An acoustic-optical fusion guidance system and method for AUV underwater dynamic docking

Abstract

The present application provides an acoustic-optical fusion guidance system and method for underwater dynamic docking of an autonomous underwater vehicle (AUV), belonging to the field of autonomous underwater robot docking technology. The system comprises a first inertial navigation system and a second inertial navigation system for measuring the absolute position, attitude, and velocity of the AUV and the docking station, respectively; a USBL array and USBL beacon, a multi-beam forward-looking imaging sonar, and a camera for measuring the AUV's measurement information at different guidance stages; and a combined filtering module for filtering and fusion using a federated Kalman filter and a maneuvering observer based on the relative motion state at the initial moment and the actual measurement information of the AUV obtained by the real-time USBL array, multi-beam forward-looking imaging sonar, and camera to obtain global state estimation information of the AUV. Compared with traditional estimation methods, the present application has higher motion state estimation accuracy and better robustness, making it more suitable for underwater dynamic docking.

Description

Technical Field

[0001] The present application belongs to the field of autonomous underwater robot docking technology, and more specifically, relates to an acoustic-optical fusion guidance system and method for AUV underwater dynamic docking. Background Art

[0002] Compared with manned equipment, autonomous underwater vehicles (AUVs) have obvious advantages in operational efficiency and safety. They can autonomously perform tasks such as marine environment monitoring, seabed topography surveys, and resource development in waters over a wider range and at greater depths.

[0003] To expand the operational range of AUVs, improve their efficiency, and meet the demands of wide-area seabed topography exploration, information collection, and specialized operations, they require regular AUV recovery for energy refueling. However, the complex and unique underwater environment presents significant challenges for AUV underwater docking and guidance, hindering their recovery.

[0004] Acoustic positioning is commonly used to guide AUVs underwater. However, in shallow water, sensor errors, random noise, and reverberation can lead to a large number of outliers in the measurement data, reducing its reliability. In dynamic docking scenarios, the success rate of AUV docking is particularly dependent on the guidance system, specifically the quality of navigation and positioning. However, existing guidance methods are insufficient to meet the high-precision positioning requirements of AUVs, making them unsuitable for dynamic docking. Summary of the Invention

[0005] In response to the defects of the existing technology, the purpose of this application is to provide an acoustic-optical fusion guidance system and method for AUV underwater dynamic docking, aiming to solve the problem that the existing guidance methods are insufficient to meet the high-precision positioning requirements of AUV and are not conducive to dynamic docking.

[0006] To achieve the above-mentioned objectives, the present application provides an acoustic-optical fusion guidance system for underwater dynamic docking of an AUV, comprising: a USBL array, a USBL beacon, a first inertial navigation system, a second inertial navigation system, a multi-beam forward-looking imaging sonar, a camera, a data initialization module, and a combined filtering module;

[0007] The USBL beacon and the first inertial navigation system are on board the AUV; the USBL array, the second inertial navigation system, the multi-beam forward-looking imaging sonar, and the camera are set up on the docking station;

[0008] The first and second inertial navigation systems are used to measure the absolute position, attitude, and velocity of the AUV and the docking station, respectively. The USBL array and USBL beacon, multi-beam forward-looking imaging sonar, and camera are used to measure the actual measurement information of the AUV during different guidance phases.

[0009] The data initialization module is used to convert the absolute position, attitude and velocity of the AUV and the docking station at the initial moment into the docking station coordinate system to obtain the relative motion state at the initial moment;

[0010] The combined filtering module is used to obtain the AUV global state estimation information by using the federated Kalman filter and the maneuvering observer for filtering fusion based on the relative motion state at the initial moment and the actual measurement information of the AUV obtained by the real-time USBL array, multi-beam forward-looking imaging sonar and camera;

[0011] The maneuver observer is used to obtain the relative maneuver estimation value based on the estimated relative motion state and the state transfer equation, and compensate the relative maneuver estimation value into the federated Kalman filter.

