process, the UAV camera is controlled to be 90 degrees pitch-down and 0 degrees azimuth (no zoom);

[0087] The picture of the parking identifier taken by the camera is subjected to weighted grayscale processing. For an image represented by an RGB three-channel color system, the coordinates of a single image pixel are (x, y), and the values corresponding to the red, green and blue channels of the pixel are R(x, y), G(x, y) and B(x, y). The weighted grayscale value Gray(x, y) is:

[0088] Gray(x, y) = w R R(x, y) + w G G(x, y) + w B B(x, y) ;

[0089] wherein w R , w G , w B are the weights of the corresponding channels;

[0090] The processed image is subjected to noise reduction using a local adaptive thresholding method. The formula of the adaptive thresholding method is as follows:

[0091]

[0092] wherein is the variance of the grayscale image, S xy is the filter window, is the variance and mean value of the pixels in the filter window, is the filtered pixel grayscale value;

[0093] The image after noise reduction is binarized by using a gray threshold method, a threshold T is first determined, and then the image is binarized according to the following formula:

[0094]

[0095] wherein G(x,y) is a pixel gray value before processing, is a pixel gray value after processing;

[0096] S43, using a template-based matching method to detect the identifier image, the unmanned aerial vehicle is built-in standard "H" identifier matching template, by comparing the correlation between the template and the region in the identification image to match the identification;

[0097] The size of the collected image S is MxN, the size of the identifier image template g to be matched in the image is mxn, S xy is a detection window, and a correlation coefficient matrix p(x,y) is calculated:

[0098]

[0099] wherein σ(S xy ,g) is the covariance of the detection window and the template g, D xy , D is the variance of the detection window and the template g, the correlation coefficient satisfies p(x,y)≤1, by selecting the region with the maximum correlation coefficient in the template and the region in the image, the identifier image is detected, thereby positioning the identifier image, and the position of the identifier image satisfies the following formula:

[0100]

[0101] wherein (x,y) is the pixel point coordinate of the identifier image.

[0102] Preferably, in step S5, the unmanned surface vehicle and the unmanned aerial vehicle are controlled to keep relatively static, the alignment error is monitored in real time, it is judged whether the landing condition is satisfied, if yes, the unmanned aerial vehicle is recovered, and if not, the step S3 is returned, which specifically includes the following steps:

[0103] S51, according to the identifier pixel point coordinate obtained in S4, horizontal offset and rotation offset are calculated;

[0104] According to the identifier image coordinate (x,y) given in step S4, the horizontal offset Δx and Δy of the unmanned aerial vehicle are solved, and the calculation formula is as follows:

[0105]

[0106] wherein Camera_Center.x and Camera_Center.y are the center point coordinates of the picture taken by the onboard camera;

[0107] The rotation offset is calculated, the direction of the identifier image template is taken as the forward direction of the UAV, the rotation offset of the UAV is calculated by comparing the angle difference between the identifier image template and the identifier image in the onboard camera picture, and the calculation formula is:

[0108] Delta theta = theta H - theta

[0109] Wherein, theta H is the direction of the identifier image template, and theta is the direction of the identifier image in the onboard camera picture.

[0110] S52, real-time monitoring of alignment error, judging landing condition and completing recovery;

[0111] When the above offset reaches the allowable threshold, it is determined that the alignment is completed, and the following formula is satisfied, which can be regarded as completing the alignment.

[0112] Delta x <= sigma x , delta y <= sigma y , delta theta <= sigma θ ;

[0113] Wherein, sigma x , sigma y , sigma θ are the allowable thresholds of horizontal offset delta x, delta y and rotation offset delta theta, which are given by human, if satisfied, the recovery can be completed, if not satisfied, return to S3.

