A method for reducing the rolling of a spherical unmanned surface vehicle and the spherical unmanned surface vehicle

Abstract

The application discloses a ball-shaped unmanned ship roll reduction method and a ball-shaped unmanned ship, and relates to the technical field of ships.The ball-shaped unmanned ship roll reduction method adopts double-shaft active mass damper (AMD) and vector propeller cooperative control, and distributes and processes low-frequency and high-frequency wave disturbance, thereby significantly improving the roll reduction effect of the ball-shaped unmanned ship (USV) in complex sea conditions.Through a clear frequency separation mechanism, the application can efficiently process low-frequency large-amplitude wave disturbance by the active mass damper, and process high-frequency instantaneous disturbance by the propeller, so that effective complementary control of the wave disturbance is realized, and the roll reduction efficiency is significantly improved.The double-mode cooperative control method not only overcomes the limitations of the traditional single control mechanism, but also significantly improves the overall stability and roll reduction effect of the system.

Description

Technical Field

[0001] This application relates to the field of ship technology, and in particular to a method for reducing the roll of a spherical unmanned surface vessel (USV) and the USV itself. Background Technology

[0002] Existing unmanned surface vessels (USVs) often face severe challenges in high sea state environments when performing high-precision tasks such as marine monitoring, exploration, or emergency rescue. In sea states 4–5, wave height is significant and wave energy is concentrated. Traditional monohull or catamaran hulls, due to insufficient seakeeping, are prone to large rolls, pitches, and heaves, resulting in a significant decrease in attitude stability and making it difficult to meet the precision requirements of marine monitoring, exploration, or rescue missions. Spherical USVs, with their unique fully symmetrical shape and omnidirectional maneuverability, demonstrate great potential in dealing with complex wave environments. Their shape, without fixed bow and stern, can reduce the directional impact of waves, theoretically possessing superior wave compliance and recovery characteristics, and is considered an important development direction for high sea state adaptability platforms. However, the key technologies of current spherical USVs are still immature, and their practical applications still have significant limitations. Existing technologies have the following limitations: Single control mechanism: Existing spherical USVs mostly rely on single thruster torque control, which excels at quickly responding to high-frequency disturbance components and stabilizing attitude through torque adjustment. However, when faced with persistent, large-amplitude disturbances generated by low-frequency large waves (long period, high energy), the thruster torque is often insufficient, failing to provide adequate and efficient compensation torque, leading to cumulative attitude deviations and even instability of the platform. A few studies have attempted to introduce single-axis active mass dampers (AMDs), generating anti-tilting torque through the movement of internal mass blocks. However, limited by internal space, the mass block's travel is limited, resulting in a small correction torque. This can only cope with some disturbances in lower sea states and cannot handle transient disturbances, thus limiting overall control effectiveness.

[0003] Lack of frequency coordination: Existing technologies lack a clear wave frequency separation mechanism and fail to effectively identify and separate low-frequency and high-frequency components that differ significantly in the wave spectrum. The processing of low-frequency and high-frequency disturbance components is mixed, resulting in low control efficiency and insufficient coordination, which seriously limits the potential of the actuator and the overall stability of the system.

[0004] Limitations of prediction and control: Existing control systems mostly rely on simple feedback algorithms (such as PID), which lack effective advance prediction and learning capabilities for nonlinear, time-varying, and highly correlated wave disturbances. The control lag is significant, which limits performance in high sea states and makes it difficult to achieve accurate, feedforward attitude stabilization in high sea states. Summary of the Invention

[0005] This application provides a method for reducing the roll of a spherical unmanned surface vessel (USV) and the USV itself, in order to overcome the problems of insufficient stability and poor roll reduction effect of existing spherical USVs under complex wave conditions. This method improves the attitude stability of the spherical USV under high sea conditions, thereby ensuring that it can reliably perform high-precision tasks such as marine monitoring, exploration or rescue.

[0006] This application provides a method for reducing the roll of a spherical unmanned surface vessel, including: The attitude and environmental information of the spherical unmanned surface vessel are obtained using multiple sensors; Based on the attitude and environmental information of the spherical unmanned surface vessel, a long short-term memory network is used to predict future wave disturbances and form a wave prediction disturbance torque. The predicted wave disturbance moment is decomposed into low-frequency disturbance components and high-frequency disturbance components; The low-frequency disturbance component is controlled by active mass damping (AMD), and the high-frequency disturbance component is controlled by the thruster.

[0007] This application also provides a spherical unmanned surface vessel, including a base and a vehicle roll reduction device; the vehicle roll reduction device includes: A central fixed platform is fixed to the base; The thrusters include a Y-direction thruster and an X-direction thruster, which are used to drive the spherical unmanned surface vessel to move in the Y and X directions; Active mass damping, installed on the lower surface of the central fixed platform, is used to suppress vibration; The active mass damping includes: The X-direction motion mechanism is installed on the lower surface of the central fixed platform; The Y-direction motion mechanism is slidably mounted on the lower surface of the X-direction motion mechanism. The Y-direction motion mechanism is arranged parallel to the central fixed platform, and the extension direction of the Y-direction motion mechanism is perpendicular to that of the X-direction motion mechanism. An X-direction motor is mounted on the central fixed platform and connected to the X-direction motion mechanism; A Y-direction motor is mounted on the central fixed platform and connected to the Y-direction motion mechanism; A mass ball is slidably mounted on the Y-direction motion mechanism.