[0012] Further preferably, the combined filtering module includes a first sub-filter, a second sub-filter, a third sub-filter, a main filter and a maneuvering observer;

[0013] The first sub-filter, the second sub-filter and the third sub-filter are connected to the USBL array, the multi-beam forward-looking imaging sonar and the camera respectively; the main filter is connected to the first inertial navigation system and the second inertial navigation system; and a maneuvering observer is connected to the main filter side;

[0014] The first sub-filter, the second sub-filter, the third sub-filter and the main filter are used to calculate the state transfer matrix, the motor drive matrix, Time relative maneuver estimate and -1 moment global optimal estimation of relative motion state, obtain k Predict the state at any moment; and use it to -1 time to estimate the relative motion state covariance matrix, process noise and state transfer matrix, and obtain the corresponding k The moment-to-moment predicted state covariance matrix; wherein the global optimal estimated relative motion state at the initial moment is the relative motion state at the initial moment;

[0015] The first sub-filter, the second sub-filter and the third sub-filter are respectively used to k Predict the status at all times, update the predicted measurement values ​​corresponding to USBL beacons and USBL arrays, multi-beam forward-looking imaging sonars and cameras, and combine k The actual measurement values ​​of the AUV of the USBL beacon and USBL array, multi-beam forward-looking imaging sonar and camera at all times, as well as the Kalman filter gain and k Predict the state covariance matrix at each moment and update the corresponding k Estimated relative motion state and estimated relative running state covariance matrix at each moment;

[0016] The main filter is used to obtain the corresponding k The relative motion state and the estimated relative motion state covariance matrix are combined with the corresponding main filter k The moment-by-moment prediction state covariance matrix and k The moment prediction state matrix is ​​calculated k AUV global state estimation information at the moment;

[0017] in, k ≥1.

[0018] Further preferably, the USBL array and the USBL beacon are used to measure the actual measurement information of the AUV in the long-range guidance stage; in the medium-range guidance stage, the multi-beam forward-looking imaging sonar is used as the main instrument, and the USBL array and the USBL beacon are used as auxiliary instruments to measure the actual measurement information of the AUV; in the short-range guidance stage, the camera is used as the main instrument, and the multi-beam forward-looking imaging sonar, the USBL array and the USBL beacon are used as auxiliary instruments to measure the actual measurement information of the AUV;

[0019] Among them, the area between the AUV and the dock is divided into long-range guidance stage, medium-range guidance stage and short-range guidance stage; the area larger than the range of the multi-beam forward-looking imaging sonar and the camera is divided into the long-range guidance stage; the area larger than the camera range and smaller than the beam forward-looking imaging sonar working range is divided into the medium-range guidance stage; the area less than or equal to the camera range is divided into the short-range guidance stage.

[0020] More preferably, k Momentary prediction status for: ;in, is the state transition matrix, 3 3-unit array; It is the maneuver drive matrix; is the sampling time; for The global optimal estimate of the relative motion state at all times, for Time relative maneuver estimate; ;in, for Auxiliary variables at time, is the observation gain matrix, is the auxiliary gain matrix, for The pseudo-inverse matrix of .

[0021] Further preferably, the predicted state covariance matrix for: ;in, for Moment i The sub-filter or main filter estimates the relative motion state covariance; For the i Process noise corresponding to the sub-filter or main filter; represents the main filter; is the state transition matrix.

[0022] In a second aspect, based on the above-mentioned acoustic-optical fusion guidance system for underwater dynamic docking of an AUV, the present application provides a corresponding acoustic-optical fusion guidance method for underwater dynamic docking of an AUV, comprising the following steps:

[0023] Step S1: Convert the absolute position, attitude, and velocity of the AUV and the docking station at the initial moment into the docking station coordinate system to obtain the relative motion state at the initial moment;

[0024] Step S2: Based on the relative motion state at the initial moment and the actual measurement information of the AUV obtained by the real-time USBL array, multi-beam forward-looking imaging sonar and camera, a federated Kalman filter and a maneuvering observer are used for filtering fusion to obtain the AUV global state estimation information;

[0025] Step S3: using the AUV global state estimation information to guide the AUV movement;

[0026] Among them, a maneuvering observer is used to obtain the relative maneuvering estimation value based on the AUV relative motion state estimator and the state transfer equation, and the relative maneuvering estimation value is compensated to the federated Kalman filter.

