Deep sea autonomous grabbing underwater robot and target object grabbing method thereof

Abstract

The invention provides a deep-sea autonomous grabbing underwater robot and a target object grabbing method thereof. The deep-sea autonomous grabbing underwater robot comprises an underwater robot body, a hydraulic pump station, a controller, a buoyancy adjusting device, a mechanical arm, an optical sensor and a foresight sonar. The buoyancy adjusting device is installed on the side portion of the underwater robot body, the mechanical arm, the optical sensor and the foresight sonar are all installed on the bow portion of the underwater robot body, and the controller and the hydraulic pump station are both carried on the underwater robot body. The hydraulic pump station is used for driving the mechanical arm, and the hydraulic pump station, the optical sensor and the foresight sonar are all in signal connection with the controller. The underwater robot is simple in structure, the acoustic sensor and the optical sensor are arranged on the underwater robot body based on the ROV or AUV platform, after a target is searched and recognized and the pose of the target is locked through optical recognition, the robot adjusts the pose and hovers, and the mechanical arm grabs the target under the disturbance condition of the robot body. And the system reliability and the operation accuracy can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of underwater robots, and in particular to a deep-sea autonomous grasping underwater robot and a target grasping method thereof, and in particular to an underwater robot capable of accurately grasping an underwater suspended target and a target grasping method thereof. Background Art

[0002] In recent years, as my country's exploration, research and understanding of the ocean field continue to deepen, my country's investment in underwater equipment is also increasing, especially for some special underwater operating environments, underwater disposal of special targets, etc., the requirements are getting higher and higher, and the demand is increasing. This patent: A deep-sea autonomous grasping underwater robot, which has the ability to grasp autonomously underwater, can autonomously grasp some specific underwater targets or suspended targets and perform corresponding disposal.

[0003] At present, deep-water target handling mainly relies on ROV and AUV equipped with mechanical arms and related operating tools to handle underwater targets in a targeted manner. However, when these underwater robots handle deep-water suspended targets, it is often difficult to accurately grasp underwater suspended targets because the underwater robots are affected by environmental factors such as the robot's automatic control accuracy and underwater water flow when they are hovering underwater. In view of the above problems, it is urgent to develop a new type of autonomous grasping robot for underwater suspended targets. Summary of the invention

[0004] In view of the defects in the prior art, the purpose of the present invention is to provide a deep-sea autonomous grasping underwater robot and a target grasping method thereof.

[0005] The deep-sea autonomous grasping underwater robot provided by the present invention comprises an underwater robot body, a hydraulic pump station, a controller, a buoyancy adjustment device, a mechanical arm, an optical sensor and a forward-looking sonar;

[0006] The buoyancy adjustment device is installed on the side of the underwater robot body, the mechanical arm, the optical sensor and the forward-looking sonar are all installed on the bow of the underwater robot body, and the controller and the hydraulic pump station are both mounted on the underwater robot body;

[0007] The hydraulic pump station is used to drive the robotic arm, and the hydraulic pump station, optical sensor and forward-looking sonar are all connected to the controller signal.

[0008] Preferably, the buoyancy adjustment device comprises a hydraulic telescopic device, a port telescopic buoyancy block and a starboard telescopic buoyancy block;

[0009] The port side telescopic buoyancy block and the starboard side telescopic buoyancy block are respectively installed at the port side and starboard side center positions of the underwater robot body through a hydraulic telescopic device, and the port side telescopic buoyancy block and the starboard side telescopic buoyancy block are symmetrically arranged.

[0010] Preferably, the hydraulic telescopic device comprises a scissor-type hydraulic telescopic rod and a hydraulic cylinder, and the hydraulic cylinder is installed inside the central frame of the underwater robot body;

[0011] One end of the scissor-type hydraulic telescopic rod is connected to the hydraulic cylinder, and the other end of the scissor-type hydraulic telescopic rod is connected to the port telescopic buoyancy block or the starboard telescopic buoyancy block;

[0012] The underwater robot body is provided with buoyancy block accommodating chambers at the port side and starboard side center positions, respectively used to accommodate the port side telescopic buoyancy block and the starboard side telescopic buoyancy block.

