A visual guidance method for dynamic recovery of unmanned surface vehicle

Abstract

The application provides a visual guidance method for dynamic recovery of an unmanned ship, which detects a visual guidance beacon arranged on a mother ship, combines the heading and attitude data of the mother ship and the unmanned ship, calculates the lateral and longitudinal distances between the unmanned ship and the mother ship, and automatically adjusts the target track of the unmanned ship, so that the unmanned ship completes track adjustment by using a track following algorithm and finally keeps at a predetermined recovery position. The application makes the unmanned ship autonomously keep the relative position relationship with the mother ship by using the visual guidance. The application autonomously controls the unmanned ship to enter the recovery position of the mother ship without manual participation and without interrupting the navigation task of the mother ship. The visual guidance is low in implementation cost, is not affected by external electromagnetic interference, and can normally work in an electromagnetic silence or electromagnetic interference countermeasure environment.

Description

Technical Field

[0001] This invention relates to visual guidance for the dynamic recovery of unmanned surface vessels (USVs), and is suitable for recovering USVs without affecting the normal navigation of the mother ship. Background Technology

[0002] The recovery of unmanned surface vessels (USVs) can be categorized into two methods based on the mother ship's status: static recovery and dynamic recovery. Static recovery requires the mother ship to cease navigation and remain almost stationary to facilitate the recovery of the USV; while dynamic recovery allows the mother ship to continue its navigation operations without needing to dock. Therefore, dynamic recovery is significant for improving the efficiency of mother ship operations. Furthermore, compared to static recovery, dynamic recovery offers a wider range of adjustable water speed for the USV, ensuring greater maneuverability.

[0003] During dynamic recovery, the unmanned surface vessel (USV) needs to maintain a stable relative position with the mother ship, i.e., enter the predetermined recovery position to facilitate the establishment of a flexible or rigid connection between the mother ship and the USV, thereby ultimately completing the recovery of the USV. During dynamic recovery, the deviation between the actual position of the USV and the recovery position needs to be controlled within a small range (generally requiring an error of less than 1 meter). If this range is exceeded, either the USV will collide with the mother ship, or the recovery will fail. Since the positioning accuracy of satellite positioning systems (such as BeiDou and GPS) is at the meter level, directly using satellite positioning cannot achieve precise control of the relative position between the USV and the mother ship. Although the positioning accuracy of high-precision satellite differential positioning meets the requirements, satellite differential positioning relies on a reference station and cannot be used anytime and anywhere, therefore it is not suitable for application in scenarios where the USV is recovered from the mother ship.

[0004] Currently, unmanned surface vessels (USVs) are generally controlled manually to reach their recovery location. During remote control, operators typically need to visually observe the USV from a suitable position outside the mother ship and manipulate it in real time based on its status and relative position to the mother ship. Manual remote control is technically challenging and has a low success rate. Furthermore, in high sea states, there are significant risks to the personal safety of operators. Therefore, there is an urgent need to develop an automated USV guidance method that can be used anytime and anywhere. Summary of the Invention

[0005] In view of this, this invention proposes a visual guidance method for the dynamic recovery of unmanned surface vessels (USVs). By detecting visual guidance beacons deployed on a mother ship and combining the attitude data of both the mother ship and the USV, the lateral and longitudinal distances between the USV and the mother ship are calculated. This allows the USV to automatically adjust its target trajectory. The USV then uses a trajectory tracking algorithm to complete the trajectory adjustment and ultimately maintain its position at the predetermined recovery location. This invention enables the USV to autonomously maintain its relative position to the mother ship through visual guidance.

[0006] The present invention is implemented as follows:

[0007] A visual guidance method for dynamic recovery of unmanned surface vessels (USVs) is proposed. The USV is equipped with a guidance imaging sensor for real-time imaging of a visual guidance beacon. The imaging sensor calibrates the visual guidance beacon, constructing a matrix D relating longitudinal distance to pixel width, where the longitudinal distance refers to the longitudinal distance y between the USV and the visual guidance beacon. i pixel width w i The longitudinal distance between the unmanned surface vessel and the visual guidance beacon is y. i At that time, the pixel width occupied by the visual guidance beacon in the image formed by the imaging sensor is represented by each row of matrix D as (y i ,w i Using the relationship matrix D, the longitudinal distance between the unmanned surface vessel (USV) and the mother ship is calculated in real time. Using a pinhole imaging model, combined with the attitude data of both the mother ship and the USV, the lateral distance between the USV and the mother ship is calculated in real time. The lateral distance is then superimposed on the current heading. The deviation from the predetermined lateral distance d0 for recovery is used as the target heading angle of the unmanned surface vessel at the next moment, and the current longitudinal distance is added to the current speed. Longitudinal distance from the planned recovery The deviation is used as the target speed for the unmanned surface vessel in the next moment, and the speed is continuously updated to complete the trajectory adjustment and finally maintain the predetermined recovery position.