[0027] Further preferably, the method for obtaining the actual measurement information of the AUV in step S2 is specifically as follows:

[0028] During the long-distance guidance stage, USBL array and USBL beacon are used to measure the actual measurement information of the AUV; during the medium-distance guidance stage, multi-beam forward-looking imaging sonar is used as the main method, USBL array and USBL beacon are used as the auxiliary methods to measure the actual measurement information of the AUV; during the short-distance guidance stage, camera is used as the main method, multi-beam forward-looking imaging sonar, USBL array and USBL beacon are used as the auxiliary methods to measure the actual measurement information of the AUV.

[0029] Further preferably, step S2 specifically includes the following steps:

[0030] Step S2.1: In the first sub-filter, the second sub-filter, the third sub-filter and the main filter, according to the state transfer matrix, the motor drive matrix, Time relative maneuver estimate and -1 moment global optimal estimation of relative motion state, obtain kPredict the status at all times; and based on -1 time to estimate the relative motion state covariance matrix, process noise and state transfer matrix, and obtain the corresponding k The moment-to-moment predicted state covariance matrix; wherein the global optimal estimated relative motion state at the initial moment is the relative motion state at the initial moment;

[0031] Step S2.2: In the first sub-filter, the second sub-filter and the third sub-filter, based on k Predict the status at all times, update the predicted measurement values ​​corresponding to USBL beacons and USBL arrays, multi-beam forward-looking imaging sonars and cameras, and combine k The actual measurement values ​​of the AUV of the USBL beacon and USBL array, multi-beam forward-looking imaging sonar and camera at all times, as well as the Kalman filter gain and k Predict the state covariance matrix at each moment and update the corresponding k Estimated relative motion state and estimated relative running state covariance matrix at each moment;

[0032] Step S2.3: In the main filter, the corresponding values ​​obtained by the first sub-filter, the second sub-filter, and the third sub-filter are k The relative motion state and the estimated relative motion state covariance matrix are combined with the corresponding main filter k The moment-by-moment prediction state covariance matrix and k The moment prediction state matrix is ​​calculated k AUV global state estimation information at time ; among them, k ≥1.

[0033] More preferably, k Momentary prediction status for: ;in, is the state transition matrix, 3 3-unit array; It is the maneuver drive matrix; is the sampling time; for The global optimal estimate of the relative motion state at all times, for Time relative maneuver estimate; ;in, for Auxiliary variables at time, is the observation gain matrix, is the auxiliary gain matrix, for The pseudo-inverse matrix of .

[0034] Further preferably, the predicted state covariance matrix for: ;in, for Moment i The sub-filter or main filter estimates the relative motion state covariance; For the i Process noise corresponding to the sub-filter or main filter; represents the main filter; is the state transition matrix.

[0035] In general, the above technical solutions conceived by this application have the following beneficial effects compared with the existing technologies:

[0036] Existing technologies estimate the AUV's motion state directly based on sensor data and filter models, without considering the relative maneuvers of the actual AUV. This leads to a mismatch between the AUV's relative motion and the filter model, resulting in deviations in the estimation of the AUV's relative motion state. The maneuvering observer provided in this application can effectively estimate the AUV's relative maneuvers, and the federated Kalman filter can effectively integrate multi-sensor information. The combined compensated federated Kalman filter can effectively estimate the AUV's motion state relative to the docking station. Compared with traditional estimation methods, this method has higher estimation accuracy and better robustness, making it more suitable for underwater dynamic docking. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 This is a schematic diagram of the composition of the sound and light fusion guidance system provided in an embodiment of the present application.

[0038] Figure 2 It is a schematic diagram of the sound and light fusion guidance method provided in an embodiment of the present application.

[0039] Figure 3 This is a schematic diagram of the compensated federated Kalman filter provided in an embodiment of the present application.

[0040] Figure 4 This is a flow chart of the sound and light fusion guidance method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0041] The embodiments of the present application are described below in conjunction with the drawings in the embodiments of the present application.