[0114] The beneficial effects of the present application are: (1) the present application considers the sea environment where the actual unmanned ship is located, and controls the unmanned machine recovery under the premise of the best speed and heading, ensures that the unmanned ship is in a relatively stable running environment when recovering the unmanned machine, and improves the success rate of recovery; (2) in the relative positioning link, the unmanned machine can obtain the position information of the unmanned ship by communication, and the relative position information of the unmanned machine-unmanned ship is obtained by using distributed cooperative technology, which further improves the positioning accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0115] Figure 1 The figure is a schematic diagram of the method steps of the present application.

[0116] Figure 2 The figure is a schematic diagram of the method flow of the present application.

[0117] Figure 3 The figure is a standard "H" identifier used in step S4 of the present application.

[0118] Figure 4 The figure is the best track diagram obtained by determining the best heading of the unmanned ship in the process of recovering the unmanned machine by the unmanned ship using the method of the present application.

[0119] Figure 5The optimal speed diagram obtained by the optimal heading determination of the unmanned ship in the process of recovering the unmanned aerial vehicle by the unmanned ship adopting the method of the application.

[0120] Figure 6 The attitude real-time monitoring curve diagram of the unmanned aerial vehicle in the process of recovering the unmanned aerial vehicle by the unmanned ship adopting the method of the application.

[0121] Figure 7 The angular velocity real-time monitoring curve diagram of the unmanned aerial vehicle in the process of recovering the unmanned aerial vehicle by the unmanned ship adopting the method of the application.

[0122] Figure 8 The three-dimensional velocity real-time monitoring curve diagram of the unmanned aerial vehicle in the process of recovering the unmanned aerial vehicle by the unmanned ship adopting the method of the application.

[0123] Figure 9 The attitude prediction trajectory diagram of the unmanned aerial vehicle in the process of recovering the unmanned aerial vehicle by the unmanned ship adopting the method of the application

[0124] Figure 10 The three-axis angular velocity prediction curve diagram of the unmanned aerial vehicle in the process of recovering the unmanned aerial vehicle by the unmanned ship adopting the method of the application.

[0125] Figure 11 The three-dimensional velocity prediction curve diagram of the unmanned aerial vehicle in the process of recovering the unmanned aerial vehicle by the unmanned ship adopting the method of the application.

[0126] Figure 12 The three-dimensional position prediction curve diagram of the unmanned aerial vehicle in the process of recovering the unmanned aerial vehicle by the unmanned ship adopting the method of the application.

[0127] Figure 13 The space position response curve diagram of the unmanned aerial vehicle in the process of recovering the unmanned aerial vehicle by the unmanned ship adopting the method of the application. DETAILED DESCRIPTION

[0128] As Figures 1-3 shown in the figure, a control method for autonomous guiding a shipborne unmanned aerial vehicle by an unmanned ship comprises the following steps:

[0129] S1, determining the optimal heading and optimal speed for executing the recovery of the unmanned aerial vehicle according to the sea environment where the unmanned ship is currently located;

[0130] S2, the unmanned ship sails according to the optimal heading and optimal speed obtained in S1, and the unmanned aerial vehicle calculates the relative position information of the unmanned ship and the unmanned aerial vehicle according to the inertial sensor and GPS carried by itself;

[0131] S3, the relative position information of the ship and the aerial vehicle is obtained through S2, the unmanned aerial vehicle is controlled to the vicinity of the parking apron of the unmanned ship, preliminary alignment is realized, and the unmanned aerial vehicle and the parking apron position are kept relatively vertical;

[0132] S4, after approaching above the unmanned ship, the unmanned aerial vehicle accurately aligns the standard mark of the unmanned ship parking apron by using the visual sensor of the unmanned aerial vehicle;

[0133] S5, the unmanned ship and the unmanned aerial vehicle are controlled to keep relative static, real-time monitoring of the alignment error is performed, it is judged whether the landing condition is met, if yes, the unmanned aerial vehicle is recovered, if not, the step S3 is returned.