[0013] Preferably, the optical sensor includes a binocular camera and / or an underwater laser detection device;

[0014] The binocular camera and the underwater laser detection device are both installed at the bow head position of the underwater robot body.

[0015] Preferably, it also includes an underwater cable cutting device, which is installed in a working basket at the bow of the underwater robot body.

[0016] Preferably, a pressure difference measuring device is installed at the lower frame position of the bow of the underwater robot body;

[0017] The pressure difference measuring device is connected to the controller and the mechanical arm respectively, and is used to detect the pressure difference between the inlet and outlet of all the actuators of the mechanical arm, and send the pressure difference information to the controller.

[0018] According to the target object grasping method of the deep-sea autonomous grasping underwater robot provided by the present invention, the deep-sea autonomous grasping underwater robot is used, and the method comprises the following steps:

[0019] Step 1: The underwater robot body searches for an underwater suspended target object through a forward-looking sonar. After the target object is found, the underwater robot body moves toward the target object until it moves to the vicinity of the target object;

[0020] Step 2: Use a binocular camera or underwater laser detection device to lock the position of the target in the deep sea, and control the underwater robot to adjust the position relationship with the target and then hover in the deep sea;

[0021] Step 3: The binocular camera or underwater laser detection device transmits the target object's posture information to the controller in real time, and the controller controls the robotic arm to perform the target object disposal action according to the target object's posture information.

[0022] Preferably, in step 1, the underwater robot body autonomously approaches the target object by manual control or by sonar detection.

[0023] Preferably, in step 1, the port side telescopic buoyancy block and the starboard side telescopic buoyancy block are both accommodated in a buoyancy block accommodating chamber of the underwater robot body;

[0024] In step 2, after the underwater robot body adjusts its position relationship with the target object, the port side telescopic buoyancy block and the starboard side telescopic buoyancy block are both symmetrically extended under the drive of the scissor-type hydraulic telescopic rod, so that the underwater robot body hovers in the deep sea.

[0025] Preferably, step 3 adopts a robot end compensation algorithm, including the following sub-steps:

[0026] Step 3.1: The controller obtains the RGB image and depth image from the optical sensor, calculates the position of the target relative to the optical sensor through the RGB image and depth image, and then obtains the position of the target relative to the base of the robotic arm through coordinate transformation;

[0027] Step 3.2: Use the RRT* algorithm to plan discretized path points based on the position of the target object relative to the robot base;

[0028] Step 3.3: Through the inverse kinematics algorithm, according to the planned path points, the joint angles of the robot arm are calculated, and the joints of the robot arm are driven to make the end of the robot arm approach the target object;

[0029] Step 3.4: using the inertial measurement unit of the manipulator base to estimate the offset of the base, and estimating the position of the target object relative to the manipulator base after the manipulator base is disturbed according to the offset;

[0030] Step 3.5: Use the single neuron adaptive PID algorithm to drive each joint of the robotic arm and compensate the end of the robotic arm;

[0031] Step 3.6: Detect the error at the end of the robot arm. When the error exceeds the preset accuracy, jump to step 3.4.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] 1. The present invention has a simple structure. Acoustic and optical sensors are configured on the underwater robot body based on the ROV or AUV platform to search and identify the target. After the optical recognition locks the target posture, the robot adjusts the posture and hovers. The mechanical arm grabs the target when the robot body is disturbed, which can effectively improve the system reliability and operation accuracy.

[0034] 2. The underwater robot body of the present invention is equipped with retractable symmetrically arranged buoyancy blocks to improve the stability of the underwater robot when it is suspended underwater, so that the underwater robot can reduce disturbances when it is hovering underwater.