[0008] This application employs a dual-axis active mass damping (AMD) and vector thruster coordinated control to distribute and process low-frequency and high-frequency wave disturbances, significantly improving the roll reduction performance of spherical unmanned surface vessels (USVs) in complex sea conditions. Through a clear frequency separation mechanism, this invention efficiently handles low-frequency, large-amplitude wave disturbances via the active mass damper, while high-frequency, transient disturbances are handled by the thrusters, thus achieving effective complementary control of wave disturbances and significantly improving roll reduction efficiency. This dual-mode coordinated control method not only overcomes the limitations of traditional single-control mechanisms but also significantly improves the overall stability and roll reduction effect of the system. Simultaneously, by combining intelligent prediction and optimization control technology using neural networks, the system can anticipate wave dynamics and optimize command execution, thereby reducing the lag of traditional control methods and improving control accuracy and dynamic response capabilities. This innovative technological breakthrough significantly improves the attitude stability of spherical USVs in high sea state environments, particularly excelling in high-precision tasks such as marine monitoring, offshore wind power maintenance, and ocean search and rescue. Meanwhile, the system's engineering optimization design, through lowering the center of gravity, static sealing, and modular design, effectively enhances the product's durability and adaptability in high sea state environments, reduces maintenance costs, and supports mass production and commercialization needs. Attached Figure Description

[0009] To more clearly illustrate the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a flowchart illustrating a method for reducing the roll of a spherical unmanned surface vessel in one embodiment of this application; Figure 2 This is a schematic diagram of the structure of a spherical unmanned surface vessel in one embodiment of this application; Figure 3 This is a schematic diagram of the anti-roll device for a spherical unmanned surface vessel in one embodiment of this application; Figure 4 This is a top view of the internal structure of a spherical unmanned surface vessel in one embodiment of this application; Figure 5 This is a bottom view of the internal structure of a spherical unmanned surface vessel in one embodiment of this application. Detailed Implementation

[0011] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of this application.

[0012] It should be noted that, in the description of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. The terms "first," "second," etc., in this application are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0013] like Figure 1 As shown, an embodiment of this application provides a method for reducing the roll of a spherical unmanned surface vessel, including the following steps: Step S1: Use multiple sensors to acquire attitude and environmental information of the spherical unmanned surface vessel; Step S2: Based on the attitude and environmental information of the spherical unmanned surface vessel, a long short-term memory network is used to predict future wave disturbances and form a predicted wave disturbance torque. Step S3: Decompose the predicted wave disturbance torque into low-frequency disturbance components and high-frequency disturbance components; Step S4: Control the low-frequency disturbance component to be processed by active mass damping (AMD), and control the high-frequency disturbance component to be processed by the thruster.

[0014] Relying solely on attitude information acquired from a single sensor can lead to accumulated errors (drift) over time, especially in turbulent sea conditions. Fusion of data from multiple sensors can improve the accuracy and robustness of attitude estimation. To provide accurate and reliable attitude information as the fundamental input to the control system, a multi-sensor fusion and attitude estimation method is introduced. This is the first step in the control system, providing reliable data for subsequent wave disturbance prediction, frequency separation, and control decisions. Attitude errors can be estimated by comparing accelerometer measurements with theoretical values ​​of the gravity vector.

[0015] In this embodiment, the acquisition of attitude and environmental information of the spherical unmanned surface vessel using multiple sensors includes: An inertial measurement unit (IMU) is used as the main attitude sensor to collect the three-axis angular velocities (ω) of the gyroscope inside the spherical unmanned surface vessel. x , ω y , ω z ) and triaxial acceleration (ax , a y , a z ); The attitude of the spherical unmanned surface vessel is estimated from a noisy measurement sequence using an extended Kalman filter based on the triaxial angular velocity and triaxial acceleration.

[0016] In this embodiment, estimating the attitude of the spherical unmanned surface vessel from a noisy measurement sequence using an extended Kalman filter based on the triaxial angular velocity and the triaxial acceleration includes: Set the state vector of the spherical unmanned surface vessel. C(q) = [q0, q1, q2, q3] Where C(q) is the attitude rotation matrix of the spherical unmanned surface vessel (USV), q0, q1, q2, and q3 represent the rotational attitude vectors of the USV in the four directions of the plane in the spatial coordinate system, and b g This represents the zero bias of the gyroscope inside the spherical unmanned surface vessel. The zero bias of the gyroscope is used to correct measurement errors in the gyroscope. The zero bias of the gyroscope inside the spherical unmanned surface vessel is indicated by its position in the x-axis, y-axis, and z-axis directions in the spatial coordinate system, where T is the transpose symbol; The attitude change model of the spherical unmanned surface vessel over time is set as follows: Where q takes the value q0, q1, q2 or q3, The time derivative of q ω represents the rate of change of attitude. m It is the angular velocity measurement value of the gyroscope, ω m The value is taken as the triaxial angular velocity ω x ω y or ω z n g It's gyroscope noise; By using the three-axis acceleration to constrain the direction of gravity, observation residuals are constructed. Used to estimate attitude error, where a m It is the acceleration measurement value of the gyroscope, a m The value is the triaxial acceleration a. x a y or a z g is the gravity vector; The attitude of the spherical unmanned surface vessel is estimated based on the time-varying attitude model of the spherical unmanned surface vessel and the observation residuals.

[0017] To overcome the cumulative errors and noise interference of a single sensor in dynamic environments, an Extended Kalman Filter (EKF) is introduced for attitude estimation. The EKF optimization estimation algorithm can estimate the state of a dynamic system from a noisy measurement sequence. EKF recursively updates the state estimate using both the system model and the observation model, thereby obtaining more accurate attitude information. Through recursive updates using EKF, the system can output high-frequency, low-latency, and drift-free attitude information under strong dynamic sea conditions, with sampling frequencies exceeding 200Hz, providing highly reliable input for subsequent control.

