A smart oil boom boom robotic arm device and its control method

Abstract

This invention relates to an intelligent oil boom boom robotic arm device and its control method, comprising a robotic arm, a hydraulic drive system, a sensing system, and a control unit. The robotic arm includes a transmission base, a steering column, a main arm, a secondary arm, a tilting head, and an end effector hook assembly. The hydraulic drive system includes a hydraulic pump station, a proportional directional solenoid valve group, and multiple actuator cylinders. The sensing system includes a gyroscope array and a camera array, forming a distributed attitude perception network and a distributed visual perception network. The control unit identifies and locates the oil boom target based on sea surface images acquired by the distributed visual perception network, performs motion planning for the robotic arm, controls the robotic arm's movement, and grasps the target. During target positioning, the distributed attitude perception network senses real-time attitude changes of various parts of the robotic arm and dynamically compensates for the robotic arm's pose error. This device can quickly, accurately, and reliably grasp oil booms intelligently, improving the efficiency and safety of oil boom deployment.

Description

Technical Field

[0001] This invention belongs to the field of oil boom boom robotic arm technology, specifically relating to an intelligent oil boom boom robotic arm device and its control method. Background Technology

[0002] In emergency response to marine oil spills, the rapid deployment and retrieval of oil booms are crucial for controlling the spread of pollution. Current mainstream technologies can be categorized into three types, but all have significant limitations and are insufficient to meet the demands of efficient operations in complex sea conditions.

[0003] Manual deployment of oil booms is the traditional mainstream solution, relying on large vessels carrying multiple operators to manually connect oil boom sections, drop pontoons, and adjust their positions (Peng Bo et al., "Research Progress in Marine Oil Spill Emergency Response Technology", 2019). For example, CN104440919A discloses a pontoon gripping manipulator that uses a hydraulic cylinder to drive a hook to clamp the pontoon. While this simplifies the single-point gripping process, it still requires personnel to operate it on deck throughout the entire process, and its function is limited to pontoon fixation, failing to address the overall deployment efficiency issue. This method is severely constrained by sea conditions, requiring operation in calm waters with wind speeds <20m / s and wave heights <0.5m. A single deployment takes over 2 hours, exhibiting significant response lag.

[0004] To enhance automation, some technologies employ remotely operated boats with mechanical structures to achieve semi-automated operations. For example, CN218022095U discloses a "remotely operated boat for rapidly deploying oil booms," which releases the boom via a stern winding module, inflates the float-type boom using an air intake, and relies on the ship's motor to drive the reel for deployment and retrieval. CN205530181U, on the other hand, has a robotic arm attached to the unmanned vessel's hull to connect to an oil spill recovery device; however, the robotic arm is only used to fix the recovery device's position and does not participate in boom grasping or attitude adjustment. While such solutions reduce the risk of personnel exposure, the deployment path is entirely dependent on the boat's course, making it impossible to dynamically correct the boom's shape.

[0005] To optimize deployment speed, CN204199260U proposes a "boom drum winding device," which integrates a planetary gear reducer and a hydraulic motor within the drum cavity, adjusting the release speed by controlling the drum's rotational speed. While this solution automates the deployment process, it lacks an end effector, making it impossible to grip, dock, or fine-tune the angle of the boom's endpoints. This results in the boom being susceptible to displacement due to ocean currents after entering the water, making it difficult to form a precise U-shaped or V-shaped containment zone.

[0006] Specifically, existing technologies suffer from the following core problems: insufficient deployment accuracy and dynamic adaptability. Remote-controlled boat winding devices (such as CN218022095U) indirectly adjust the fence position through the boat's trajectory. When constructing complex enclosure shapes, multiple boats need to coordinate or the course needs repeated adjustments, resulting in a response delay exceeding 30 minutes. Furthermore, the mechanical functions are limited and lack coordination. Existing technologies mostly focus on optimizing local aspects. The human-assisted robotic arm (CN104440919A) can only perform single-point gripping, and the remote-controlled boat robotic arm (CN205530181U) is only used for fixed purposes; neither has achieved a complete "identification-positioning-grabbing-compensation" system. The entire process is closed-loop; manual operation poses safety hazards. Even with highly automated solutions (such as CN205530181U), manual operation is still required to complete the initial connection of the oil boom to the vessel, end fixation, and abnormal handling. In severe sea conditions with wave height ≥1.5m, the risk of operators falling into the water reaches 38%. Attitude compensation and environmental robustness are lacking. Traditional robotic arms mostly use a single gyroscope for attitude correction. When the hull rolls ±15°, the end positioning error exceeds 32cm, and the hydraulic system response delay is >100ms, which cannot offset wave disturbances in real time. This causes the alignment deviation between the grab and the target to increase linearly with wave height.

[0007] The aforementioned technical bottlenecks result in an oil boom deployment success rate of less than 50% in sea state 3 (wave height ≤ 1.5m), and the efficiency is only one-third that of manual operation, making it difficult to meet the practical needs of unmanned emergency response. Therefore, developing a robotic arm device for oil boom vessels with distributed attitude perception, real-time dynamic compensation, and intelligent grasping functions has become the key to overcoming the limitations of existing technologies. Summary of the Invention

[0008] The purpose of this invention is to provide an intelligent oil boom boom robotic arm device and its control method. This device can quickly, accurately, and reliably grasp oil booms, thereby improving the efficiency and safety of oil boom deployment.

[0009] The technical solution adopted to achieve the present invention is: an intelligent oil boom boat robotic arm device, including a robotic arm, a hydraulic drive system, a sensing system and a control unit;

[0010] The robotic arm includes a transmission base, a steering column, a large arm, a small arm, a flipping head, and an end gripper assembly connected in sequence.

[0011] The hydraulic drive system includes a hydraulic pump station, a proportional reversing solenoid valve group, and multiple actuator cylinders to provide power to the robotic arm. The multiple actuator cylinders include a column drive cylinder, a boom drive cylinder, a forearm drive cylinder, a tilting head drive cylinder, and a hook drive cylinder. The hydraulic pump station is connected to the proportional reversing solenoid valve group through hydraulic lines, and the proportional reversing solenoid valve group is connected to each actuator cylinder through hydraulic lines to realize multi-degree-of-freedom movement and grasping of the robotic arm.