[0035] 3. The present invention controls the position of the end of the robotic arm through the kinematic model of the robotic arm, and realizes real-time compensation of the end of the robotic arm in combination with an inertial measurement unit.

[0036] 4. The present invention optimizes the inlet and outlet pressure difference of the robot arm actuator through the pressure difference measuring device, thereby improving the accuracy of the robot arm force feedback control. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0038] Figure 1 A schematic diagram of the structure of the deep-sea autonomous grasping underwater robot of the present invention from one angle;

[0039] Figure 2 This is a schematic diagram of the structure of the deep-sea autonomous grasping underwater robot from another angle of the present invention;

[0040] Figure 3 It is a schematic diagram of the structure of the deep-sea autonomous grasping underwater robot of the present invention when the buoyancy block is extended;

[0041] Figure 4 It is a structural block diagram of the end compensation system of the robot arm in the present invention;

[0042] Figure 5 It is a flowchart of the end compensation algorithm of the robot arm in the present invention;

[0043] Figure 6 It is a flowchart of the single neuron adaptive PID algorithm in the present invention;

[0044] Figure 7 Schematic diagram of the path planning of the RRT* algorithm in the present invention.

[0045] The figure shows:

[0046] Underwater robot body 1 Forward-looking sonar 7

[0047] Hydraulic pump station 2 Underwater laser detection device 8

[0048] Starboard telescopic buoyancy block 3 Pressure differential measuring device 9

[0049] Mechanical arm 4 Port side telescopic buoyancy block 10

[0050] Underwater cable cutting device 5 Scissor type hydraulic telescopic rod 11

[0051] Binocular camera 6 Hydraulic cylinder 12 DETAILED DESCRIPTION

[0052] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0053] The present invention discloses a deep-sea autonomous grasping underwater robot and a method for grasping a target object. Acoustic and optical sensors are arranged on the underwater robot body based on an ROV or AUV platform to search for and identify the target. After optical recognition and locking of the target object's posture, the robot adjusts its posture and hovers. The mechanical arm grasps the target object when the robot body is disturbed, which can effectively improve the system reliability and operation accuracy.

[0054] According to the deep-sea autonomous grasping underwater robot provided by the present invention, Figure 1 , 2 As shown, it includes an underwater robot body 1, a hydraulic pump station 2, a controller, a buoyancy adjustment device, a mechanical arm 4, an optical sensor and a forward-looking sonar 7; the buoyancy adjustment device is installed on the side of the underwater robot body 1, the mechanical arm 4, the optical sensor and the forward-looking sonar 7 are all installed on the bow of the underwater robot body 1, and the controller and the hydraulic pump station 2 are both mounted on the underwater robot body 1; the hydraulic pump station 2 is used to drive the mechanical arm 4, and the hydraulic pump station 2, the optical sensor and the forward-looking sonar 7 are all connected to the controller signal. The underwater robot mainly uses ROV or AUV as the main platform. By carrying optical, acoustic and other sensors, underwater seven-function mechanical arms and related operating tools on the underwater robot, it can perform related disposal functions on suspended targets in the deep sea.

[0055] like Figure 3 As shown, the buoyancy adjustment device includes a hydraulic telescopic device, a port side telescopic buoyancy block 10 and a starboard side telescopic buoyancy block 3; the port side telescopic buoyancy block 10 and the starboard side telescopic buoyancy block 3 are respectively installed at the port side and starboard side center positions of the underwater robot body 1 through the hydraulic telescopic device, and the port side telescopic buoyancy block 10 and the starboard side telescopic buoyancy block 3 are symmetrically arranged.