[0134] In the embodiment, MATLAB 2019b is used as a simulation calculation software, a cooperative landing scene of the unmanned aerial vehicle and the unmanned ship is simulated, and the control method for autonomously guiding the shipborne unmanned aerial vehicle to be recovered by the unmanned ship is simulated.

[0135] In the embodiment, the unmanned ship is initially located at a starting position of (0, 0, 0), and the heading and speed are selected by judging the sea scene; the unmanned aerial vehicle is initially located at a starting position of (2km, 2km, 1.5m), and keeps a horizontal hovering state (attitude values: φ = 0°, θ = 0°, ψ = 0°. Angular velocity values: x-axis = 0°, y-axis = 0°, z-axis = 0°. Speed: vx = 0m / s, vy = 0m / s, vz = 0m / s).

[0136] Figure 4 The best heading curve of the unmanned ship in the process of recovering the unmanned aerial vehicle is given according to the determination method of the heading and speed of the unmanned ship in step 1. Figure 4 It can be seen that the unmanned ship starts to move at the origin (0km, 0km) of the world coordinate system, and the track is at an angle of 45° with the x-axis.

[0137] Figure 5 The best heading curve of the unmanned ship in the process of recovering the unmanned aerial vehicle is given according to the determination method of the heading and speed of the unmanned ship in step 1. Figure 5 It can be seen that the unmanned ship keeps a uniform straight-line motion in the process of recovery.

[0138] Figure 6 The real-time monitoring curve of the attitude of the unmanned aerial vehicle in the process of recovering the unmanned aerial vehicle is given, and the attitude values of three axes are given respectively. Figure 6 It can be seen that the unmanned aerial vehicle is in the initial horizontal attitude, and finally approaches the unmanned ship at 11.7s and flies vertically by implementing the method of the invention.

[0139] Figure 7 The real-time monitoring curve of the angular velocity of the unmanned aerial vehicle in the process of recovering the unmanned aerial vehicle is given, and the angular velocity curves of the x-axis, y-axis and z-axis of the unmanned aerial vehicle are given respectively. Figure 7 It can be seen that the unmanned aerial vehicle is in the initial hovering state, and finally approaches the unmanned ship at 2.52s by implementing the method of the invention.

[0140] Figure 8The three-dimensional velocity real-time monitoring curve of the unmanned plane in the process of unmanned plane recovery is given, and the curves of three-axis velocities are given respectively. Figure 8 It can be seen that after the unmanned plane vertically approaches the unmanned ship, the horizontal approach is maintained, and the unmanned plane lands on the unmanned ship after 16.82s.

[0141] Figure 9 The Kalman filter for the attitude motion prediction of the unmanned plane in step 3 of the application is described, and the predicted curves of the roll angle, pitch angle and yaw angle of the unmanned plane are dotted lines. From Figure 9 It can be seen that the attitude prediction trajectory of the unmanned plane finally overlaps with the actual value.

[0142] Figure 10 The predicted curve of the angular velocity of the unmanned plane described in step 3 of the application is given, and the predicted curves of three-axis angular velocities are given respectively. From Figure 10 It can be seen that the angular velocity of the unmanned plane overlaps with the actual value.

[0143] Figure 11 The predicted curve of the angular velocity of the unmanned plane described in step 3 of the application is given, and the predicted curves of three-axis angular velocities are given respectively. From Figure 11 It can be seen that the three-dimensional velocity of the unmanned plane finally overlaps with the actual value.

[0144] Figure 12 The predicted curve of the angular velocity of the unmanned plane described in step 3 of the application is given, and the predicted curves of three-axis angular velocities are given respectively. From

[0145] Figure 13 The three-dimensional motion trajectory curve of the unmanned ship autonomous guiding shipborne unmanned plane recovery control of the application is given, and the motion trajectory of the unmanned ship is a dotted line, and the motion trajectory of the unmanned plane is a dotted line. It can be seen from the figure that the unmanned plane finally completes landing at (0.9964km, 0.9964km).