[0012] The sensing system includes a gyroscope array and a camera array. The gyroscope array includes multiple gyroscopes respectively installed on the transmission base, steering column, upper arm, lower arm and tilting head to form a distributed attitude sensing network to sense the attitude changes of each part of the robotic arm. The camera array includes multiple sets of binocular cameras respectively installed on the upper arm, lower arm and tilting head to form a distributed visual sensing network to acquire sea surface images.

[0013] The control unit is electrically connected to the gyroscope array, camera array, and proportional solenoid valve group, respectively. The control unit identifies and locates the target to be grasped based on the sea surface image acquired by the distributed visual perception network, performs motion planning for the robotic arm, and then controls the movement of the robotic arm to grasp the target by adjusting the proportional solenoid valve group. During the target positioning process, the control unit senses the attitude changes of various parts of the robotic arm in real time caused by the swaying of the hull due to the wave surge based on the distributed attitude perception network, and then dynamically compensates for the posture error of the robotic arm.

[0014] Furthermore, the transmission base consists of a flange base and a gearbox mounted thereon; the gearbox is fixedly installed on the deck of the hull via the flange base at the bottom, and the gearbox is connected to the column drive cylinder and the transmission gear at the bottom of the steering column via its internal rack and gear transmission assembly, so as to drive the steering column to rotate horizontally under the drive of the column drive cylinder.

[0015] Furthermore, the upper end of the steering column is rotatably connected to the rear end of the boom, and the front end of the boom is rotatably connected to the rear end of the forearm; the tilting head includes a first link, a second link, and a hook connecting seat, the rear end of the first link is rotatably connected to the front of the forearm, the front end of the first link is rotatably connected to the rear end of the second link, and the front and rear ends of the hook connecting seat are rotatably connected to the front end of the second link and the front end of the forearm, respectively; the end hook assembly includes a hook rotating hydraulic motor, a hook bracket, and two hooks, the hook bracket is mounted on the hook connecting seat via the hook rotating hydraulic motor, and the two hooks are arranged opposite each other and rotatably connected to the hook bracket; the hydraulic pump station is connected to the hook rotating hydraulic motor via hydraulic pipelines and a proportional reversing solenoid valve group to drive its rotation;

[0016] The boom drive cylinder is installed between the steering column and the boom to drive the boom to rotate relative to the steering column; the forearm drive cylinder is installed between the boom and the forearm to drive the forearm to rotate relative to the boom; the tilting head drive cylinder is installed between the forearm and the tilting head to drive the tilting head to rotate relative to the forearm; the hook drive cylinder is installed between the two hooks to drive the two hooks to open and close to perform the gripping action.

[0017] Furthermore, both the boom and forearm are lightweight, hollow structural components; the rear end of the boom drive cylinder is rotatably connected to the lower end of the steering column, and the front end of the boom drive cylinder's telescopic rod is rotatably connected to the rear of the boom; the rear end of the forearm drive cylinder is rotatably connected to the middle of the boom, and the front end of the forearm drive cylinder's telescopic rod is rotatably connected to the rear of the forearm; the rear end of the tilting head drive cylinder is rotatably connected to the middle and rear of the forearm, and the front end of the tilting head drive cylinder's telescopic rod is rotatably connected to the rotating joint between the first and second connecting rods.

[0018] Furthermore, the pitch angle range of the upper arm relative to the steering column is 0-90°; the pitch angle range of the lower arm relative to the upper arm is 0-60°; and the pitch angle range of the tilting head relative to the lower arm is 0-50°.

[0019] Furthermore, a pressure sensor is provided on the hook of the end gripper assembly, and the control unit is electrically connected to the pressure sensor to detect the force on the hook and adjust the gripping force accordingly.

[0020] Furthermore, the control unit includes:

[0021] The power module is used to convert input electrical energy into power at different voltage levels and to supply power to various electrical components and modules in the device.

[0022] The signal receiving module supports multiple communication protocols and is used to receive data from the gyroscope, binocular camera, and pressure sensor on the end effector assembly, as well as control commands sent by the wireless remote controller.

[0023] The data processing module is used to process, calculate and analyze the data received by the signal receiving module, and then output corresponding control commands to the signal output module.

[0024] The signal output module is used to output control signals to each proportional solenoid valve to control the movement of the robotic arm and perform grasping operations.

[0025] The present invention also provides a control method for the above-mentioned intelligent oil boom boat robotic arm device, comprising:

[0026] S1. System parameter calibration; setting the global coordinate system with the hull's center of mass as the origin. The attitude angles of the hull coordinate system are defined according to the global coordinate system. The system measures the hull's attitude angles around the x, y, and z axes using gyroscopes on the steering column and converts them into attitude angle vectors. A robotic arm base coordinate system is established with a set position on the transmission base as the origin. Within this coordinate system, the kinematic chain state of each joint of the robotic arm is defined. The angular velocity and rotation angle of each joint are measured by gyroscopes on each joint, used to detect and provide feedback on the motion state of each joint. The robotic arm's recovery pose is preset and calibrated. The end effector coordinate system is set based on the robotic arm's attitude position and size information. The hull coordinate system attitude angles are defined. Each gyroscope in the distributed attitude perception network performs self-calibration and initialization.

[0027] S2. Target Recognition and Localization: Multiple binocular cameras in the distributed visual perception network scan the target on the sea surface, namely the oil boom buoys, acquiring sea surface images through the binocular cameras. Then, the YOLOv5-SPF model algorithm is used to identify the target from the acquired sea surface images. The target data identified by each group of binocular cameras in the distributed visual perception network are processed by a spatiotemporal alignment fusion algorithm to locate the coordinates of the oil boom buoys. ;