[0008] Furthermore, visual guidance beacons are deployed on the sides or stern of the mother ship.

[0009] Furthermore, the calculation process for the longitudinal distance between the unmanned surface vessel and the mother ship at the current moment is as follows:

[0010] Obtain the pixel width w of the visual guidance beacon in the image captured by the unmanned surface vessel's guidance imaging sensor at the current moment, and the horizontal coordinate x of the center of the visual guidance beacon in the image;

[0011] Find the two elements W1 and W2 that are numerically closest to w in the relation matrix D, and their corresponding vertical distances Y1 and Y2;

[0012] Using (Y1, W1) and (Y2, W2), we can obtain a linear relationship between the vertical distance and the pixel width occupied by the visual guide beacon in the image. Then, we can use this linear relationship to obtain the vertical distance corresponding to the pixel width w at the current moment.

[0013] Furthermore, the calculation process for the lateral distance between the unmanned surface vessel and the mother ship at the current moment is as follows:

[0014] Based on the current longitudinal distance between the unmanned surface vessel and the mother ship Let the first right angle be the lateral distance between the unmanned surface vessel and the mother ship at the current moment. Construct a right triangle with leg 2 as the second leg; based on the pinhole imaging model, estimate the angle corresponding to leg 2, and use trigonometric functions to obtain the lateral distance between the unmanned surface vessel and the mother ship at the current moment.

[0015] Furthermore, the angles corresponding to the second leg of the right triangle Where w is the pixel width occupied by the visual guidance beacon in the image formed by the unmanned surface vessel's guidance imaging sensor at the current moment, x is the horizontal coordinate of the center of the visual guidance beacon in the image formed at the current moment, a1 is the heading angle of the unmanned surface vessel at the current moment, and a2 is the heading angle of the mother ship at the current moment.

[0016] Furthermore, the target heading angle at the next moment The calculation method is as follows:

[0017]

[0018] Where, k a Here, is the target heading control parameter, and atan is the arctangent function, used to suppress the rate of change of the target heading angle when the lateral distance deviation is large.

[0019] Furthermore, the target speed of the unmanned surface vessel at the next moment. The calculation method is as follows:

[0020]

[0021] Where, k v For control parameters. v is the current speed of the unmanned surface vessel.

[0022] Furthermore, a visual guidance beacon is created using a circular graphic, with the circular graphic being black and white and the width ratio of the circular graphic being 1:1:3:1:1.

[0023] Beneficial effects

[0024] (1) High autonomy: This invention controls the unmanned surface vessel to enter the recovery position of the mother ship in an autonomous manner without human intervention.

[0025] (2) High efficiency: This invention enables the mother ship to recover the unmanned surface vessel while it is underway, without interrupting the mother ship's navigation mission.

[0026] (3) Strong applicability: This invention can be applied to both side recovery and dock recovery methods.

[0027] (4) The present invention adopts visual guidance, has low implementation cost, and is not affected by external electromagnetic interference. It can work normally in environments such as electromagnetic silence or electromagnetic interference. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of an optical beacon;

[0029] Figure 2 This is a schematic diagram of the deployment of visual guidance beacons during side recovery in the embodiment;

[0030] Figure 3 This is a schematic diagram of the deployment of visual guidance beacons during dock recovery;

[0031] Figure 4 This is a schematic diagram illustrating the principle of lateral distance estimation;

[0032] Figure 5 This is a schematic diagram of the visual guidance beacon in the embodiment;

[0033] Figure 6 This is a schematic diagram of the target heading in the embodiment; Detailed Implementation

[0034] (1) Create visual guidance beacons:

[0035] A visual guidance beacon is created using a circular graphic, with the circular graphic being black and white and the width ratio of the circular ring being 1:1:3:1:1. Figure 1 As shown. The circular graphic is printed on a flat surface, and the overall side length of the printed circular image is L, typically 1 to 2 meters. Visual guidance beacons can also be constructed using light sources such as LEDs, as long as the image proportions and size meet the above constraints.