[0146] From the simulation results described above, it can be seen that the unmanned ship autonomous guiding shipborne unmanned plane recovery control method proposed in the application can quickly and accurately recover the unmanned plane.

Claims

1. A method for autonomous guiding of a shipboard UAV by an unmanned surface vehicle, characterized in that, Comprise the following steps: S1, according to the current unmanned ship is located in the sea area environment judgment execution unmanned plane recycling best heading and best speed; Specifically, it comprises the following steps: S11, the weight assignment is carried out on the longitudinal and lateral roll motion affecting the stability of the unmanned ship; The influence degree z of the determined longitudinal and lateral roll motion on the unmanned ship navigation is calculated by the following formula: z = w1z1 + w2z2; Wherein, z1 and z2 represent the roll and pitch of the unmanned ship respectively, w1 and w2 represent the weight of the roll and pitch motion respectively; S12, statistics longitudinal and lateral motion data, and according to S11 weighting value; statistics in the sampling time T in the collection of n groups of longitudinal and lateral motion and lateral motion data, and according to S11 weighting value to obtain the comprehensive influence factor z i , i = 1, 2,..., n, while, record each z i Corresponding to the heading x i And the speed of y i ; S13, fitting the relationship between the comprehensive influence factor obtained in S12 and the heading and speed using the least square method; fitting the relationship between the z i and the heading x i and the speed y i and the time using the least square method; fitting the function value z i on the n*n rectangular point domain (x, y) composed of the collected heading x n and speed y T , x = [x1, x2,..., x n ] T , y = [y1, y2,..., y i ] l ; fixing the speed set y, constructing n least square fitting polynomials for the heading set x wherein each is constructed from the following recursion formula: Let obtain According to the least square principle Then construct the least square fitting polynomial of the speed y where each ψ l (y) is constructed from the following recursion formula, l = 0, 1,..., q - 1: According to the least square principle Finally, the fitting polynomial of the longitudinal and lateral roll influence value and the heading and speed of the unmanned ship is obtained S14, according to the fitting result of S13, the minimum heading and speed of the longitudinal and lateral roll are obtained; The minimum point of the function f(x,y) of S13 fitting is the best heading and speed; S2, the unmanned ship sails according to the best heading and best speed obtained by S1, and the unmanned plane calculates the relative position information of the unmanned ship and the unmanned plane according to the inertial sensor and GPS carried by itself; S3, the relative position information of the ship and the plane is obtained through S2, the unmanned plane is controlled to the vicinity of the unmanned ship parking apron, the preliminary alignment is realized, and the unmanned plane and the parking apron position are kept relatively vertical; S4, after the unmanned plane approaches the upper part of the unmanned ship, the visual sensor of the unmanned plane is used to accurately align the standard mark of the unmanned ship parking apron; S5, control the unmanned ship and the unmanned plane to keep relatively static, real-time monitor the alignment error, judge whether the landing condition is met, if yes, complete the unmanned plane recycling, if not, return to S3.

2. The unmanned surface vehicle autonomous guidance of shipboard unmanned aircraft recovery control method of claim 1, wherein, In step S2, the unmanned ship sails according to the best heading and best speed obtained by S1, and the unmanned plane calculates the relative position information of the unmanned ship and the unmanned plane according to the inertial sensor and GPS carried by itself, which comprises the following steps: S21, the unmanned plane obtains the current attitude information of the machine body through the inertial measurement unit; The attitude quaternion k moment data is obtained by integrating the angular velocity w(k-1) at the k-1 moment obtained by the inertial measurement unit Specifically as follows: Wherein, T represents the sampling time, and the attitude matrix E(q(k-1)) is represented as follows: where I3is a three-dimensional identity matrix, Matrix S(q v (k-1)) is expressed as follows: S22, the rotation matrix of the navigation coordinate system and the airborne coordinate system is calculated; The rotation matrix is obtained from the attitude data at k time obtained by S21, and is specifically as follows: Wherein, S23, the position, attitude and speed of the unmanned plane in the navigation coordinate system are calculated; The position information p(k) and speed information v(k) of the unmanned plane in the navigation coordinate system are obtained through GPS, and the angular velocity information is obtained through the rotation matrix calculated by S22, which is specifically as follows: w n (k) = R(q(k))w(k); where w n (k) is the angular velocity of the UAV in the navigation frame.