[0028] S3. Motion Planning and Attitude Compensation: First, establish a motion disturbance model of the hull, then calculate the offset of the robotic arm base caused by changes in the hull attitude. Then, the end effector error is calculated by fusing data from five gyroscopes. Parameter compensation used for target trajectory generation; based on the coordinates of the oil boom buoys. and end effector error The target coordinates are obtained after fusion and compensation. Next, motion planning is performed on each joint of the robotic arm; the maximum speed of the robotic arm is calculated based on the maximum flow rate of the proportional directional solenoid valve, so that the speed of each joint of the robotic arm can be indirectly controlled by controlling the opening of the proportional directional solenoid valve; data from 5 gyroscopes are integrated, and a Jacobian matrix error propagation model of the robotic arm kinematic chain is established. Inverse kinematics algorithm is used to perform inverse kinematics calculation on each joint, and the required rotation angle of each joint at different times is planned during the process of the end effector reaching the position of the oil boom float; the obtained joint motion angle signal is converted into the control of the opening of the proportional directional solenoid valve through feedforward-feedback composite control, which indirectly controls the motion angle of each joint of the robotic arm; during the execution of the hydraulic drive system, sliding mode control is used to suppress hydraulic disturbances and adaptive dynamic adjustment of hydraulic cylinder movement is used for load change change; when the end effector reaches the top of the oil boom float, the robotic arm and hook are controlled to grab the target downwards, and then move in reverse to put the target back to the preset recovery position.

[0029] Furthermore, the implementation method of the YOLOv5-SPF model algorithm is as follows:

[0030] 1) Utilizing spatial pyramid pooling for rapid layering Multi-scale pooling is performed on the input feature map Fin to extract multi-scale features;

[0031] 2) Deformable convolutional networks using deformable kernels To improve the detection accuracy of deformed targets by handling non-rigid deformation of the target;

[0032] 3) Utilizing coordinate attention mechanism Attention weighting is applied to the spatial coordinate information of the input feature map Fin to enhance the spatial location features of the target.

[0033] 4) By using the feature concatenation algorithm Concat, the output feature maps of SPPF, DCNv2, and CoordAtt are concatenated along the channel dimension to integrate three features: multi-scale, deformation adaptation, and spatial attention.

[0034] 5) Use the convolutional layer Conv to perform convolution operations on the stitched feature map, adjust the number of channels and integrate the features to obtain the final output feature Fout, which is the bounding box data of the oil boom pontoon.

[0035] Furthermore, during motion planning and attitude compensation, a dynamic coordinate transformation algorithm is used to calculate the offset of the robotic arm base caused by changes in the hull attitude. The calculation formula is:

[0036]

[0037] in, For the rotation tensor, derived from the hull attitude angle generate; This represents the lever vector from the center of mass of the hull to the base of the robotic arm; Measure noise for the gyroscope;

[0038] The end effector error is calculated by fusing data from five gyroscopes. The calculation formula is as follows:

[0039]

[0040] in, Indicates the first The end-position error estimated by each gyroscope Indicates the weighting coefficient. Indicates the first Measurement variance of each gyroscope It is a natural constant. The time decay coefficient, This is the current timestamp;

[0041] Based on the coordinates of the oil boom pontoons and end effector error The target coordinates are obtained after fusion and compensation. The calculation formula is as follows:

[0042]

[0043] Then, the motion planning of each joint of the robotic arm is performed using the fifth-order polynomial method to generate the trajectory from the current position to the compensated target position, the expression of which is:

[0044]

[0045] in, Represents the total time of the trajectory, coefficient Determined by boundary conditions, ;

[0046] The maximum speed of the robotic arm is:

[0047]

[0048] in, Indicates the valve's maximum flow rate. Represents the flow coefficient. Indicates the maximum opening area of ​​the valve. This indicates the maximum pressure difference across the valve. Indicates the density of hydraulic oil. Indicates the effective area of ​​the piston;

[0049] In feedforward-feedback composite control, the controller output signal is represented as:

[0050]

[0051] in, Indicates the trajectory tracking error. The proportional and differential gain matrices;

[0052] Sliding mode control is used to suppress hydraulic disturbances, and its expression is:

[0053]

[0054] in, This represents the final output signal of the sliding mode controller. This represents the gain coefficient in the control law. Indicates the sliding surface. Indicates the slope of the sliding surface. This represents the original trajectory tracking error. express The derivative of Indicates the boundary layer thickness. Represents a saturation function;

[0055] When using adaptive dynamic adjustment of hydraulic cylinder movement based on load mutation, if the oil pressure mutation of the hook drive cylinder exceeds a set threshold, then:

[0056]

[0057] in, This represents the updated proportional gain. Indicates the base proportional gain. Indicates the amount of load change. , This represents the load capacity converted from hydraulic pressure. This represents the initial value of the load force converted from hydraulic pressure. ,in The inner diameter of the hydraulic cylinder. The diameter of the piston rod. The pressure difference between the two chambers of the hydraulic cylinder; This is the adjustment coefficient; This indicates the amount of change in the hull's attitude.

[0058] Compared with existing technologies, the present invention has the following beneficial effects: The present invention provides an intelligent oil boom vessel robotic arm device and its control method. Through innovative design of the structure of the robotic arm, hydraulic drive system, sensing system and control unit, and combined with software implementation of a control method based on autonomous identification and positioning and dynamic attitude compensation, it can achieve fast, accurate and reliable intelligent grasping of oil booms, replacing manual deck operation and realizing unmanned deployment. It not only greatly shortens the time for grasping and deploying oil booms and effectively improves the efficiency of oil boom deployment, but also effectively improves work safety. It can also adapt to continuous operation in sea state 3. Therefore, it has strong practicality and broad application prospects. Attached Figure Description

[0059] Figure 1 This is a schematic diagram of the structure of the intelligent oil boom boom robotic arm device provided in an embodiment of the present invention. Figure 1 ;

[0060] Figure 2 This is a schematic diagram of the structure of the intelligent oil boom boom robotic arm device provided in an embodiment of the present invention. Figure 2 ;

[0061] Figure 3 This is a schematic diagram of the hydraulic drive system in an embodiment of the present invention;

[0062] Figure 4 This is a flowchart illustrating the implementation of the control method for the intelligent oil boom boat robotic arm device provided in this embodiment of the invention.