[0036] In this embodiment, a 1.5m visual guidance beacon is fabricated on a square whiteboard with a side length of 2m, such as... Figure 5 As shown.

[0037] (2) Imaging calibration:

[0038] To reduce the impact of imaging distortion on guidance and simultaneously decrease the computational load during real-time processing, this invention performs imaging calibration, establishing a correspondence between the longitudinal distance y between the unmanned surface vessel (USV) guidance imaging sensor and the visual guidance beacon, and the pixel width w occupied by the visual guidance beacon in the image captured by the USV guidance imaging sensor.

[0039] Let the predetermined initial longitudinal distance between the unmanned surface vessel (USV) guidance imaging sensor and the visual guidance beacon be y0, and the predetermined recovery longitudinal distance between the USV guidance imaging sensor and the visual guidance beacon be y0. Position the unmanned surface vessel's (USV) guidance imaging sensor at point y0, pointing it toward the visual guidance beacon, and save the captured image I0. Obtain the pixel width w0 occupied by the visual guidance beacon in image I0. Then, position the USV's guidance imaging sensor at point y0 in steps s (typically 1m). i=y0-si, and facing the visual guidance beacon, save the captured image I. i , acquire image I i The pixel width w occupied by the visual guidance beacon i ; final acquisition The corresponding pixel width

[0040] Generate distance y i With pixel width w i The correspondence matrix D is M rows and 2 columns. (representing rounding up), each row of matrix D can be represented as (y i ,w i ).

[0041] In this embodiment, the predetermined initial longitudinal distance y0 between the unmanned surface vessel's guidance imaging sensor and the visual guidance beacon is 50m, and the predetermined retrieval longitudinal distance is... The distance is 5m, which means that the unmanned surface vessel should begin visual guidance 50m behind the mother ship and be recovered 5m from the guidance beacon.

[0042] The unmanned surface vessel's guidance imaging sensor was placed at a distance of 50m and aimed at the visual guidance beacon to acquire an image. The visual guidance beacon occupied a pixel width of 49 pixels in the image.

[0043] Then, the unmanned surface vessel's guidance imaging sensor is positioned in 1-meter increments at the y-axis. i =50-0.5i, and facing the visual guidance beacon, acquire image I. i The pixel width w occupied by the visual guidance beacon i If i is 10, the pixel width w 10 It is 62;

[0044] The pixel width at 5m The value is 498.

[0045] Finally, using the above data, a 46-row, 2-column correspondence matrix D is generated, as shown in the table below.

[0046] 50 49 49 51 48 52 47 53 46 54 45 55 ··· ··· 18 139 17 147 ··· ··· 5 498

[0047] (3) Deploy visual guidance beacons:

[0048] Deploy visual guidance beacons on the mother ship. For mother ships recovering from the side, the deployment method of the visual guidance beacons is as follows: Figure 2 As shown; for mother ships recovering from dock landings, the deployment method of visual guidance beacons is as follows: Figure 3 As shown. During deployment, the normal to the visual guidance beacon plane must maintain a high degree of parallelism with the mother ship's axis, generally requiring an angle of less than 10°.

[0049] Taking hull-side recovery as an example, the visual guidance beacon is 5m from the mother ship's side, and the angle between the normal to the visual guidance beacon's plane and the mother ship's axis is 5°. If the predetermined recovery position is 4m from the mother ship's side, then the predetermined lateral recovery distance d0 = 1m. For mother ships recovering in the dock, the visual guidance beacon is deployed at the stern of the mother ship.

[0050] (4) Estimate the longitudinal and lateral distances between the unmanned surface vessel's guidance imaging sensor and the visual guidance beacon at the current moment:

[0051] Based on the positioning and heading attitude information of the mother ship and itself, the unmanned surface vessel (USV) enters the predetermined starting position guided by vision. The maximum allowable positioning error for both the USV and the mother ship is 30 meters.

[0052] The longitudinal distance is estimated as follows:

[0053] Let the lateral field of view of the guidance imaging sensor on the unmanned surface vessel be α, and the number of pixels in the image be (W, H). For the image captured by the guidance imaging sensor, a visual guidance beacon is detected using a row-scan detection method, and the pixel width w of the visual guidance beacon and the x-coordinate of the beacon's center in the image are calculated. The value closest to w is then found in the second column of matrix D. i Let y be the corresponding element W1 and W2. i Let Y1 and Y2 be the two values.