3. The unmanned surface vehicle autonomous guidance of shipboard unmanned aircraft recovery control method of claim 1, wherein, In step S3, the relative position information of the ship and the plane is obtained through S2, the unmanned plane is controlled to the vicinity of the unmanned ship parking apron, the preliminary alignment is realized, and the unmanned plane and the parking apron position are kept relatively vertical, which comprises the following steps: S31, the real-time attitude and position data of the current unmanned plane are used to predict the position information of the unmanned plane in the navigation coordinate system by using Kalman filter; The position information of the unmanned plane at k+1 time is predicted by using Kalman filter, including quaternion, speed and position as follows: X(k + 1) = [q(k + 1), v(k + 1), p(k + 1)] T Wherein, q(k+1) is the quaternion attitude value at k+1 time, v(k+1) is the velocity value at k+1 time, p(k+1) is the three-dimensional space position at k+1 time The state prediction equation is X(k+1)=FX(k); Wherein, F is the state transition matrix, as follows I3 is a three-dimensional unit matrix; the covariance P(k+1) propagation matrix is: P(k+1) = FP(k) F T + Q; Wherein, Q is the state noise covariance matrix; the measurement equation is obtained by using GPS data: z(k+1)=HX(k+1)+R; where z(k + 1) = [v(k + 1), p(k + 1)] T R is the measurement noise covariance and H is the measurement matrix, as follows Then the update prediction equation is as follows: S32, the relative position information obtained by S2 is used to preset the trajectory of the unmanned aerial vehicle approaching the unmanned ship; The position point of the unmanned ship in the navigation coordinate system at the current time is P obtained by GPS t (k), the unmanned aerial vehicle position prediction information P(k+1) obtained by S31, and planning the unmanned aerial vehicle track, specifically by setting n track points (x i ,y i ) satisfying: S33, the unmanned aerial vehicle moves to the upper side of the unmanned ship platform according to the preset trajectory.