[0063] In the diagram: 1-Transmission base; 2-Steering column; 3-Upright arm; 4-Lean arm; 5-Tilting head; 6-End hook assembly; 7-Column drive cylinder; 8-Upright arm drive cylinder; 9-Lean arm drive cylinder; 10-Tilting head drive cylinder; 11-Hook drive cylinder; 12-First gyroscope; 13-Second gyroscope; 14-Third gyroscope; 15-Fourth gyroscope; 16-Fifth gyroscope; 17-First set of binocular cameras; 18-Second set of binocular cameras; 19-Third set of binocular cameras; 20-Flange base; 21- - Gearbox; 22- First connecting rod; 23- Second connecting rod; 24- Grappling hook connecting seat; 25- Grappling hook rotary hydraulic motor; 26- Grappling hook bracket; 27- Grappling hook; 28- Oil reservoir; 29- Hydraulic pump; 30- Pressure regulating valve; 31- Hydraulic distributor; 32- Oil pressure sensor; 33- Solenoid pressure regulating valve; V1- First proportional reversing solenoid valve; V2- Second proportional reversing solenoid valve; V3- Third proportional reversing solenoid valve; V4- Fourth proportional reversing solenoid valve; V5- Fifth proportional reversing solenoid valve; V6- Sixth proportional reversing solenoid valve. Detailed Implementation

[0064] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0065] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0066] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0067] like Figure 1-2 As shown, this embodiment provides an intelligent oil boom boat robotic arm device, including a robotic arm, a hydraulic drive system, a sensing system, and a control unit.

[0068] The robotic arm includes a transmission base 1, a steering column 2, a large arm 3, a small arm 4, a flipping head 5, and an end gripping hook assembly 6, which are connected in sequence.

[0069] The hydraulic drive system includes a hydraulic pump station, a proportional directional solenoid valve assembly, and multiple actuator cylinders to provide power to the robotic arm. The multiple actuator cylinders include a column drive cylinder 7, a boom drive cylinder 8, a forearm drive cylinder 9, a tilting head drive cylinder 10, a hook rotation hydraulic motor 25, and a hook drive cylinder 11. The hydraulic pump station is connected to the proportional directional solenoid valve assembly via hydraulic lines, and the proportional directional solenoid valve assembly is connected to each actuator cylinder via hydraulic lines to achieve multi-degree-of-freedom movement and grasping of the robotic arm.

[0070] Figure 3 This is a schematic diagram of the hydraulic drive system in this embodiment. Figure 3 As shown, in this embodiment, the hydraulic pump station consists of an oil reservoir 28, a hydraulic pump 29, a pressure regulating valve 30, and a hydraulic distributor 31. The hydraulic pump station is connected to the column drive cylinder 7, the boom drive cylinder 8, the forearm drive cylinder 9, the tilting head drive cylinder 10, the hook rotation hydraulic motor 25, and the hook drive cylinder 11 via proportional directional solenoid valves V1, V2, V3, V4, V5, and V6 and corresponding hydraulic pipelines. By adjusting the opening of each proportional directional solenoid valve group, the corresponding cylinders are driven to work. An oil pressure sensor 32 and an electromagnetic pressure regulating valve 33 are installed on the hydraulic pipeline between the proportional directional solenoid valve V6 and the hook drive cylinder 11 to detect and regulate the oil pressure of the hook drive cylinder 11.

[0071] The sensing system includes a gyroscope array and a camera array. The gyroscope array includes multiple gyroscopes 12, 13, 14, 15, and 16 respectively installed on the transmission base, steering column 2, upper arm 3, lower arm 4, and tilting head 5 to form a distributed attitude sensing network to sense attitude changes of various parts of the robotic arm. The camera array includes multiple sets of binocular cameras 17, 18, and 19 respectively installed on the upper arm 3, lower arm 4, and tilting head 5 to form a distributed visual sensing network to acquire sea surface images.

[0072] The control unit is electrically connected to the gyroscope array, camera array, and proportional solenoid valve group, respectively. The control unit identifies and locates the target to be grasped based on the sea surface image acquired by the distributed visual perception network, performs motion planning for the robotic arm, and then controls the movement of the robotic arm to grasp the target by adjusting the proportional solenoid valve group. During the target positioning process, the control unit senses the attitude changes of various parts of the robotic arm in real time caused by the swaying of the hull due to the wave surge based on the distributed attitude perception network, and then dynamically compensates for the posture error of the robotic arm.

[0073] The transmission base 1 consists of a flange base 20 and a gearbox 21 mounted thereon. The gearbox 21 is fixedly installed on the deck of the hull via the flange base 20 at the bottom. The gearbox 21 is connected to the column drive cylinder 7 and the transmission gear at the bottom of the steering column 2 via its internal rack and gear transmission assembly, so as to drive the steering column to rotate horizontally under the drive of the column drive cylinder.

[0074] The upper end of the steering column 2 is rotatably connected to the rear end of the boom 3, and the front end of the boom 3 is rotatably connected to the rear end of the forearm 4. The flipping head 5 includes a first connecting rod 22 (arc-shaped structure in this embodiment), a second connecting rod 23, and a hook connecting seat 24. The rear end of the first connecting rod 22 is rotatably connected to the front part of the forearm 4, and the front end of the first connecting rod 22 is rotatably connected to the rear end of the second connecting rod 23. The front and rear ends of the hook connecting seat 24 are rotatably connected to the front end of the second connecting rod 23 and the front end of the forearm 4, respectively. The end hook assembly 6 includes a hook rotating hydraulic motor 25, a hook bracket 26, and two hooks 27. The hook bracket 26 is mounted on the hook connecting seat 24 via the hook rotating hydraulic motor 25. The two hooks 27 are arranged opposite each other and rotatably connected to the hook bracket 26, respectively. The hydraulic pump station is connected to the hook rotating hydraulic motor 25 via hydraulic pipelines and a proportional reversing solenoid valve group to drive its rotation.

[0075] The boom drive cylinder 8 is installed between the steering column 2 and the boom 3 to drive the boom to rotate relative to the steering column; the forearm drive cylinder 9 is installed between the boom 3 and the forearm 4 to drive the forearm to rotate relative to the boom; the tilting head drive cylinder 10 is installed between the forearm 4 and the tilting head 5 to drive the tilting head to rotate relative to the forearm; the hook drive cylinder 11 is installed between the two hooks to drive the two hooks 27 to open and close to perform the gripping action.