[0054] In the short term, it is assumed that there is a linear relationship between the longitudinal distance between the unmanned surface vessel (USV) guidance imaging sensor and the visual guidance beacon and the pixel width w occupied by the visual guidance beacon in the image captured by the USV guidance imaging sensor. Therefore, the longitudinal distance between the USV guidance imaging sensor and the visual guidance beacon at the current moment is estimated using the following formula:

[0055]

[0056] The lateral distance is estimated as follows:

[0057] Let the heading angles of the unmanned surface vessel and the mother ship at the current moment be a1 and a2, respectively. Figure 4 As shown. The lateral distance between the unmanned surface vessel's guidance imaging sensor and the visual guidance beacon at the current moment is:

[0058] d=y×tan(∠1) (2)

[0059] Figure 4 In the image coordinate system, a2' / / a2, o is the origin of the imaging sensor coordinate system, ∠1=∠AOC=∠AOB+∠BOC, and ∠AOB is calculated in the image coordinate system as follows: At this point, a negative value for ∠1 indicates that the visual guidance beacon is on the left side of the imaging sensor; conversely, a positive value indicates that the visual guidance beacon is on the right side of the imaging sensor.

[0060] According to the pinhole imaging model, regardless of whether it is positive or negative, ∠1 can be uniformly estimated by the following formula:

[0061]

[0062] Therefore, the lateral distance between the unmanned surface vessel's guidance imaging sensor and the visual guidance beacon at the current moment is finally estimated using the following formula:

[0063]

[0064] In this embodiment, the lateral field of view of the guidance imaging sensor on the unmanned surface vessel is 30°, and the number of pixels in the imaging image is (1920, 1080).

[0065] Based on the positioning and heading attitude information of the mother ship and itself, the unmanned surface vessel (USV) enters the predetermined starting position for visual guidance. Due to the positioning errors of the USV and the mother ship, the USV is actually 55m away from the mother ship. At this time, the pixel width of the visual guidance beacon in the image formed by the guidance imaging sensor is 45, and the center of the visual guidance beacon is at the horizontal coordinate of 916. The two elements in the second column of matrix D that are closest in value to 45 are found to be 49 and 51, corresponding to y... i The values ​​are 50 and 49.

[0066] Estimate the distance between the unmanned surface vessel and the visual guidance beacon at the current moment:

[0067]

[0068] At the current moment, the headings of the unmanned surface vessel (USV) and the mother ship are 14° and 12° respectively. Estimate the lateral distance between the USV and the mother ship at the current moment:

[0069]

[0070] (5) Guidance:

[0071] The target heading angle and target speed of the unmanned surface vessel (USV) are calculated and continuously updated for the next moment. The USV then autonomously navigates by following a track until it is captured by the mother ship's recovery system.

[0072] By adjusting the unmanned surface vessel's (USV) course, it is guided to autonomously navigate to a position with a lateral distance equal to the predetermined recovery lateral distance d0. Specifically, this involves superimposing the current lateral distance onto the current course. The deviation from the predetermined lateral distance d0 for recovery is taken as the target heading angle of the unmanned surface vessel at the next moment, and is calculated as follows:

[0073]

[0074] Where, k a is the target heading control parameter, with a value range of [0.01, 5]; atan is the arctangent function, used to suppress the rate of change of the target heading angle when the lateral distance deviation is large, thereby ensuring the navigation stability and safety of the unmanned surface vessel.

[0075] By adjusting the speed of the unmanned surface vessel (USV), it is guided autonomously to a longitudinal distance equal to the predetermined recovery longitudinal distance. The state is determined by adding the current longitudinal distance to the current speed. Longitudinal distance from the planned recovery The deviation is used as the target speed of the unmanned surface vessel at the next moment, and the calculation method is as follows:

[0076]

[0077] Where k v This is a control parameter, and its value range is [0.001, 1].

[0078] Update the target speed and target heading angle of the unmanned surface vessel (USV). The USV then navigates autonomously in a fixed heading manner according to the target speed and heading angle. Repeat steps (4) and (5) until the USV is captured by the mother ship's recovery device.