[0042] AUV underwater dynamic docking generally consists of a mobile docking station and the AUV. The docking is completed when the AUV successfully enters the docking station. The guidance system is responsible for the navigation and positioning of the AUV, aiming to guide the AUV into the docking station smoothly, and is the key to the docking system.

[0043] System components such as Figure 1As shown, the present application provides an acoustic-optical fusion guidance system, in which the AUV is equipped with an ultra-short baseline beacon (USBL) and an inertial navigation system (INS). The USBL can realize underwater acoustic communication between the AUV and the dock, and the INS can sense the absolute position, attitude and speed of the AUV; the dock is equipped with a USBL array, a multi-beam forward-looking imaging sonar (MFLS), a binocular camera and an INS; the relative slant range and position of the AUV can be obtained through acoustic signal propagation between the USBL array and the USBL beacon, and the MFLS can obtain a two-dimensional image of the AUV through acoustic propagation, and further obtain the two-dimensional position of the AUV; the camera can obtain an image of the AUV through optical photography, and further obtain the three-dimensional position of the AUV; the INS can sense the absolute position, speed and attitude of the dock.

[0044] The three sensors for measuring the relative position of AUVs each have their own advantages and disadvantages. The USBL has a wide range and is easily affected by the inherent defects of underwater acoustic propagation, and its positioning is prone to outliers. However, its slant range measurement accuracy is relatively reliable, and its close-range positioning accuracy is not high. The MFLS is also an acoustic measurement instrument and is also susceptible to the influence of underwater acoustic signals. Its range is smaller than that of the USBL. However, since it is based on visual imaging and then detects and locates the AUV's position information, when the AUV's features are accurately detected, the positioning has fewer outliers and is more accurate than the USBL. The binocular camera is an optical perception device and is not affected by sound propagation, but its range is indeed very small and it is very dependent on the visibility of the underwater environment. Its advantage is that it has higher positioning accuracy than the USBL and MFLS.

[0045] To maximize the use of the above-mentioned acoustic and optical sensors, such as Figure 2 As shown, the AUV underwater dynamic docking guidance system designed in this application is divided into three stages, namely the long-range guidance stage, the medium-range guidance stage and the close-range guidance stage; the long-range guidance stage relies only on the ultra-short baseline positioning system; the medium-range guidance stage is mainly based on multi-beam forward-looking imaging sonar, supplemented by the ultra-short baseline positioning system; the close-range guidance stage relies on binocular cameras, multi-beam forward-looking imaging sonar and ultra-short baseline positioning system as auxiliary means.

[0046] As mentioned above, the AUV underwater docking guidance system provided by the present application integrates three types of acoustic and optical navigation sensors, namely USBL, MFLS and binocular camera. Their information data measurement values ​​are different and their accuracy varies. Therefore, in order to obtain high-precision navigation and positioning information, it is also necessary to combine and filter the measurement values ​​observed by each sensor. For the combined navigation system, compared with using a Kalman filter to centrally process the information of all navigation sensors, the Federated Kalman Filter (FKF) can overcome the former's shortcomings such as high state dimension, large computational complexity and low fault tolerance. Therefore, the present application adopts FKF as a state estimation method. FKF is a typical information fusion technology. , which consists of multiple sub-filters and a main filter, and adopts two-stage decentralized filtering; finally, the optimal state estimation is obtained through the data fusion method; but for the dynamic docking of AUV underwater, since the AUV and the dock are in relative motion and the motion is uncertain, the traditional FKF cannot effectively obtain the relative positioning of the AUV, resulting in the guidance system not working well; therefore, this application designs a maneuvering observer, which obtains the relative maneuvering estimate based on the AUV relative motion state estimator and the state transfer equation, and then feeds back the compensation to the FKF, and finally obtains a more accurate motion state estimation of the AUV; this method is called compensated federated Kalman filtering (CFKF).