[0018] In this embodiment, the step of predicting future wave disturbances and forming wave prediction disturbance torques using a long short-term memory network based on the attitude and environmental information of the spherical unmanned surface vessel includes: A wave prediction model is constructed based on a Long Short-Term Memory (LSTM) network. The attitude deviation, angular velocity, and wave moment estimates at historical moments are used as historical measured sea state data. The wave prediction model is then supervised and trained using the historical measured sea state data and high sea state wave data generated through computational fluid dynamics simulation to obtain the trained wave prediction model. Based on the trained wave prediction model and the attitude and environmental information of the spherical unmanned surface vessel, future wave disturbances are predicted, and wave prediction disturbance torque is formed.

[0019] To achieve an anticipatory response to wave disturbances, the system constructs a wave prediction model based on Long Short-Term Memory (LSTM) networks. LSTM is suitable for processing time series with long-term dependencies, effectively capturing the nonlinear and non-stationary characteristics of waves, predicting future wave disturbances in advance, and providing a basis for feedforward control.

[0020] This provides feedforward information for the MPC (Model Predictive Control) discussed later, which can effectively reduce control lag and improve the system's response speed and control accuracy. The accuracy of its wave prediction directly affects the performance of MPC, and thus the performance of the entire control system.

[0021] Network Input: The input to the LSTM network consists of historical attitude deviations, angular velocities, and wave moment estimates. This information contains dynamic changes in wave perturbations, which helps the network learn the evolution patterns of waves.

[0022] Network structure: A two-layer stacked LSTM is used to improve the network's modeling capabilities. The design of the number of hidden layer nodes and the output of the fully connected layer needs to be adjusted according to the complexity of the actual problem and the amount of data.

[0023] Training strategy: Supervised learning using historical measured sea state data and high sea state wave data generated by CFD (Computational Fluid Dynamics) simulations can improve the network's generalization ability. The choice of loss function and optimizer needs to be adjusted according to the characteristics of the actual problem.

[0024] In this embodiment, the step of decomposing the wave prediction disturbance torque into low-frequency disturbance components and high-frequency disturbance components includes: A second-order Butterworth filter is used to set up a low-pass path and a high-pass path. The transfer functions corresponding to the low-pass path and the high-pass path are as follows: ;in, H LP (s) Let be the transfer function of the low-pass filter. H HP (s) Let be the transfer function of the high-pass filter. oh c The cutoff frequency, oh c This indicates the boundary between a low-pass filter and a high-pass filter. s For the complex variables of the Laplace transform; Based on the low-pass and high-pass paths, the real-time attitude deviation and wave prediction disturbance torque are decomposed in the frequency domain to obtain low-frequency disturbance components and high-frequency disturbance components.

[0025] Different actuators have different frequency response characteristics, requiring disturbances of different frequencies to be assigned to appropriate actuators for handling. By using a digital filter bank, real-time attitude deviations and predicted disturbance torques are decomposed in the frequency domain, enabling a coordinated control strategy where low-frequency disturbances are handled by dual-axis active mass damping (AMD) and high-frequency disturbances by the thrusters. Frequency separation is a key step in achieving coordinated control, improving control efficiency and system stability.

[0026] In this embodiment, controlling the low-frequency disturbance component is handled by active mass damping (AMD), and controlling the high-frequency disturbance component is handled by the thruster, including: The attitude error after low-pass filtering and the attitude error after high-pass filtering are determined based on the attitude error, respectively, according to the low-pass filtering operation and the high-pass filtering operation. Obtain the attitude error after low-pass filtering Obtain the attitude error after high-pass filtering. , where θ error (t) represents the attitude error, which indicates the difference between the actual attitude and the desired attitude of the spherical unmanned surface vessel. [ ] indicates a low-pass filter operation. [ ] indicates a high-pass filter operation; The attitude error after low-pass filtering and the attitude error after high-pass filtering are used as input references for active mass damping and thruster, and the control input for the coordinated control of the spherical unmanned surface vessel by the active mass damping and the thruster in a future preset time period is optimized. The control weights of the active mass damper and the thruster are dynamically adjusted according to the energy distribution of the low-frequency disturbance component and the high-frequency disturbance component, and the spherical unmanned surface vessel is controlled in coordination with the control input and control weights of the active mass damper and the thruster.

[0027] In this embodiment, optimizing the control input for the coordinated control of the active mass damping and the thruster of the spherical unmanned surface vessel within a preset future time period includes: constructing a predictive model based on the rigid body dynamics and actuator dynamics of the spherical unmanned surface vessel using model predictive control (MPC). The predictive model is... ;in, or =[ ϕ,θ,ψ ] T Let be an Euler angle vector, representing the roll angle ϕ, pitch angle θ, and bow angle ψ of the spherical unmanned surface vessel; Let be the first derivative of the Euler angle, representing the angular velocity; M is the second derivative of the Euler angle, representing angular acceleration; or C( ) is the inertia matrix, used to describe the inertial characteristics of a spherical unmanned surface vessel; or, ) is the Coriolis force matrix, used to describe the Coriolis force generated by the motion of a spherical unmanned surface vessel; D( or, ) is the damping matrix, used to describe the damping force acting on the spherical unmanned surface vessel; g( or The restoring torque is used to describe the restoring torque generated when a spherical unmanned surface vessel is disturbed. t AMD The control torque generated by active mass damping; t thr The thrust generated by the propulsion unit; t wave Predict the disturbance moment for the wave; Obtain the mechanism constraints of the spherical unmanned surface vessel, which include the travel of the active mass damper's mass block, the thruster thrust, the rate of change of thrust, and state constraints; wherein the travel of the active mass damper's mass block is: x AMD ∈[− x max , x max ], y AMD ∈[− y max, y max The thrust of the thruster is... F i ∈[0, F max The thrust variation rate is |Δ F i |≤Δ F max The state constraints are: or ∈ H , H Within the safe attitude range; Based on the aforementioned institutional constraints, the optimization problem is solved in each control cycle. ;in, u t:t+N-1 To control the input sequence, it includes the active mass damping position and thruster thrust for the next N steps; or t+k For predicted future moments k Euler angles; or ref For the desired Euler angles; Q This is the weight matrix for attitude error, used to adjust the accuracy of attitude control; R The weight matrix is ​​used to control the input and limit the energy of the control input. S The weight matrix is ​​used to control the rate of change of the input, and is used to limit the smoothness of the control input; Based on the output of solving the optimization problem, the control input for the spherical unmanned surface vessel to be controlled in coordination by the active mass damping and the thruster within a preset time period in the future is determined.