[0076] In this embodiment, both the upper arm 3 and the lower arm 4 are lightweight, hollow structural components. The rear end of the cylinder body of the upper arm drive cylinder 8 is rotatably connected to the lower end of the steering column 2, and the front end of the telescopic rod of the upper arm drive cylinder 8 is rotatably connected to the rear of the upper arm 3; the rear end of the cylinder body of the lower arm drive cylinder 9 is rotatably connected to the middle of the upper arm 3, and the front end of the telescopic rod of the lower arm drive cylinder 9 is rotatably connected to the rear of the lower arm 4; the rear end of the cylinder body of the tilting head drive cylinder 10 is rotatably connected to the middle and rear of the lower arm 4, and the front end of the telescopic rod of the tilting head drive cylinder 10 is rotatably connected to the rotatable joint between the first connecting rod 22 and the second connecting rod 23.

[0077] Through the above design, this device can achieve the following: the pitch angle range of the upper arm relative to the steering column is 0-90°; the pitch angle range of the lower arm relative to the upper arm is 0-60°; and the pitch angle range of the tilting head relative to the lower arm is 0-50°.

[0078] In this embodiment, a pressure sensor is provided on the hook of the end gripper assembly 6, and the control unit is electrically connected to the pressure sensor to detect the force on the hook and adjust the magnitude of the gripping force.

[0079] In this embodiment, the control unit includes:

[0080] 1) Power module, used to convert input electrical energy into power of different voltage levels and to power various electrical components and modules in the device;

[0081] 2) Signal receiving module, which supports multiple communication protocols, is used to receive data from the gyroscope, binocular camera and pressure sensor on the end effector, as well as control commands sent by the wireless remote controller.

[0082] 3) Data processing module, used to process, calculate and analyze the data received by the signal receiving module, and then output corresponding control commands to the signal output module;

[0083] 4) Signal output module, used to output control signals to each proportional solenoid valve to control the movement of the robotic arm and perform grasping operations.

[0084] like Figure 4 As shown, this embodiment also provides a control method for the above-mentioned intelligent oil boom boat robotic arm device, including:

[0085] S1, System Parameter Calibration

[0086] 1) After the robotic arm is installed on the ship, set the global coordinate system with the ship's center of mass as the origin. The attitude angles of the hull coordinate system are defined according to the global coordinate system. The attitude angles of the hull around the x, y, and z axes (usually corresponding to roll, pitch, and yaw) are measured by the gyroscope on the transmission base and converted into attitude angle vectors, which serve as feedback for subsequent attitude compensation.

[0087] 2) The position is set on the transmission base (generally the center position of the flange base, and...). via lever arm vector (Associated calculations) The robot arm base coordinate system is set as the origin, and the motion chain state of each joint of the robot arm is defined in the robot arm base coordinate system. The angular velocity and rotation angle of each joint are measured by the gyroscope on each joint of the robot arm, which is used to detect and provide feedback on the motion state of each joint of the robot arm.

[0088] 3) Preset and calibrate the robotic arm's retrieval pose. Specifically, the robotic arm is set in a predetermined posture (referred to as the zero-position posture, with the joint angle θ=0° in this posture). The camera captures a deck marker point (which also serves as the preset retrieval position). The retrieval pose is calculated using a hand-eye calibration algorithm. Specifically, by utilizing the rotation matrix of the camera coordinate system relative to the robotic arm's end-effector coordinate system (hand-eye rotation) and the translation vector of the camera coordinate system relative to the robotic arm's end-effector coordinate system (hand-eye translation), combined with minimizing projection error and regularization, the transformation matrix of the camera relative to the robotic arm's end-effector is solved to achieve the preset retrieval pose calibration (i.e., the accurate mapping between "the target position seen by the camera" and "the position where the robotic arm needs to move").

[0089] 4) Set the coordinate system of the end effector based on the posture and size information (arm span) of the robotic arm.

[0090] 5) Define the attitude angles of the hull coordinate system. After the coordinates are determined, each gyroscope in the distributed attitude perception network performs self-calibration and initialization.

[0091] S2, Target Identification and Localization

[0092] A distributed visual perception network is constructed using three sets of 5-megapixel binocular cameras (a total of 6 cameras). The effective field of view is expanded to 120° through a multi-view geometric network, eliminating the blind spot of a single camera.

[0093] After the system is powered on, it begins initialization and self-test. Once the self-test is normal, it enters the predetermined roving scanning and recognition program. The robotic arm begins omnidirectional roving swings according to the pre-programmed sequence. Three sets of binocular cameras in the distributed visual perception network scan the target on the sea surface, namely the oil boom buoys, acquiring sea surface images through the binocular cameras, and then identifying the target from the acquired sea surface images using the YOLOv5-SPF model algorithm. The implementation method of the YOLOv5-SPF model algorithm is as follows:

[0094] 1) Utilizing spatial pyramid pooling for rapid layering Multi-scale pooling (such as 1×1, 2×2, 3×3 pooling) is performed on the input feature map Fin to extract multi-scale features (adapting to targets of different sizes, such as different viewing angles of buoys);

[0095] 2) Deformable convolutional networks employing deformable convolutional kernels (the position of the convolutional kernels adjusts according to the shape of the target). It processes non-rigid deformations of targets (such as the tilting or rotation of pontoons in water) to improve the detection accuracy of deformed targets;

[0096] 3) Utilizing coordinate attention mechanism Attention weighting is applied to the spatial coordinate information (x / y axis) of the input feature map Fin to enhance the spatial location features of the target (such as the edges and corners of the pontoon) and reduce background interference;

[0097] 4) The output feature maps of SPPF, DCNv2, and CoordAtt are concatenated along the channel dimension using the feature concatenation algorithm Concat (e.g., SPPF outputs channel C1, DCNv2 outputs channel C2, and CoordAtt outputs channel C3, resulting in channel C1+C2+C3 after concatenation) to integrate three features: multi-scale, deformation adaptation, and spatial attention.

[0098] 5) Use the convolutional layer Conv to perform convolution operations on the stitched feature map, adjust the number of channels and integrate the features to obtain the final output feature Fout, which is the bounding box data of the oil boom pontoon.

[0099] A major innovation of this method lies in the application of the YOLOv5-SPF model algorithm, which solves the problem of detection accuracy of traditional YOLOv5 in multi-scale targets, deformable targets, and complex backgrounds, and improves the detection accuracy of targets such as pontoons (96.7% mAP@0.5 for pontoon detection) and inference speed (42ms / frame).