[0056] The hydraulic telescopic device includes a scissor-type hydraulic telescopic rod 11 and a hydraulic cylinder 12, and the hydraulic cylinder 12 is installed inside the central frame of the underwater robot body 1; one end of the scissor-type hydraulic telescopic rod 11 is connected to the hydraulic cylinder 12, and the other end of the scissor-type hydraulic telescopic rod 11 is connected to the port telescopic buoyancy block 10 or the starboard telescopic buoyancy block 3; the underwater robot body 1 is provided with a buoyancy block accommodating cavity at the port and starboard center positions, respectively used to accommodate the port telescopic buoyancy block 10 and the starboard telescopic buoyancy block 3. In view of the problem of how to improve the stability of the underwater robot body 1 when it is hovering in the deep sea, in addition to improving the distance between the gravity and buoyancy centers of the underwater robot body 1 and improving the stability of the layout and control of the underwater robot propulsion system, retractable buoyancy blocks are also designed at the port and starboard center positions of the underwater robot body 1 to improve the stability of the robot body itself when the underwater robot grabs the target underwater.

[0057] The optical sensor includes a binocular camera 6 and / or an underwater laser detection device 8; the binocular camera 6 and the underwater laser detection device 8 are both installed at the head position of the bow of the underwater robot body 1. By configuring acoustic and optical sensors at the head position of the underwater robot body 1, deep-water suspended targets are searched. After the target is found and approached, the target is locked by an optical device, and the robot autonomously adjusts the position relationship with the target and then hovers, and the mechanical arm 4 autonomously grabs the target. The underwater robot also includes an underwater cable cutting device 5, which is installed in the working basket at the bow of the underwater robot body 1.

[0058] A pressure difference measuring device 9 is installed at the lower frame position of the bow of the underwater robot body 1; the pressure difference measuring device 9 is connected to the controller and the mechanical arm 4 respectively, and is used to detect the pressure difference between the inlet and outlet of all actuators of the mechanical arm 4, and send the pressure difference information to the controller. The mechanical arm 4 mounted on the underwater robot body 1 has a force feedback function, and the mechanical arm force feedback function is realized through the pressure difference measuring device 9 at the inlet and outlet of the mechanical arm actuator.

[0059] The robotic arm 4 mounted on the underwater robot body can realize the compensation function of the robotic arm's hand gripping end through relevant algorithms. When the underwater robot is hovering, the robot's own posture changes due to the influence of control accuracy and environmental factors. Under the disturbance of the robot body itself, the robotic arm can dynamically compensate for the interference of the robot body, thereby realizing the function of the robotic arm's hand gripping to accurately grasp the target object.

[0060] Example 1

[0061] This embodiment discloses a deep-sea autonomous grasping underwater robot, wherein the underwater robot body 1 provides an installation platform for all underwater equipment.

[0062] The starboard telescopic buoyancy block 3 and the port telescopic buoyancy block 10 are respectively installed at the center positions of the port and starboard of the underwater robot body 1. Figure 3 As shown, the port and starboard telescopic buoyancy blocks realize the extension and retraction functions through a scissor-type hydraulic telescopic mechanism, which is composed of a scissor-type hydraulic telescopic rod 11 and a hydraulic cylinder 12. After the underwater robot needs to hover and autonomously grab the suspended target, the starboard telescopic buoyancy block 3 and the port telescopic buoyancy block 10 are synchronously opened by the scissor-type hydraulic telescopic device, so that the port and starboard buoyancy blocks are fully deployed, increasing the flow area along the height direction of the underwater robot, providing the stability of the underwater robot body 1 hovering, and reducing the interference effect on the seven-function mechanical arm grabbing the target; the scissor-type hydraulic telescopic mechanism has the characteristics of compact structure, long telescopic distance, high stability, etc., and is suitable for installation on the underwater robot body 1, and occupies less space.