[0047] Before performing CFKF, since the measurement data of each sensor is based on the carrier coordinate system, the measurement data needs to be unified into the docking station coordinate system; ;in, Unify the measurement data of each navigation sensor to the measurement value in the docking station coordinate system; Measure AUV slant range values ​​for USBL; Measure the two-dimensional coordinates of AUV for MFLS; Measure the AUV 3D coordinates for the camera; The rotation and translation matrix for converting the USBL coordinate system to the docking station coordinate system; The rotation and translation matrices for transforming the MFLS coordinate system to the docking station coordinate system; The rotation and translation matrices for converting the binocular camera coordinate system to the docking station coordinate system can be determined based on the sensor installation position; and the AUV motion state also needs to be converted to the relative motion state in the docking station coordinate system; ;in, The three-axis position and velocity of the AUV measured by inertial navigation; The three-axis position and speed of the docking station; is the position and speed of the AUV relative to the dock; is the coordinate transformation matrix, which can be determined based on the AUV and docking station attitudes.

[0048] CFKF consists of the following four steps, the structure diagram is as follows Figure 3 shown.

[0049] Step S1: Information distribution:

[0050] The overall information weight of each filter is allocated according to the variance upper bound method: ;in, and To globally optimally estimate the relative motion state covariance and process noise, and Estimate the relative motion state covariance and process noise for the sub-filter, is the weight factor, satisfying: .

[0051] In this application, long-distance guidance , during mid-range guidance , close range guidance .

[0052] Step S2: Time update:

[0053] The main filter and the sub-filter are updated independently in time.

[0054] First update the forecast status : ;in, is the state transition matrix, 3 3-unit array; It is the maneuver drive matrix; is the sampling time; for Estimate the relative motion state at all times, for Estimate relative maneuver at any moment; estimate the maneuver observer as follows ;in, for Auxiliary variables at time, is the observation gain matrix, is the auxiliary gain matrix, which is set according to the actual situation. for The pseudo-inverse matrix of .

[0055] Calculate the predicted state covariance matrix : .

[0056] Step S3: Measurement update: Use the subsystem measurement value to measure and update its corresponding sub-filter; for linear systems, the sub-filter can use a conventional Kalman filter, and for nonlinear systems, it is necessary to use a nonlinear filter, such as an extended Kalman filter (EKF), an unscented Kalman filter (UKF), a cubic Kalman filter (CKF) and a particle filter (PF), etc.; in the embodiment of the present application, in order to reduce the amount of calculation and facilitate implementation and deployment, a more conventional EKF is used as a sub-filter.

[0057] The sub-filters update the predicted measurement values ​​separately: .

[0058] Calculate the Kalman filter gain separately : ;in, is the pseudo measurement matrix; is the measurement noise variance matrix; is the measurement equation.

[0059] Then the estimated relative motion state and covariance matrix are updated: .

[0060] The main filter has no measurement values, so no measurement update is performed.

[0061] Step S4: Information fusion: The estimated information of all sub-filters and the main filter is fused according to the following rules to calculate the global state estimation information: .

[0062] The global optimal estimation of the relative motion state after CFKF filtering is obtained through underwater acoustic communication Send to AUV to guide AUV movement, such as Figure 4 shown.

[0063] Compared with the existing technology, this application has the following advantages:

[0064] Existing technologies estimate the AUV's motion state directly based on sensor data and filter models, without considering the relative maneuvers of the actual AUV. This leads to a mismatch between the AUV's relative motion and the filter model, resulting in deviations in the estimation of the AUV's relative motion state. The maneuvering observer provided in this application can effectively estimate the AUV's relative maneuvers, and the federated Kalman filter can effectively integrate multi-sensor information. The combined compensated federated Kalman filter can effectively estimate the AUV's motion state relative to the docking station. Compared with traditional estimation methods, this method has higher estimation accuracy and better robustness, making it more suitable for underwater dynamic docking.

[0065] It should be understood that expressions such as "include" and "may include" used in this application indicate the existence of the disclosed functions, operations, or constituent elements, and do not limit one or more additional functions, operations, and constituent elements. In this application, terms such as "include" and / or "have" may be interpreted as indicating specific characteristics, numbers, operations, constituent elements, components, or combinations thereof, but may not be interpreted as excluding the existence or possibility of adding one or more other characteristics, numbers, operations, constituent elements, components, or combinations thereof.

[0066] In the description of the embodiments of the present application, it should be noted that, unless otherwise clearly specified and limited, the term "connection" should be understood in a broad sense.