[0028] In this embodiment, dynamically adjusting the control weights of the active mass damping and the thruster based on the energy distributions of the low-frequency disturbance component and the high-frequency disturbance component includes: The short-time energy of the low-frequency disturbance component is obtained based on the energy distribution of the low-frequency disturbance component and the high-frequency disturbance component. E LP and the short-time energy of the high-frequency disturbance component E HP ; pass Adjusting the control weight of the active mass damping l AMD With the control weight of the thruster l thr .

[0029] To improve the robustness and adaptability of the system, enabling it to cope with various complex sea conditions and ensuring that the AMD and the thruster are always in optimal coordination under different sea conditions, avoiding energy waste or control conflicts, it is necessary to dynamically adjust the control weights of the two based on their different control capabilities and real-time disturbance energy distribution to achieve optimal coordinated control.

[0030] When low-frequency energy is dominant, the contribution weight of AMD is increased in MPC optimization; conversely, the thruster response is enhanced. This mechanism ensures that the two are always in optimal coordination under different sea states, avoiding energy waste or control conflicts.

[0031] Through the construction of the aforementioned theoretical framework, this invention achieves fully intelligent closed-loop control from perception to execution, possessing strong adaptability to complex wave environments in high sea states, and is expected to significantly improve the attitude stability and control accuracy of the vehicle. The superiority of this control method has been verified in simulation environments and prototype testing, laying a solid theoretical foundation for subsequent engineering applications.

[0032] The aforementioned method for reducing the roll of a spherical unmanned surface vessel (USV) employs a dual-axis active mass damping (AMD) and vector thruster coordinated control to distribute and process low-frequency and high-frequency wave disturbances, significantly improving the roll reduction effect of the USV in complex sea states. Through a clear frequency separation mechanism, this invention efficiently handles low-frequency, large-amplitude wave disturbances by the active mass damper, while high-frequency, transient disturbances are handled by the thruster, thus achieving effective complementary control of wave disturbances and significantly improving roll reduction efficiency. This dual-mode coordinated control method not only overcomes the limitations of traditional single control mechanisms but also significantly improves the overall stability and roll reduction effect of the system. Simultaneously, by combining intelligent prediction and optimization control technology using neural networks, the system can anticipate wave dynamics and optimize command execution, thereby reducing the lag of traditional control methods and improving control accuracy and dynamic response capabilities. This innovative technological breakthrough significantly improves the attitude stability of the spherical USV in high sea state environments, particularly excelling in high-precision tasks such as marine monitoring, offshore wind power maintenance, and ocean search and rescue. Meanwhile, the system's engineering optimization design, through lowering the center of gravity, static sealing, and modular design, effectively enhances the product's durability and adaptability in high sea state environments, reduces maintenance costs, and supports mass production and commercialization needs.

[0033] In one embodiment, such as Figure 2 , Figure 3 , Figure 4 , Figure 5 As shown, a spherical unmanned surface vessel (USV) anti-roll device 10 is provided, including a base and a vehicle anti-roll device; the vehicle anti-roll device includes: Central fixed platform 1, suitable for fixing to the base of the aircraft; The thrusters include a Y-direction thruster 21 and an X-direction thruster 22, which drive the spherical unmanned surface vessel to move in the Y and X directions, respectively; an active mass damper 3 is installed on the lower surface of the central fixed platform 1. The active mass damping 3 includes: The X-direction motion mechanism 31 is installed on the lower surface of the central fixed platform 1; The Y-direction motion mechanism 33 is slidably mounted on the lower surface of the X-direction motion mechanism 31. The Y-direction motion mechanism 33 is arranged parallel to the central fixed platform 1, and the extension direction of the Y-direction motion mechanism 33 is perpendicular to that of the X-direction motion mechanism 31. The X-direction motor 32 is mounted on the central fixed platform 1 and connected to the X-direction motion mechanism 31; The Y-direction motor 34 is mounted on the central fixed platform 1 and connected to the Y-direction motion mechanism 33; The mass ball 35 is slidably mounted on the Y-direction motion mechanism 33.

[0034] The central fixed platform provides stable support for the thrusters, motors, and active mass dampers, ensuring the system's structural stability under wave disturbances. Independent motion mechanisms in the X and Y directions, along with corresponding motor controls, allow the mass sphere to slide flexibly in two orthogonal directions, generating anti-tilting moments for different directions and effectively compensating for low-frequency, high-amplitude waves. Simultaneously, the thrusters handle rapid responses to high-frequency transient disturbances, complementing the active mass damping system and ensuring that low-frequency and high-frequency disturbances are handled by the most suitable actuators, significantly improving roll reduction efficiency. This dual-axis collaborative structure not only enhances the system's dynamic response and control precision but also significantly improves stability in high sea states through frequency separation and mechanical structure optimization. Furthermore, the sliding structure and modular design give the system excellent durability and adaptability, enabling reliable operation in complex sea conditions and ensuring the attitude stability and operational reliability of the spherical unmanned surface vessel when performing high-precision tasks such as marine monitoring, exploration, and rescue.