[0063] A hydraulic pump station 2 is installed inside the central frame of the underwater robot body 1. A seven-function robotic arm 4 is installed on the starboard side of the lower frame of the bow of the underwater robot body 1. An underwater cable cutting device 5 is placed in the bow working basket of the underwater robot body 1. The seven-function robotic arm 4 can also grab the underwater cable cutting device 5 in the sampling basket, and cut targets such as cables suspended in the deep sea when the underwater robot body 1 is in a hovering state. A binocular camera 6, a forward-looking sonar 7 and an underwater laser detection device 8 are installed at the head position of the bow of the underwater robot 1. A robotic arm pressure difference measuring device 9 is also installed at the lower frame position of the bow of the underwater robot body 1

[0064] The underwater robot body 1 uses the forward-looking sonar 7 to conduct a large-scale search for deep-water suspended targets. After the target is found, the underwater robot body 1 can approach the target autonomously by manual or sonar detection. After the underwater robot body 1 locks the target and approaches the target through the forward-looking sonar 7, it locks the position of the target in the deep sea through the binocular camera 6 or the underwater laser detection device 8, controls the robot body 1 to adjust the position relationship with the target and hover in the deep sea, and the binocular camera 6 or the underwater laser detection device 8 transmits the locked target position to the control system of the robot arm 4 in real time. The control system of the robot arm 4 real-time handles the target position of the robot arm 4, thereby realizing the function of the robot arm 4 handling the target.

[0065] Preferably, the underwater robot body 1 can also autonomously grasp and dispose of the target in a turbid seawater environment. The underwater laser detection device 8 can lock the target position in real time, and transmit the target position to the control system of the mechanical arm 4 in real time, so as to control the mechanical arm 4 to complete the function of autonomously grasping the target in a turbid seawater environment.

[0066] According to the target object grasping method of the deep-sea autonomous grasping underwater robot provided by the present invention, the deep-sea autonomous grasping underwater robot is used, and the method comprises the following steps:

[0067] Step 1: The underwater robot body 1 searches for underwater suspended targets through the forward-looking sonar 7. After the target is found, the underwater robot body 1 moves toward the target until it moves to the vicinity of the target; wherein the underwater robot body 1 approaches the target autonomously through manual control or through sonar detection. During the movement, the port side telescopic buoyancy block 10 and the starboard side telescopic buoyancy block 3 are both stored in the buoyancy block accommodating chamber of the underwater robot body 1;

[0068] Step 2: The position of the target object in the deep sea is locked by the binocular camera 6 or the underwater laser detection device 8, and the underwater robot body 1 is controlled to adjust its position relationship with the target object and then hover in the deep sea; after the underwater robot body 1 adjusts its position relationship with the target object, the port side telescopic buoyancy block 10 and the starboard side telescopic buoyancy block 3 are both symmetrically extended under the drive of the scissor-type hydraulic telescopic rod 11, so that the underwater robot body 1 hovers in the deep sea.

[0069] Step 3: The binocular camera 6 or the underwater laser detection device 8 transmits the target object posture information to the controller in real time, and the controller controls the robotic arm 4 to perform the target object disposal action according to the target object posture information.

[0070] Among them, step 3 adopts the robot end compensation algorithm, including the following sub-steps:

[0071] Step 3.1: The controller obtains the RGB image and the depth image from the optical sensor, calculates the position and posture of the target object relative to the optical sensor through the RGB image and the depth image, and then obtains the position and posture of the target object relative to the base of the robot arm 4 through coordinate transformation;

[0072] Step 3.2: Using the RRT* algorithm, discretized path points are planned according to the position of the target object relative to the base of the robot 4;

[0073] The path planning of the robot 4 is a key technology to achieve autonomous grasping, which guarantees the grasping efficiency and quality. Path planning can be divided into search-based path planning, sampling-based path planning and intelligent algorithm-based path planning. The robot in this application performs path planning based on sampling, and uses the RRT* algorithm to search for the optimal path, which is an optimization of the RRT algorithm.

[0074] The goal of the RRT* algorithm is to find a nearly optimal path from a starting point to a destination. It does this by iteratively expanding the search tree, optimizing the path cost in the process. In each iteration, the algorithm performs the following steps:

[0075] (1) Random sampling: randomly select a new point q in the search space rand , as candidate directions for expanding the tree.

[0076] (2) Node selection: Find an existing node q in the tree near , so that it is consistent with the new sampling point q rand The distance between them is minimal.