[0067] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. An acoustic and optical fusion guidance system for underwater dynamic docking of AUV, characterized by: include: USBL array, USBL beacon, first inertial navigation system, second inertial navigation system, multi-beam forward-looking imaging sonar, camera, data initialization module and combined filtering module; The USBL beacon and the first inertial navigation system are on board the AUV; the USBL array, the second inertial navigation system, the multi-beam forward-looking imaging sonar, and the camera are set up on the docking station; The first and second inertial navigation systems are used to measure the absolute position, attitude, and velocity of the AUV and the docking station, respectively. The USBL array and USBL beacon, multi-beam forward-looking imaging sonar, and camera are used to measure the actual measurement information of the AUV during different guidance phases. The data initialization module is used to convert the absolute position, attitude and velocity of the AUV and the docking station at the initial moment into the docking station coordinate system to obtain the relative motion state at the initial moment; The combined filtering module is used to obtain the AUV global state estimation information by using the federated Kalman filter and the maneuvering observer for filtering fusion based on the relative motion state at the initial moment and the actual measurement information of the AUV obtained by the real-time USBL array, multi-beam forward-looking imaging sonar and camera; The maneuver observer is used to obtain a relative maneuver estimate based on the estimated relative motion state and the state transfer equation, and to compensate the relative maneuver estimate into the federated Kalman filter; The combined filtering module includes a first sub-filter, a second sub-filter, a third sub-filter, a main filter and a maneuvering observer; The first sub-filter, the second sub-filter and the third sub-filter are connected to the USBL array, the multi-beam forward-looking imaging sonar and the camera respectively; the main filter is connected to the first inertial navigation system and the second inertial navigation system; and a maneuvering observer is connected to the main filter side; The first sub-filter, the second sub-filter, the third sub-filter and the main filter are used to calculate the state transfer matrix, the motor drive matrix, Time relative maneuver estimate and -1 moment global optimal estimation of relative motion state, obtain k Predict the state at any moment; and use it to -1 time to estimate the relative motion state covariance matrix, process noise and state transfer matrix, and obtain the corresponding k The moment-to-moment predicted state covariance matrix; wherein the global optimal estimated relative motion state at the initial moment is the relative motion state at the initial moment; The first sub-filter, the second sub-filter and the third sub-filter are respectively used to k Predict the status at all times, update the predicted measurement values ​​corresponding to USBL beacons and USBL arrays, multi-beam forward-looking imaging sonars and cameras, and combine k The actual measurement values ​​of the AUV of the USBL beacon and USBL array, multi-beam forward-looking imaging sonar and camera at all times, as well as the Kalman filter gain and k Predict the state covariance matrix at each moment and update the corresponding k Estimated relative motion state and estimated relative running state covariance matrix at each moment; The main filter is used to obtain the corresponding k The relative motion state and the estimated relative motion state covariance matrix are combined with the corresponding main filter k The moment-by-moment prediction state covariance matrix and k The moment prediction state matrix is ​​calculated k AUV global state estimation information at the moment; in, k Momentary prediction status for: in, is the state transition matrix, 3 3-unit array; It is the maneuver drive matrix; is the sampling time; for The global optimal estimate of the relative motion state at all times, for Time relative maneuver estimate; in, for Auxiliary variables at time, is the observation gain matrix, is the auxiliary gain matrix, for The pseudo-inverse matrix of k ≥1.

2. The acousto-optic fusion guidance system according to claim 1, characterized in that: The USBL array and USBL beacon are used to measure the actual measurement information of the AUV during the long-range guidance phase. During the medium-range guidance phase, the multi-beam forward-looking imaging sonar is used as the main instrument, and the USBL array and USBL beacon are used as auxiliary instruments to measure the actual measurement information of the AUV. During the close-range guidance phase, the camera is used as the primary instrument, supplemented by multi-beam forward-looking imaging sonar, USBL array, and USBL beacon to measure the actual measurement information of the AUV; Among them, the area between the AUV and the dock is divided into long-range guidance stage, medium-range guidance stage and short-range guidance stage; the area larger than the range of the multi-beam forward-looking imaging sonar and the camera is divided into the long-range guidance stage; the area larger than the camera range and smaller than the beam forward-looking imaging sonar working range is divided into the medium-range guidance stage; the area less than or equal to the camera range is divided into the short-range guidance stage.