[0035] Furthermore, the plane where the X-direction thruster 32 is located is perpendicular to the central fixed platform 1, the plane where the Y-direction thruster 31 is located is perpendicular to the plane where the X-direction thruster 32 is located, and the plane where the Y-direction thruster 31 is located is perpendicular to the central fixed platform 1.

[0036] Furthermore, the X-direction motion mechanism 31 includes a first synchronous pulley 311, an X-direction synchronous belt 312, a first guide rod 313, and a first idler pulley 314. The first synchronous pulley 311 and the first idler pulley 314 are arranged at both ends of the lower surface of the central fixed platform 1 along the X-direction. The X-direction synchronous belt 312 is arranged around the first synchronous pulley 311 and the first idler pulley 314. The two first guide rods 313 are respectively arranged parallel to each other on both sides of the extension direction of the X-direction synchronous belt 312. The X-direction motor 32 drives the X-direction synchronous belt 312 through the first synchronous pulley 311. The X-direction synchronous belt 312 pulls the Y-direction motion mechanism 33 to move linearly along the first guide rod 313.

[0037] Furthermore, the X-direction motion mechanism 31 also includes at least two first photoelectric components 315, which are correspondingly disposed at both ends of the first guide rod 313. The at least two first photoelectric components 315 are used to limit the travel of the Y-direction motion mechanism 33 in the X direction.

[0038] Furthermore, the X-direction motion mechanism 31 also includes an X-axis slide 316, which is slidably mounted on the two first guide rods 313, and the Y-direction motion mechanism 33 is mounted on the X-axis slide 316.

[0039] Furthermore, the Y-direction motion mechanism 33 includes a second synchronous pulley 331, a Y-direction synchronous belt 332, a second guide rod 333, and a second idler pulley 334. The second synchronous pulley 331 and the second idler pulley 334 are arranged at both ends of the lower surface of the X-axis slide table 316 along the Y-direction. The Y-direction synchronous belt 332 is arranged around the second synchronous pulley 331 and the second idler pulley 334. The two second guide rods 333 are respectively arranged parallel to each other on both sides of the extension direction of the Y-direction synchronous belt 332. The Y-direction motor 32 drives the Y-direction synchronous belt 332 through the second synchronous pulley 331. The Y-direction synchronous belt 332 pulls the mass ball 35 to move linearly along the second guide rod 333.

[0040] Furthermore, the Y-direction motion mechanism 33 also includes at least two second photoelectric components 335, which are correspondingly disposed at both ends of the second guide rod 333. The at least two second photoelectric components 335 are used to limit the travel of the mass ball 35 in the X direction.

[0041] Furthermore, the X-direction motion mechanism 31 also includes a cable carrier 317, the interior of which is adapted to house cables and air pipes. One end of the cable carrier 317 is fixed, and the other end of the cable carrier 317 moves with the X-axis slide table 316.

[0042] The spherical unmanned surface vessel further includes a controller, which is electrically connected to the thruster and the active mass damping X-axis motor 22 and Y-axis motor 24. The controller includes: Memory, used to store computer programs; A processor is used to implement the steps of the aforementioned spherical unmanned surface vessel roll reduction method when executing the computer program.

[0043] The operation process of the aforementioned spherical unmanned surface vessel anti-roll device is as follows.

[0044] Achieving XY motion in the horizontal plane (1) X-axis motion Drive source: X-direction motor 33 drives X-direction synchronous belt 312 through first synchronous pulley 311.

[0045] Transmission process: The synchronous belt is fixed on the first synchronous pulley 311 and the first idler pulley 314 at both ends of the X-direction motion mechanism. The X-direction motor 33 drives the X-direction synchronous belt 312 via the first synchronous pulley 311, thereby pulling the Y-axis motion mechanism 33 to move linearly along the guide rod 313.

[0046] Guiding and support: The first guide rod 313 ensures straight movement, and the first idler pulley 314 is used to tension the synchronous belt.

[0047] Photoelectric limit switch: The first photoelectric component 315 is used to limit the maximum stroke of the Y-direction motion mechanism 33 along the first guide rod 313 in the X direction to ensure the safety of the component.

[0048] (3) Y-axis motion Drive source: Y-direction motor 34 drives Y-direction synchronous belt 332 through second synchronous pulley 331.

[0049] Transmission and guidance: The Y-direction synchronous belt 332 drives the Y-direction slide table to move along the second guide rod 333, and the second idler pulley 334 tensions the Y-direction synchronous belt 332.

[0050] Photoelectric limit switch: The second photoelectric component 335 is used to limit the maximum travel of the mass ball 35 along the second guide rod 333 in the Y direction to ensure the safety of the component.

[0051] (3) Motion superposition The Y-direction motion mechanism 33 is mounted on the X-axis slide 316 to achieve two-dimensional planar motion.

[0052] The spherical unmanned surface vessel also includes a tow chain 317. Cables and air pipes are placed inside the tow chain 317, with one end fixed and the other end moving with the X-axis slide 316 to prevent the cables from getting tangled or stretched and damaged.

[0053] An example of a complete motion cycle is shown below.

[0054] Suppose the platform needs to move from the origin to the target point (x1, y1), and then executes the following steps: S11. The control system sends commands to the X and Y motors.