[0100] Then, the target data identified by each group of binocular cameras in the distributed visual perception network are processed using a spatiotemporal alignment fusion algorithm to locate the coordinates of the oil boom pontoons. Specifically, first, a view index i is defined, with three sets of binocular cameras, totaling six cameras, each camera corresponding to a view i; then, the triangulation function Triangulate(ui,vi) is used in conjunction with the parallax of the binocular cameras to calculate the three-dimensional coordinates X of the oil boom pontoon. target Next, by integrating motion planning algorithms and gyroscope-based dynamic compensation algorithms, the target is precisely locked, and preparation is made to capture the oil boom pontoon.

[0101] S3, Motion Planning and Posture Compensation

[0102] Motion planning and attitude compensation work together, with compensation and motion planning implemented based on changes in the hull's attitude angle.

[0103] 1) Motion attitude compensation calculations are performed based on the influence of waves on the hull, serving as feedforward compensation for motion planning. First, a hull motion disturbance model is established. Then, a dynamic coordinate transformation algorithm is used to calculate the offset of the robotic arm base caused by changes in hull attitude. The calculation formula is:

[0104]

[0105] in, For the rotation tensor, derived from the hull attitude angle generate; This represents the lever vector (mechanical fixed parameter) from the center of mass of the hull to the base of the robotic arm. To measure the noise of the gyroscope.

[0106] Then, the end effector error is calculated by fusing data from five gyroscopes. This is used for parameter compensation in target trajectory generation; its calculation formula is:

[0107]

[0108] in, Indicates the first The end-position error estimated by each gyroscope This represents the weighting coefficient, which comprehensively measures the reliability and data timeliness of the gyroscope, and determines the weighting coefficient. The contribution of the end-position error estimated by each gyroscope to the fusion result. Indicates the first The measurement variance of each gyroscope (reflecting sensor reliability). It is a natural constant. This is the time decay factor (typical value 0.1~0.5). This is the current timestamp (in seconds).

[0109] Based on the coordinates of the oil boom pontoons and end effector error The target coordinates are obtained after fusion and compensation. The calculation formula is as follows:

[0110]

[0111] Then, the motion planning of each joint of the robotic arm is performed using the fifth-order polynomial method to generate the trajectory from the current position to the compensated target position, the expression of which is:

[0112]

[0113] in, Represents the total trajectory time (usually specified as 15 seconds), coefficient ( Determined by boundary conditions:

[0114] Start / End Point: , ;

[0115] Start / End Speed: , ;

[0116] Initial / End Acceleration: , .

[0117] Next, the maximum speed of the robotic arm is calculated based on the maximum flow rate of the proportional solenoid valve, so that the movement speed of each joint of the robotic arm can be indirectly controlled by controlling the opening of the proportional solenoid valve in the future.

[0118] The maximum speed of the robotic arm is:

[0119]

[0120] in, This indicates the valve's maximum flow rate (m³ / s). Indicates the flow coefficient (determined by valve structure). This indicates the maximum opening area of ​​the valve (m²). This indicates the maximum pressure difference (Pa) across the valve. This indicates the density of hydraulic oil (kg / m³). This represents the effective area of ​​the piston (m²).

[0121] By integrating data from five gyroscopes, and establishing a Jacobian matrix error propagation model for the robotic arm's kinematics chain, an inverse kinematics algorithm (Levenberg-Marquardt algorithm) is used to perform inverse kinematics calculations on each joint, thus planning the required rotation angle of each joint at different times during the process of the end effector reaching the location of the oil boom float.

[0122] Solve for joint angles :

[0123] in, Represents the positive kinematic function. Jacobian matrix ( ), This is the damping coefficient (to prevent singularities).

[0124] The obtained joint motion angle signals are converted into the opening degree of the proportional directional solenoid valve through feedforward-feedback composite control, thereby indirectly controlling the motion angles of each joint of the robotic arm. In feedforward-feedback composite control, the controller output signal is represented as:

[0125]

[0126] Among them, feedback items:

[0127] Indicates trajectory tracking error;

[0128] Let be the proportional and differential gain matrices.

[0129] Feedforward term:

[0130] This refers to the end effector error.

[0131] During the execution of the hydraulic drive system, sliding mode control is used to suppress hydraulic disturbances and adaptive dynamic adjustment of the hydraulic cylinder movement is adopted based on load changes.

[0132] Sliding mode control is used to suppress hydraulic disturbances, and its expression is:

[0133]

[0134] in, This represents the final output signal of the sliding mode controller. This represents the gain coefficient in the control law. Indicates the sliding surface. This indicates the slope of the sliding surface (>0). This represents the original trajectory tracking error. express The derivative of Indicates the boundary layer thickness (to suppress chattering). This represents a saturation function.

[0135] When the hydraulic cylinder movement is dynamically adjusted adaptively to adapt to sudden load changes, if the oil pressure of the hook drive cylinder (measured by an oil pressure sensor installed on the front side of the hook drive cylinder) changes beyond a set threshold, then:

[0136]

[0137] in, This represents the updated proportional gain. This represents the base proportional gain (a pre-set initial value to ensure control stability under normal conditions). This represents the load change (N). , This represents the load capacity converted from hydraulic pressure. This represents the initial value of the load force converted from hydraulic pressure. ,in The inner diameter of the hydraulic cylinder is (m). The diameter of the piston rod is (m). The pressure difference between the two chambers of the hydraulic cylinder (Pa); The adjustment factor (unit: 1 / (N·rad), where N is the unit of force and rad is the unit of radians). This represents the change in the hull's attitude (rad).

[0138] When the end effector reaches above the oil boom pontoon, it controls the robotic arm and hook to grab the target downwards, and then moves in reverse to place the target back to the preset recovery position.