4. The unmanned surface vehicle autonomous guidance of shipboard unmanned aircraft recovery control method of claim 1, wherein, In step S4, after the unmanned aerial vehicle approaches the upper side of the unmanned ship, the visual sensor of the unmanned aerial vehicle is used to accurately aim at the standard identifier of the unmanned ship parking apron, which specifically includes the following steps: S41, the visual sensor carried by the unmanned aerial vehicle is subjected to distortion correction to reduce the distortion of the collected image caused by the defects of the lens of the camera lens; Consider a point P in the world coordinate system with three-dimensional spatial coordinates (X w ,Y w ,Z w ), which is converted into a point P in the camera coordinate system (X C ,Y C ,Z C ) using a rotation matrix R 3×3 and a translation matrix T 3×1 . The conversion relationship is shown in the following formula: The camera intrinsic matrix K is obtained by camera calibration, and P is converted into the pixel coordinate system of the image as follows: wherein Z C is a scale ratio, given by the camera manufacturer; [u v 1] T is a pixel coordinate of the image, f x , f y is a focal length, and u0, v0 are distortion factors. The three-dimensional coordinates of the vertex space of the identification code are P i = [X i , Y i , Z i ] T The pixel point projected onto the image is p i = [u i , v i ] T The conversion relationship between the space point and the pixel point is obtained where z i is the depth information, which is converted into a matrix form as z i p i = K [RP i + T]; The optimal pose R, T of the camera is solved by iteratively solving the following equation, that is, the distortion correction is completed; S42, the angle of the unmanned aerial vehicle camera is kept unchanged, the image taken by the unmanned aerial vehicle is preprocessed, and the pixel point coordinates of the identifier image position are obtained; The "H" identifier is selected as the parking identifier, which is surrounded by a black circle on a white background, has distinctive features, and is easy to be recognized by the camera; during the identification process, the unmanned aerial vehicle camera is controlled to be at -90 degrees pitch and 0 degrees azimuth; The picture of the parking identifier taken by the camera is subjected to weighted grayscale processing; for an image represented by an RGB three-channel color system, the coordinates of a single image pixel are (x, y), the values corresponding to the red, green and blue channels of the pixel are R(x, y), G(x, y) and B(x, y), and the weighted grayscale value Gray(x, y) is obtained as follows: Gray(x, y) = w R R(x, y) + w G G(x, y) + w B B(x, y); where w R , w G , w B are weights for the corresponding channels; The processed image is subjected to noise reduction by using a local adaptive thresholding method, and the adaptive thresholding method formula is as follows: wherein, S is the variance of the gray scale image, xy is the filter window, S is the variance and mean of the pixels within the filter window, is the filtered pixel gray scale value; The noise-reduced image is binarized by using the grayscale threshold method; first, a threshold T is determined, and then the image is binarized according to the following formula: G(x, y) is the pixel gray value before processing, G(x, y) is the pixel gray value before processing, S43, a template-based matching method is used to detect the identifier image; the unmanned aerial vehicle has a built-in standard "H" identifier matching template, which matches the identifier by comparing the correlation of the template and the region in the identifier image. The size of the captured image S is M x N, and the size of the identifier image template g to be matched in the image is m x n, S xy For the detection window, the correlation coefficient matrix p(x, y) is calculated: where σ(S xy , g) is the covariance of the detection window and the template g, D xy , D is the variance of the detection window and the template g, and the correlation coefficient satisfies |p(x, y)|≤1. The identifier image is detected by selecting the region with the maximum correlation coefficient in the template and the region in the image, and the position of the identifier image satisfies the following equation: Wherein, (x, y) is the pixel point coordinate of the identifier image.

5. The unmanned surface vehicle autonomous guidance of shipboard unmanned aircraft recovery control method of claim 1, wherein, In step S5, the unmanned ship and the unmanned aerial vehicle are controlled to be relatively stationary, the alignment error is monitored in real time, it is judged whether the landing condition is met, if yes, the unmanned aerial vehicle is recovered, and if not, the process returns to S3, which specifically includes the following steps: S51, according to the identifier pixel point coordinates obtained in S4, the horizontal offset and the rotational offset are calculated; According to the identifier image coordinates (x, y) given in step S4, the horizontal offset Δx and Δy of the unmanned aerial vehicle are solved, and the calculation formula is as follows: Wherein, Camera_Center.x, Camera_Center.y are the coordinates of the center point of the picture taken by the airborne camera; The rotation offset is calculated, the direction of the identifier image template is taken as the forward direction of the unmanned aerial vehicle, the rotation offset of the unmanned aerial vehicle is calculated by comparing the angle difference between the identifier image template and the identifier image in the picture of the airborne camera, and the calculation formula is: Δθ = θ H -θ; where θ H is the direction of the identifier image template, and θ is the direction of the identifier image in the onboard camera picture. S52, real-time monitoring of alignment error, judging landing condition and completing recovery; When the above offset reaches the allowable threshold, it is determined that the alignment is completed, and the following formula is considered to be completed alignment; Δx < σ x , Δy < σ y , Δθ < σ θ ; where σ x ,σ y ,σ θ are the allowed thresholds for the horizontal offsets Δx, Δy, and the rotational offset Δθ, respectively, given by the human, and if they are met, the recycling is completed, otherwise the process returns to S3.

Images