[0055] S12. The X-axis motor drives the X-axis slide to position x1 via a synchronous belt.

[0056] S13. The Y-axis motor drives the Y-axis slide (with the X-axis mechanism) to position y1 via a synchronous belt.

[0057] S14. If vibration occurs during the movement, the AMD system starts up and drives the mass ball to perform reverse compensation movement to stabilize the platform.

[0058] S15. After reaching the target position, each shaft maintains its position (motor closed-loop holding torque), and the cable chain bends accordingly.

[0059] This application has the following characteristics: Dual-mode collaboration: The first-ever dual-axis AMD and vector thruster collaborative control targets low-frequency and high-frequency wave disturbances respectively, overcoming the limitations of a single system and significantly improving the anti-roll effect.

[0060] Explicit frequency separation: Through a clear frequency division mechanism, the allocation and processing of low-frequency and high-frequency disturbances are optimized to achieve efficient complementary control.

[0061] Intelligent proactive control: Combining neural network prediction and optimization control, it can detect wave dynamics in advance, reduce phase lag, and improve control accuracy.

[0062] Engineering optimization: Enhanced high sea state durability and mission adaptability through lowering the center of gravity, static sealing, and modular design, while reducing mass production costs.

[0063] This application, compared to existing technologies, employs a dual-axis active mass damping (AMD) and vector thruster coordinated control to distribute and process low-frequency and high-frequency wave disturbances, significantly improving the roll reduction effect of spherical unmanned surface vessels (USVs) in complex sea states. Through a clear frequency separation mechanism, this invention efficiently handles low-frequency, large-amplitude wave disturbances via the active mass damper, while high-frequency, transient disturbances are handled by the thrusters, thus achieving effective complementary control of wave disturbances and significantly improving roll reduction efficiency. This dual-mode coordinated control method not only overcomes the limitations of traditional single control mechanisms but also significantly improves the overall stability and roll reduction effect of the system. Simultaneously, by combining intelligent prediction and optimization control technology using neural networks, the system can anticipate wave dynamics and optimize command execution, thereby reducing the lag of traditional control methods and improving control accuracy and dynamic response capabilities. This innovative technological breakthrough significantly improves the attitude stability of spherical USVs in high sea state environments, particularly excelling in high-precision tasks such as marine monitoring, offshore wind power maintenance, and ocean search and rescue. Meanwhile, the system's optimized engineering design, through lowering the center of gravity, static sealing, and modular design, effectively enhances the product's durability and adaptability in high sea state environments, reduces maintenance costs, and supports mass production and commercialization needs. This includes: High sea state performance: Through multi-sensor fusion perception and adaptive anti-interference algorithm optimization, it can capture the dynamic changes of sea waves in real time and accurately adjust the attitude parameters of the vehicle carrier, perfectly adapting to the needs of various high sea state missions such as marine exploration, offshore wind power operation and maintenance, and ocean search and rescue. Its performance comprehensively surpasses the limits of existing single control systems.

[0064] Synergistic gain: Adopting an innovative design of "dual-mode synergy + frequency separation", it achieves a control effect of 1+1>2, and the anti-shake efficiency far exceeds that of single module or implicit synergistic technology.

[0065] Control Precision: Integrating intelligent prediction algorithms and optimized control strategies, it learns wave variation patterns under different sea conditions through big data analysis, predicts the impact of waves on the vehicle in advance, and dynamically optimizes control parameters to achieve a closed-loop, highly efficient response of "prediction-adjustment-feedback". Intelligent prediction and optimized control reduce lag, resulting in more precise dynamic response and adaptability to nonlinear wave environments.

[0066] The aforementioned spherical unmanned surface vessel (USV) roll reduction device employs a dual-axis active mass damping (AMD) and vector thruster coordinated control to distribute and process low-frequency and high-frequency wave disturbances, significantly improving the roll reduction effect of the USV in complex sea states. Through a clear frequency separation mechanism, this invention efficiently handles low-frequency, large-amplitude wave disturbances by the active mass damper, while high-frequency, transient disturbances are handled by the thruster, thus achieving effective complementary control of wave disturbances and significantly improving roll reduction efficiency. This dual-mode coordinated control method not only overcomes the limitations of traditional single control mechanisms but also significantly improves the overall stability and roll reduction effect of the system. Simultaneously, by combining intelligent prediction and optimization control technology using neural networks, the system can anticipate wave dynamics and optimize command execution, thereby reducing the lag of traditional control methods and improving control accuracy and dynamic response capabilities. This innovative technological breakthrough significantly improves the attitude stability of the spherical USV in high sea state environments, particularly excelling in high-precision tasks such as marine monitoring, offshore wind power maintenance, and ocean search and rescue. Meanwhile, the system's engineering optimization design, through lowering the center of gravity, static sealing, and modular design, effectively enhances the product's durability and adaptability in high sea state environments, reduces maintenance costs, and supports mass production and commercialization needs.

[0067] For a description of the features in the embodiment corresponding to the spherical unmanned surface vessel roll reduction device, please refer to the relevant description of the embodiment corresponding to the spherical unmanned surface vessel roll reduction method, which will not be repeated here.

[0068] Embodiments of this application also provide a computer-readable storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above embodiments of the spherical unmanned surface vessel roll reduction method when running: The attitude and environmental information of the spherical unmanned surface vessel are obtained using multiple sensors; Based on the attitude and environmental information of the spherical unmanned surface vessel, a long short-term memory network is used to predict future wave disturbances and form a wave prediction disturbance torque. The predicted wave disturbance moment is decomposed into low-frequency disturbance components and high-frequency disturbance components; The low-frequency disturbance component is controlled by active mass damping (AMD), and the high-frequency disturbance component is controlled by the thruster.