[0079] In this embodiment, the horizontal distance at the current moment is used. And the predetermined lateral distance for recovery d0 = 1, target heading control parameter k a Taking a value of 2 generates the target heading of the unmanned surface vessel at the next moment, such as... Figure 6 The above,

[0080]

[0081] Using the current longitudinal distance between the unmanned surface vessel and the visual guidance beacon Given a mother ship speed of 10 knots and a control parameter k value of 0.1, calculate the target speed of the unmanned surface vessel.

[0082]

[0083] Update the target heading and target speed of the unmanned surface vessel (USV). The USV will autonomously navigate in a heading-maintaining manner. Subsequently, the USV will catch up with the mother ship and increase the lateral distance between them. Repeat steps (4) and (5) continuously. The USV will automatically maintain the predetermined recovery position on the side of the mother ship until it is captured by the recovery device, at which point visual guidance ends.

Claims

1. A visual guidance method for dynamic recovery of unmanned surface vehicle, wherein a guidance imaging sensor is arranged on the unmanned surface vehicle to perform real-time imaging on a visual guidance beacon, and the method comprises the following steps: The visual guidance beacon is imaged and calibrated using an imaging sensor, and a matrix D relating longitudinal distance to pixel width is constructed, where longitudinal distance refers to the longitudinal distance y between the unmanned surface vessel and the visual guidance beacon. i pixel width w i The longitudinal distance between the unmanned surface vessel and the visual guidance beacon is y. i At that time, the pixel width occupied by the visual guidance beacon in the image formed by the imaging sensor is represented by each row of matrix D as (y i ,w i Using the relationship matrix D, the longitudinal distance between the unmanned surface vessel (USV) and the mother ship is calculated in real time. Using a pinhole imaging model, combined with the attitude data of both the mother ship and the USV, the lateral distance between the USV and the mother ship is calculated in real time. The lateral distance is then superimposed on the current heading. The deviation from the predetermined lateral distance d0 for recovery is used as the target heading angle of the unmanned surface vessel at the next moment, and the current longitudinal distance is added to the current speed. Longitudinal distance from the planned recovery The deviation is used as the target speed for the unmanned surface vessel in the next moment, and the speed is continuously updated to complete the trajectory adjustment and finally maintain the predetermined recovery position. ​ The calculation process of the longitudinal distance between the unmanned ship and the mother ship at the current time is as follows, Obtain the pixel width w of the visual guide beacon in the image formed by the unmanned ship guide imaging sensor at the current time, and the horizontal coordinate x of the center of the visual guide beacon in the image; Find two elements W1 and W2 in the relationship matrix D that are closest to w in value, and the corresponding longitudinal distances Y1 and Y2; A linear relationship between the longitudinal distance and the pixel width of the visual guidance beacon in the image is obtained using (Y1, W1), (Y2, W2), and the longitudinal distance corresponding to the pixel width w at the current time is obtained through the linear relationship The calculation process of the lateral distance between the unmanned ship and the mother ship at the current time is as follows, a longitudinal distance between the unmanned vehicle and the mother ship at the current time a lateral distance between the unmanned vehicle and the mother ship at the current time a right angle side two, a right triangle is constructed; according to a pinhole imaging model, an angle corresponding to the right angle side two is estimated, and a trigonometric function relationship is used to obtain the lateral distance between the unmanned vehicle and the mother ship at the current time angle of the two opposite right sides wherein w is the pixel width of the visual guide beacon in the image taken by the unmanned vehicle guide imaging sensor at the current time, x is the horizontal coordinate of the center of the visual guide beacon in the image taken at the current time, a1 is the heading angle of the unmanned vehicle at the current time, and a2 is the heading angle of the mother ship at the current time. 2.The visual guidance method for dynamic recovery of unmanned surface vehicle according to claim 1, wherein: The visual guide beacon is arranged on the side or tail of the mother ship.

3. The visual guidance method for dynamic recovery of unmanned surface vehicle according to claim 1 or 2, characterized in that: Next time target heading angle The calculation is as follows: where k a is a target heading control parameter, and atan is an arctangent function used to suppress the rate of change of the target heading angle when the lateral distance deviation is large.

4. The visual guidance method for dynamic recovery of unmanned surface vehicle according to claim 3, wherein: Next time unmanned ship target speed The calculation method is: where k v is a control parameter and v is the current speed of the USV.

5. The visual guidance method for dynamic recovery of unmanned surface vehicle according to claim 4, wherein: The visual guide beacon is made of a circular ring pattern, and the circular ring pattern adopts black and white pattern, and the circular ring width ratio is 1:1:3:1:1.

Images