[0077] (3) Node expansion: from q near Along the q rand Create a new node q in the direction new , this new node is a one-step extension in the target direction.

[0078] (4) Cost calculation: Calculate the cost from the starting point to q new The path cost, that is, the cumulative cost.

[0079] (5) Reselect the parent node: new Find a set of nodes Q nearby near , and try to use some of these nodes as q new If this can reduce the number of nodes that reach q new cost.

[0080] (6) Rewiring: If through q new Can reduce q near The cost of other nodes in the network is to reconnect these nodes so that q new Become their parent node.

[0081] When the initial position and target position of the robot are known, the algorithm can plan a series of path points {x1, x2, ..., x n |x i =(p x ,p y ,p z ),i=1,2,...,n}.

[0082] Step 3.3: By using the inverse kinematics algorithm, according to the planned path points, the joint angles of the robot arm 4 are calculated, and the joints of the robot arm 4 are driven to make the end of the robot arm 4 approach the target object;

[0083] Take one of the path points x i Take the example of kinematics solution, knowing x i =(p x ,p y ,p z ) and given the roll angle α, pitch angle β and yaw angle γ of the path point, the corresponding rotation matrix can be obtained:

[0084]

[0085] The homogeneous transformation matrix of the corresponding posture can be obtained through the three-dimensional coordinates of the path point and the rotation matrix:

[0086]

[0087] Where: n x , n y , n z is the direction vector of the x-axis of the robot end link coordinate system in the base coordinate system; x , o y , o z is the direction vector of the y-axis of the robot end link coordinate system in the base coordinate system; a x , a y , a z is the direction vector of the z-axis of the robot end link coordinate system in the base coordinate system; p x , p y , p z is the coordinate of the robot end link coordinate system in the base coordinate system.

[0088] Based on the following inverse kinematics model and considering the constraints of the robot's own mechanical structure, the six joint angles corresponding to this posture can be solved:

[0089] θ1=Atan2(p y ,p x )

[0090]

[0091] Where: a i ,d i is the DH parameter of the robot arm, a i is the length of the common normal line between the joint i and i+1 axis, d i is the distance from the origin of the i-1th coordinate system to Z i-1 Axis and X i The intersection of the axes along the Z i-1 Distance measured by axis

[0092]

[0093] Due to the limitation of the working range of the fifth axis, s5≠0, the solution is

[0094]

[0095] Since θ3 has two possible values, θ2 has two possible solutions.

[0096]

[0097] At this point, the inverse solution is obtained for the path points in the joint space, that is, the six joint angles corresponding to the path points. Similarly, the joint angles corresponding to all path points planned by the RRT* algorithm can be solved.

[0098] Step 3.4: using the inertial measurement unit of the base of the robot arm 4 to estimate the offset of the base, and estimating the position of the target object relative to the base of the robot arm 4 after the base of the robot arm 4 is disturbed according to the offset;

[0099] When the robotic arm on the underwater vehicle ROV operates autonomously, the base is mostly in a floating state. Affected by the flow of seawater and the operation of the robotic arm, the base of the robotic arm deviates, resulting in the inability of the end gripper to accurately grasp the target. Therefore, a compensation strategy needs to be introduced to improve the grasping quality. In order to measure the offset of the base of the robotic arm, an inertial measurement unit (IMU) is installed on the base. The acceleration signal is measured by the IMU. After processing the signal, the position change of the object, that is, the error of the end gripper, can be solved. Then, the error compensation of the end position of the robotic arm is performed in combination with a single neuron PID controller to achieve accurate positioning of the end.

[0100] Step 3.5: Use the single neuron adaptive PID algorithm to drive each joint of the robot arm 4 and compensate the end of the robot arm 4;

[0101] After using the RRT* algorithm and inverse kinematics solution, the path points in the joint space are obtained. Two path points are taken in turn to perform "linear" interpolation on the joint angles. Then the interpolated joint angles are input and the single neuron PID algorithm is used to control the joint angles. Next, the control method is described using one of the joints of the robotic arm as an example.