[0069] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard disk, magnetic disk, or optical disk.

[0070] Embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above embodiments of the spherical unmanned surface vessel roll reduction method: The attitude and environmental information of the spherical unmanned surface vessel are obtained using multiple sensors; Based on the attitude and environmental information of the spherical unmanned surface vessel, a long short-term memory network is used to predict future wave disturbances and form a wave prediction disturbance torque. The predicted wave disturbance moment is decomposed into low-frequency disturbance components and high-frequency disturbance components; The low-frequency disturbance component is controlled by active mass damping (AMD), and the high-frequency disturbance component is controlled by the thruster.

[0071] Embodiments of this application also provide another computer program product, including a non-volatile computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the steps in any of the above embodiments of the spherical unmanned surface vessel roll reduction method. The attitude and environmental information of the spherical unmanned surface vessel are obtained using multiple sensors; Based on the attitude and environmental information of the spherical unmanned surface vessel, a long short-term memory network is used to predict future wave disturbances and form a wave prediction disturbance torque. The predicted wave disturbance moment is decomposed into low-frequency disturbance components and high-frequency disturbance components; The low-frequency disturbance component is controlled by active mass damping (AMD), and the high-frequency disturbance component is controlled by the thruster.

[0072] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0073] The foregoing has provided a detailed description of a method for reducing the roll of a spherical unmanned surface vessel (USV) and the USV itself. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of these embodiments are only intended to aid in understanding the method and its core concepts. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the scope of protection of the claims.

Claims

1. A method for reducing the roll of a spherical unmanned surface vessel, characterized in that, include: The attitude and environmental information of the spherical unmanned surface vessel are obtained using multiple sensors; Based on the attitude and environmental information of the spherical unmanned surface vessel, a long short-term memory network is used to predict future wave disturbances and form a wave prediction disturbance torque. The predicted wave disturbance moment is decomposed into low-frequency disturbance components and high-frequency disturbance components; The low-frequency disturbance component is controlled by active mass damping, while the high-frequency disturbance component is controlled by the thruster. The control weights of the active mass damper and the thruster are dynamically adjusted according to the energy distribution of the low-frequency disturbance component and the high-frequency disturbance component, and the spherical unmanned surface vessel is controlled in coordination with the control input and control weights of the active mass damper and the thruster.

2. The method for reducing the roll of a spherical unmanned surface vessel according to claim 1, characterized in that, The method of acquiring attitude and environmental information of the spherical unmanned surface vessel using multiple sensors includes: An inertial measurement unit is used as the main attitude sensor to collect the three-axis angular velocity and three-axis acceleration of the gyroscope inside the spherical unmanned surface vessel; The attitude of the spherical unmanned surface vessel is estimated from a noisy measurement sequence using an extended Kalman filter based on the triaxial angular velocity and triaxial acceleration.

3. The method for reducing the roll of a spherical unmanned surface vessel according to claim 2, characterized in that, The step of estimating the attitude of the spherical unmanned surface vessel from a noisy measurement sequence using an extended Kalman filter based on the triaxial angular velocity and the triaxial acceleration includes: Set the state vector of the spherical unmanned surface vessel. C(q) = [q0, q1, q2, q3] Where C(q) is the attitude rotation matrix of the spherical unmanned surface vessel (USV), q0, q1, q2, and q3 represent the rotational attitude vectors of the USV in the four directions of the plane in the spatial coordinate system, and b g This represents the zero bias of the gyroscope inside the spherical unmanned surface vessel. The zero bias of the gyroscope is used to correct measurement errors in the gyroscope. The zero bias of the gyroscope inside the spherical unmanned surface vessel is indicated by its position in the x-axis, y-axis, and z-axis directions in the spatial coordinate system, where T is the transpose symbol; The attitude change model of the spherical unmanned surface vessel over time is set as follows: Where q takes the value q0, q1, q2 or q3, The time derivative of q ω represents the rate of change of attitude. m It is the angular velocity measurement value of the gyroscope, ω m The value is taken as the triaxial angular velocity ω x ω y or ω z n g It's gyroscope noise; By using the three-axis acceleration to constrain the direction of gravity, observation residuals are constructed. Used to estimate attitude error, where a m It is the acceleration measurement value of the gyroscope, a m The value is the triaxial acceleration a. x a y or a z g is the gravity vector; The attitude of the spherical unmanned surface vessel is estimated based on the time-varying attitude model of the spherical unmanned surface vessel and the observation residuals.

4. The method for reducing the roll of a spherical unmanned surface vessel according to claim 1, characterized in that, The step of predicting future wave disturbances and forming wave prediction disturbance torques based on the attitude and environmental information of the spherical unmanned surface vessel using a long short-term memory network includes: A wave prediction model is constructed based on a long short-term memory network; The attitude deviation, angular velocity, and wave moment estimates at historical moments are used as historical measured sea state data. The wave prediction model is then supervised and trained using the historical measured sea state data and high sea state wave data generated through computational fluid dynamics simulation to obtain the trained wave prediction model. Based on the trained wave prediction model and the attitude and environmental information of the spherical unmanned surface vessel, future wave disturbances are predicted, and wave prediction disturbance torque is formed.