3. The acousto-optic fusion guidance system according to claim 1 or 2, characterized in that: Forecast state covariance matrix for: in, for Moment i The sub-filter or main filter estimates the relative motion state covariance; For the i Process noise corresponding to the sub-filter or main filter; represents the main filter; is the state transition matrix.

4. An acoustic-optic fusion guidance method for underwater dynamic docking of an AUV based on the acoustic-optic fusion guidance system according to claim 1, characterized in that: The following steps are involved: Step S1: Convert the absolute position, attitude, and velocity of the AUV and the docking station at the initial moment into the docking station coordinate system to obtain the relative motion state at the initial moment; Step S2: Based on the relative motion state at the initial moment and the actual measurement information of the AUV obtained by the real-time USBL array, multi-beam forward-looking imaging sonar and camera, a federated Kalman filter and a maneuvering observer are used for filtering fusion to obtain the AUV global state estimation information; Step S3: using the AUV global state estimation information to guide the AUV movement; Among them, a maneuvering observer is used to obtain the relative maneuvering estimation value based on the AUV relative motion state estimator and the state transfer equation, and the relative maneuvering estimation value is compensated to the federated Kalman filter; Step S2 specifically includes the following steps: Step S2.1: In the first sub-filter, the second sub-filter, the third sub-filter and the main filter, according to the state transfer matrix, the motor drive matrix, Time relative maneuver estimate and -1 moment global optimal estimation of relative motion state, obtain k Predict the status at all times; and based on -1 time to estimate the relative motion state covariance matrix, process noise and state transfer matrix, and obtain the corresponding k The moment-to-moment predicted state covariance matrix; wherein the global optimal estimated relative motion state at the initial moment is the relative motion state at the initial moment; Step S2.2: In the first sub-filter, the second sub-filter and the third sub-filter, based on k Predict the status at all times, update the predicted measurement values ​​corresponding to USBL beacons and USBL arrays, multi-beam forward-looking imaging sonars and cameras, and combine k The actual measurement values ​​of the AUV of the USBL beacon and USBL array, multi-beam forward-looking imaging sonar and camera at all times, as well as the Kalman filter gain and k Predict the state covariance matrix at each moment and update the corresponding k Estimated relative motion state and estimated relative running state covariance matrix at each moment; Step S2.3: In the main filter, the corresponding values ​​obtained by the first sub-filter, the second sub-filter, and the third sub-filter are k The relative motion state and the estimated relative motion state covariance matrix are combined with the corresponding main filter k The moment-by-moment prediction state covariance matrix and k The moment prediction state matrix is ​​calculated k AUV global state estimation information at the moment; k Momentary prediction status for: in, is the state transition matrix, 3 3-unit array; It is the maneuver drive matrix; is the sampling time; for The global optimal estimate of the relative motion state at all times, for Time relative maneuver estimate; in, for Auxiliary variables at time, is the observation gain matrix, is the auxiliary gain matrix, for The pseudo-inverse matrix of k ≥1.

5. The acousto-optic fusion guidance method according to claim 4, characterized in that: The method for obtaining the actual measurement information of the AUV in step S2 is specifically as follows: In the long-range guidance phase, USBL array and USBL beacon are used to measure the actual measurement information of the AUV; in the medium-range guidance phase, multi-beam forward-looking imaging sonar is used as the main method, supplemented by USBL array and USBL beacon, to measure the actual measurement information of the AUV; During the close-range guidance phase, the camera is used as the main instrument, supplemented by multi-beam forward-looking imaging sonar, USBL array and USBL beacon to measure the actual measurement information of the AUV.

6. The acousto-optic fusion guidance method according to claim 4 or 5, characterized in that: Forecast state covariance matrix for: in, for Moment i The sub-filter or main filter estimates the relative motion state covariance; For the i Process noise corresponding to the sub-filter or main filter; represents the main filter; is the state transition matrix.

Images