[0139] This device has demonstrated its reliability in multiple simulated exercises across various sea areas, adapting to continuous operations under sea state 3 and showcasing the development direction of intelligent marine emergency equipment. In actual operations, the system exhibits excellent environmental adaptability and control precision, stably completing oil boom grabbing and docking tasks even in complex sea conditions with gusts reaching level 3 and wave heights exceeding 0.5 meters. In multi-machine collaborative operation scenarios, the devices achieve state synchronization through a highly interference-resistant frequency-hopping communication protocol, further improving joint operation efficiency. Recent tests show that the system can operate continuously for 24 hours without failure in nearshore environments with high salt spray concentrations, with key components achieving an IP68 protection level. Through iterative optimization of the grabbing strategy using reinforcement learning algorithms, the system has acquired the ability to autonomously adapt to oil booms of different sizes. The success rate of the hydraulic robotic arm in completing multi-angle, multi-pose docking tasks in dynamic environments has increased to 98.5%. In anti-interference tests, the device operated continuously for two hours in a strong electromagnetic interference environment with a communication packet loss rate of less than 0.1%, significantly enhancing data link stability. The system's startup performance in low-temperature environments was also verified. At -10℃, the hydraulic oil flow remained stable, and the actuator response speed showed no significant decrease.

[0140] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method of controlling an intelligent boom arm apparatus of a boat machine, characterized by, The intelligent oil fence boat mechanical arm device comprises a mechanical arm, a hydraulic drive system, a sensing system and a control unit; The mechanical arm comprises a transmission base, a steering column, a large arm, a small arm, a turnover head and a terminal hook assembly which are connected in sequence; The hydraulic drive system comprises a hydraulic pump station, a proportional reversing electromagnetic valve group and a plurality of execution oil cylinders for providing power for the mechanical arm; the plurality of execution oil cylinders comprises a column driving oil cylinder, a large arm driving oil cylinder, a small arm driving oil cylinder, a turnover head driving oil cylinder and a hook driving oil cylinder; the hydraulic pump station is connected with the proportional reversing electromagnetic valve group through a hydraulic pipeline; the proportional reversing electromagnetic valve group is connected with each execution oil cylinder through a hydraulic pipeline to realize multi-degree-of-freedom movement and grabbing of the mechanical arm; The sensing system comprises a gyroscope array and a camera array; the gyroscope array comprises a plurality of gyroscopes arranged on the transmission base, the steering column, the large arm, the small arm and the turnover head respectively to form a distributed posture sensing network to sense posture changes of each part of the mechanical arm; the camera array comprises a plurality of binocular cameras arranged on the large arm, the small arm and the turnover head respectively to form a distributed visual sensing network to obtain sea surface images; The control unit is electrically connected with the gyroscope array, the camera array and the proportional reversing electromagnetic valve group respectively; the control unit identifies and locates the target to be grabbed based on the sea surface images obtained by the distributed visual sensing network, plans the movement of the mechanical arm, and then controls the movement of the mechanical arm and the grabbing of the target by regulating the proportional reversing electromagnetic valve group. The control method of the intelligent oil fence boat mechanical arm device comprises: S1, system parameter calibration; set the global coordinate system with the center of the boat body as the origin , and define the boat body coordinate system attitude angle according to the global coordinate system , measure the attitude angle of the boat body around the x, y, z axes through the gyroscope on the steering column, and convert it into an attitude angle vector; set the mechanical arm base coordinate system with the set position on the transmission base as the origin, and define the motion chain state of each joint of the mechanical arm in the mechanical arm base coordinate system, and each joint state vector is measured by the gyroscope on each joint of the mechanical arm angular velocity and rotation angle, which is used to detect and feedback the motion state of each joint of the mechanical arm; preset the recovery pose of the mechanical arm and calibrate it; set the end effector coordinate system according to the attitude position and size information of the mechanical arm; define the boat body coordinate system attitude angle; each gyroscope in the distributed attitude perception network is self-calibrated and initialized; S2, target recognition and positioning; a plurality of binocular cameras in the distributed visual perception network scan the target on the sea surface, i.e. the oil containment boom float, acquire the sea surface image through the binocular camera, and then recognize the target from the acquired sea surface image through the YOLOv5-SPPF model algorithm; the target data recognized by each group of binocular cameras in the distributed visual perception network is processed through a space-time alignment fusion algorithm to locate the coordinates of the oil containment boom float ; S3. Motion Planning and Attitude Compensation: First, establish a motion disturbance model of the hull, then calculate the offset of the robotic arm base caused by changes in the hull attitude. Then, the end effector error is calculated by fusing data from five gyroscopes. Parameter compensation used for target trajectory generation; based on the coordinates of the oil boom buoys. and end effector error The target coordinates are obtained after fusion and compensation. Then, motion planning is performed on each joint of the robotic arm; In the process of motion planning and posture compensation, the dynamic coordinate transformation algorithm is used to calculate the base offset of the manipulator caused by the change of the ship posture , and the calculation formula is: wherein, is the rotation tensor, by the hull attitude angles generated; denotes the link vector from the hull center of mass to the manipulator base; is the gyroscope measurement noise; Utilizing five gyroscope data fusion to calculate end effector error The calculation formula is: wherein, represents the end position error estimated by the represents a weight coefficient, represents the measurement variance of the is a natural constant, is a time decay coefficient, is a current timestamp;​​ According to the coordinates of the oil containment boom buoy And end effector error Fusion to obtain the compensated target coordinates The calculation formula is: 。 2. The method of claim 1, wherein, The transmission base is composed of a flange base and a gear box arranged thereon; the gear box is fixedly installed on the deck of the boat body through the flange base at the bottom; the gear box is matched with the column driving oil cylinder and the transmission gear at the bottom of the steering column through the internal rack and pinion transmission assembly to drive the steering column to rotate horizontally under the driving of the column driving oil cylinder.