[0102] (1) Calculation of joint errors

[0103]

[0104] where e (k) is the actual angle of the joint angle. is the desired angle of the joint angle.

[0105] This single neuron PID adopts position PID, which should have a proportional term Integral Item and the differential term

[0106]

[0107] (2) Supervised Hebb Learning and Normalization

[0108]

[0109] Where ηp , η i , η d They are the learning rates of the proportional term, integral term, and differential term of the neuron PID respectively. are the weights of the proportional term, integral term, and differential term of this iteration respectively. (k-1) This is the last output of the single neuron PID algorithm.

[0110] After obtaining the above updated connection weights After that, a normalization operation similar to unitization is required, that is:

[0111]

[0112] It is not difficult to understand that the proportional term x1, integral term x2 and differential term x3 are constantly updated, multiplied and then accumulated to update the weights w1, w2, w3, and finally a normalization process is performed.

[0113] (3) Output update

[0114]

[0115] The integration function has no nonlinear requirements and is directly proportional to the output, so the control law update is:

[0116] u (k) =u (k-1) +∈·net (k)

[0117] Where ∈ is the proportional coefficient of the neuron. (k-1) It is the output of the single neuron PID algorithm in this iteration.

[0118] In each control cycle after the “linear” interpolation of the joint angle, the single neuron PID algorithm continuously iteratively adjusts the weights and outputs the joint torque.

[0119] Step 3.6: Detect the error at the end of the robot arm 4. When the error exceeds the preset accuracy, jump to step 3.4.

[0120] Example 2

[0121] This embodiment selects the controller for implementing the above algorithm, and uses the EBOX-58 series industrial computer, which can work in harsh environments with low temperature and humid environment, and has excellent stability. Figure 5As shown in the figure, the controller obtains RGB and depth images from the depth camera, calculates the position of the target object relative to the camera through target detection and depth information, and then obtains the position of the target object relative to the base of the robotic arm through coordinate transformation. The inverse kinematics algorithm solves each joint angle, and the RRT* algorithm is used to plan the discrete path points, and the joint angles of the robotic arm are driven to make the end of the robotic arm approach the target object. Since the end of the robotic arm is affected by the seabed environment when it approaches the target object, the base will shift irregularly. At this time, the data measured by the IMU is read to estimate the true position of the target object, and then the single neuron adaptive PID algorithm is used to drive the joint angle to compensate the end to achieve target tracking. When the end of the robotic arm generates an error again and exceeds a certain accuracy, the controller uses the IMU data to compensate the end again.

[0122] In the description of the present application, it should be understood that the terms "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present application.

[0123] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A deep-sea autonomous grasping underwater robot, characterized in that: It comprises an underwater robot body (1), a hydraulic pump station (2), a controller, a buoyancy adjustment device, a mechanical arm (4), an optical sensor and a forward-looking sonar (7); The buoyancy adjustment device is installed on the side of the underwater robot body (1), the mechanical arm (4), the optical sensor and the forward-looking sonar (7) are all installed on the bow of the underwater robot body (1), and the controller and the hydraulic pump station (2) are both mounted on the underwater robot body (1); The hydraulic pump station (2) is used to drive the mechanical arm (4); the hydraulic pump station (2), the optical sensor and the forward-looking sonar (7) are all connected to the controller signal.

2. The deep-sea autonomous grasping underwater robot according to claim 1, characterized in that: The buoyancy adjustment device comprises a hydraulic telescopic device, a port telescopic buoyancy block (10) and a starboard telescopic buoyancy block (3); The port side telescopic buoyancy block (10) and the starboard side telescopic buoyancy block (3) are respectively installed at the port side and starboard side center positions of the underwater robot body (1) through a hydraulic telescopic device, and the port side telescopic buoyancy block (10) and the starboard side telescopic buoyancy block (3) are symmetrically arranged.