5. The method for reducing the roll of a spherical unmanned surface vessel according to claim 1, characterized in that, The step of decomposing the predicted wave disturbance moment into low-frequency disturbance components and high-frequency disturbance components includes: A second-order Butterworth filter is used to set up a low-pass path and a high-pass path. The transfer functions corresponding to the low-pass path and the high-pass path are as follows: , ;in, H LP (s) Let be the transfer function of the low-pass filter. H HP (s) Let be the transfer function of the high-pass filter. ω c The cutoff frequency, ω c This indicates the boundary between a low-pass filter and a high-pass filter. s For the complex variables of the Laplace transform; Based on the low-pass and high-pass paths, the real-time attitude deviation and wave prediction disturbance torque are decomposed in the frequency domain to obtain low-frequency disturbance components and high-frequency disturbance components.

6. The method for reducing the roll of a spherical unmanned surface vessel according to claim 5, characterized in that, The control of the low-frequency disturbance component is handled by active mass damping, and the control of the high-frequency disturbance component is handled by the thruster, including: The attitude error after low-pass filtering and the attitude error after high-pass filtering are determined based on the attitude error, respectively, according to the low-pass filtering operation and the high-pass filtering operation. Obtain the attitude error after low-pass filtering Obtain the attitude error after high-pass filtering. , where θ error (t) represents the attitude error, which indicates the difference between the actual attitude and the desired attitude of the spherical unmanned surface vessel. [ ] indicates a low-pass filter operation. [ ] indicates a high-pass filter operation; The attitude error after low-pass filtering and the attitude error after high-pass filtering are used as input references for active mass damping and thruster, and the control input for the coordinated control of the spherical unmanned surface vessel by the active mass damping and the thruster in a future preset time period is optimized.

7. The method for reducing the roll of a spherical unmanned surface vessel according to claim 6, characterized in that, The optimization of the control input for the coordinated control of the active mass damping and the thruster of the spherical unmanned surface vessel within a preset future time period includes: A predictive model is constructed based on the rigid body dynamics and actuator dynamics of the spherical unmanned surface vessel using a model predictive control approach. The predictive model is as follows: ;in, Let be an Euler angle vector, representing the roll angle of the spherical unmanned surface vessel. ϕ Pitch angle θ and bow angle ψ ; Let be the first derivative of the Euler angle, representing the angular velocity; Let be the second derivative of the Euler angle, representing angular acceleration; This is the inertia matrix, used to describe the inertial characteristics of a spherical unmanned surface vessel; This is the Coriolis force matrix, used to describe the Coriolis force generated by the motion of a spherical unmanned surface vessel; This is the damping matrix, used to describe the damping force acting on the spherical unmanned surface vessel. Restoring torque is used to describe the restoring torque generated when a spherical unmanned surface vessel is disturbed. The control torque generated by active mass damping; The thrust generated by the propulsion unit; Predict the disturbance moment for the wave; Obtain the mechanism constraints of the spherical unmanned surface vessel, which include the travel of the active mass damper's mass block, the thruster thrust, the rate of change of thrust, and state constraints; wherein the travel of the active mass damper's mass block is: , The thrust of the thruster is The thrust variation rate is The state constraints are: , H Within the safe attitude range; Based on the aforementioned institutional constraints, the optimization problem is solved in each control cycle. ;in, To control the input sequence, it includes the active mass damping position and thruster thrust for the next N steps; For predicted future moments k Euler angles; For the desired Euler angles; Q This is the weight matrix for attitude error, used to adjust the accuracy of attitude control; R The weight matrix is ​​used to control the input and limit the energy of the control input. S The weight matrix is ​​used to control the rate of change of the input, and is used to limit the smoothness of the control input; Based on the output of solving the optimization problem, the control input for the spherical unmanned surface vessel to be controlled in coordination by the active mass damping and the thruster within a preset time period in the future is determined.

8. The method for reducing the roll of a spherical unmanned surface vessel according to claim 6, characterized in that, The step of dynamically adjusting the control weights of the active mass damping and the thruster based on the energy distribution of the low-frequency disturbance component and the high-frequency disturbance component includes: The short-time energy of the low-frequency disturbance component is obtained based on the energy distribution of the low-frequency disturbance component and the high-frequency disturbance component. E LP and the short-time energy of the high-frequency disturbance component E HP ; pass Adjusting the control weight of the active mass damping λ AMD With the control weight of the thruster λ thr .

9. A spherical unmanned surface vessel, characterized in that, Includes a base and a vehicle roll reduction device; the vehicle roll reduction device includes: The central fixed platform (1) is fixed to the base; The thruster (2) includes a Y-direction thruster (21) and an X-direction thruster (22), the thruster (2) being used to drive the spherical unmanned surface vessel to move in the Y and X directions; An active mass damper (3) is installed on the lower surface of the central fixed platform (1) to suppress vibration; The active mass damping (3) includes: The X-direction motion mechanism (31) is installed on the lower surface of the central fixed platform (1); The Y-direction motion mechanism (33) is slidably mounted on the lower surface of the X-direction motion mechanism (31). The Y-direction motion mechanism (33) is arranged parallel to the central fixed platform (1), and the extension direction of the Y-direction motion mechanism (33) is perpendicular to that of the X-direction motion mechanism (31). The X-direction motor (32) is mounted on the central fixed platform (1) and connected to the X-direction motion mechanism (31). The Y-direction motor (34) is mounted on the central fixed platform (1) and connected to the Y-direction motion mechanism (33). A mass ball (35) is slidably mounted on the Y-direction motion mechanism (33). The spherical unmanned surface vessel also includes a controller electrically connected to the thruster and the active mass damping X-direction motor (32) and Y-direction motor (34). The controller includes: Memory, used to store computer programs; A processor, configured to implement the steps of the spherical unmanned surface vessel roll reduction method as described in any one of claims 1 to 8 when executing the computer program.

Images