3. The method of claim 1, wherein, The upper end of the steering column is rotationally connected with the rear end of the large arm, and the front end of the large arm is rotationally connected with the rear end of the small arm; the turnover head comprises a first connecting rod, a second connecting rod and a hook connecting seat; the rear end of the first connecting rod is rotationally connected with the front part of the small arm, the front end of the first connecting rod is rotationally connected with the rear end of the second connecting rod, and the front and rear ends of the hook connecting seat are respectively rotationally connected with the front end of the second connecting rod and the front end of the small arm; the terminal hook assembly comprises a hook rotating hydraulic motor, a hook support and two hooks; the hook support is installed on the hook connecting seat through the hook rotating hydraulic motor, and the two hooks are arranged opposite to each other and are rotationally connected with the hook support respectively; the hydraulic pump station is connected with the hook rotating hydraulic motor through the hydraulic pipeline and the proportional reversing electromagnetic valve group to drive the rotation thereof. The large arm driving oil cylinder is installed between the steering column and the large arm to drive the large arm to rotate relative to the steering column; the small arm driving oil cylinder is installed between the large arm and the small arm to drive the small arm to rotate relative to the large arm; the turnover head driving oil cylinder is installed between the small arm and the turnover head to drive the turnover head to rotate relative to the small arm; and the hook driving oil cylinder is installed between the two hooks to drive the two hooks to open and close to perform the grabbing action.

4. The method of claim 3, wherein the method further comprises: The large arm and the small arm are both hollow structural members of lightweight design; the rear end of the cylinder body of the large arm driving oil cylinder is rotationally connected with the lower end of the steering column, and the front end of the telescopic rod of the large arm driving oil cylinder is rotationally connected with the rear part of the large arm; the rear end of the cylinder body of the small arm driving oil cylinder is rotationally connected with the middle part of the large arm, and the front end of the telescopic rod of the small arm driving oil cylinder is rotationally connected with the rear part of the small arm; the rear end of the cylinder body of the turnover head driving oil cylinder is rotationally connected with the middle-rear part of the small arm, and the front end of the telescopic rod of the turnover head driving oil cylinder is rotationally connected at the rotation connection joint of the first connecting rod and the second connecting rod.

5. The method of claim 3, wherein the method further comprises: The pitch angle range of the large arm relative to the steering column is 0-90°; the pitch angle range of the small arm relative to the large arm is 0-60°; and the pitch angle range of the turnover head relative to the small arm is 0-50°.

6. The method of claim 1, wherein, The pressure sensor is arranged on the grab hook of the terminal grab hook assembly, and the control unit is electrically connected with the pressure sensor to detect the force borne by the grab hook through the pressure sensor and to regulate and control the size of the grabbing force.

7. The method of claim 1, wherein, The control unit comprises: a power supply module for converting input electric energy into power supply of different voltage levels and supplying power to each electrical component and module in the device; a signal receiving module supporting multiple communication protocols for receiving data acquired by the gyroscopes, the binocular camera and the pressure sensor on the grab hook of the terminal grab hook assembly, and receiving control instructions sent by a wireless remote controller; a data processing module for processing, calculating and analyzing the data received by the signal receiving module and then outputting corresponding control instructions to the signal output module; a signal output module for outputting control signals to each proportional reversing electromagnetic valve to control the movement of the mechanical arm and the execution of the grabbing operation.

8. The method of claim 1, wherein, Step S3 further comprises: calculating the maximum speed of the movement of the mechanical arm according to the maximum flow of the proportional reversing electromagnetic valve to indirectly control the movement speed of each joint of the mechanical arm by subsequently controlling the opening of the proportional reversing electromagnetic valve; fusing the data of the five gyroscopes, establishing a Jacobian matrix error propagation model of the kinematic chain of the mechanical arm, using inverse kinematics algorithm to solve the inverse kinematics of each joint, and planning the rotation angle required by each joint at different times during the process of the end effector reaching the position of the floating pontoon of the oil containment boom; converting the obtained joint movement angle signal into the opening of the proportional reversing electromagnetic valve through feedforward-feedback compound control to indirectly control the movement angle of each joint of the mechanical arm; and using sliding mode control to suppress hydraulic disturbance and using load mutation adaptive dynamic adjustment of hydraulic cylinder movement during the execution process of the hydraulic driving system; when the end effector reaches above the floating pontoon of the oil containment boom, controlling the mechanical arm and the grab hook to grab the target downward, and then moving reversely to put the target back to the pre-set recovery position.

9. The method of claim 1, wherein, The implementation method of the YOLOv5-SPPF model algorithm is: 1) Fast layer using spatial pyramid pooling Multi-scale pooling is performed on the input feature map Fin to extract multi-scale features; 2) deformable convolution network using deformable convolution kernel , process the non-rigid deformation of the target to improve the detection accuracy of the deformed target; 3) Coordinate attention mechanism The spatial coordinate information of the input feature map Fin is weighted by attention, and the spatial position features of the target are enhanced. 4) The output feature maps of SPPF, DCNv2 and CoordAtt are spliced along the channel dimension through a feature splicing algorithm Concat to fuse three kinds of features of multi-scale, deformation adaptation and spatial attention; 5) The final output feature Fout is obtained by using the convolution layer Conv to convolve the spliced feature map, adjusting the number of channels and integrating the features, which is the bounding box data of the floating pontoon of the oil containment boom.

10. The method of claim 8, wherein the method further comprises: The trajectory from the current position to the compensated target position is generated by using the quintic polynomial method for motion planning of each joint of the robot arm, and its expression is: wherein denotes the total time of the trajectory, the coefficient is determined by the boundary conditions, ; The maximum speed of the robot arm motion is: wherein, represents the maximum flow of the valve, represents the flow coefficient, represents the maximum opening area of the valve, represents the maximum pressure difference across the valve, represents the density of the hydraulic oil, represents the effective area of the piston; In the feedforward-feedback composite control, the controller output signal is expressed as: wherein, represents a trajectory tracking error, is a proportional, derivative gain matrix; The hydraulic disturbance is suppressed by using the sliding mode control, and its expression is: wherein represents the final output signal of the sliding mode controller, represents a gain coefficient in the control law, represents a sliding surface, represents a sliding surface slope, represents an original trajectory tracking error, represents a derivative of represents a boundary layer thickness, represents a saturation function; When the oil pressure of the grab hook driving cylinder suddenly changes and exceeds the set threshold, the load mutation self-adaptive dynamic adjustment is adopted for the hydraulic cylinder motion, and there are: wherein, represents an updated proportional gain, represents a base proportional gain, represents a load change amount, , represents an oil pressure converted load force, represents an oil pressure converted load force initial value, wherein is an oil cylinder inner diameter, is a piston rod diameter, is an oil cylinder two chamber pressure difference; is a regulation coefficient; represents a hull attitude change amount.

Images