3. The deep-sea autonomous grasping underwater robot according to claim 2, characterized in that: The hydraulic telescopic device comprises a scissor-type hydraulic telescopic rod (11) and a hydraulic cylinder (12), wherein the hydraulic cylinder (12) is installed inside the central frame of the underwater robot body (1); One end of the scissor-type hydraulic telescopic rod (11) is connected to the hydraulic cylinder (12), and the other end of the scissor-type hydraulic telescopic rod (11) is connected to the port telescopic buoyancy block (10) or the starboard telescopic buoyancy block (3); The underwater robot body (1) is provided with buoyancy block accommodating chambers at the port side and starboard side center positions, respectively used to accommodate the port side telescopic buoyancy block (10) and the starboard side telescopic buoyancy block (3).

4. The deep-sea autonomous grasping underwater robot according to claim 1, characterized in that: The optical sensor comprises a binocular camera (6) and / or an underwater laser detection device (8); The binocular camera (6) and the underwater laser detection device (8) are both installed at the bow head position of the underwater robot body (1).

5. The deep-sea autonomous grasping underwater robot according to claim 1, characterized in that: It also comprises an underwater cable cutting device (5), which is installed in a working basket at the bow of the underwater robot body (1).

6. The deep-sea autonomous grasping underwater robot according to claim 1, characterized in that: A pressure difference measuring device (9) is installed at the lower frame position of the bow of the underwater robot body (1); The pressure difference measuring device (9) is connected to the controller and the mechanical arm (4) respectively, and is used to detect the pressure difference between the inlet and outlet of all actuators of the mechanical arm (4), and send the pressure difference information to the controller.

7. A method for grasping a target object by a deep-sea autonomous grasping underwater robot, characterized in that: The deep-sea autonomous grasping underwater robot according to any one of claims 1 to 6 is used, comprising the following steps: Step 1: The underwater robot body (1) searches for an underwater suspended target object through a forward-looking sonar (7). After the target object is found, the underwater robot body (1) moves toward the target object until it moves to the vicinity of the target object; Step 2: The position of the target object in the deep sea is locked by a binocular camera (6) or an underwater laser detection device (8), and the underwater robot body (1) is controlled to adjust the position relationship with the target object and then hover in the deep sea; Step 3: The binocular camera (6) or the underwater laser detection device (8) transmits the position information of the target object to the controller in real time, and the controller controls the robotic arm (4) to perform the disposal action of the target object according to the position information of the target object.

8. The method for grasping a target object by a deep-sea autonomous grasping underwater robot according to claim 7, characterized in that: In step 1, the underwater robot body (1) autonomously approaches the target object by manual control or by sonar detection.

9. The method for grasping a target object by a deep-sea autonomous grasping underwater robot according to claim 7, characterized in that: In step 1, the port side telescopic buoyancy block (10) and the starboard side telescopic buoyancy block (3) are both stored in the buoyancy block accommodating chamber of the underwater robot body (1); In step 2, after the underwater robot body (1) adjusts its positional relationship with the target object, the port side telescopic buoyancy block (10) and the starboard side telescopic buoyancy block (3) are both symmetrically extended under the drive of the scissor-type hydraulic telescopic rod (11), so that the underwater robot body (1) hovers in the deep sea.

10. The method for grasping a target object by a deep-sea autonomous grasping underwater robot according to claim 7, characterized in that: Step 3 uses the end-of-arm compensation algorithm, which includes the following sub-steps: Step 3.1: The controller obtains an RGB image and a depth image from the optical sensor, calculates the position of the target object relative to the optical sensor through the RGB image and the depth image, and then obtains the position of the target object relative to the base of the robot arm (4) through coordinate transformation; Step 3.2: Using the RRT* algorithm, discretized path points are planned according to the position of the target object relative to the base of the robot arm (4); Step 3.3: Using an inverse kinematics algorithm, the joint angles of the robot arm (4) are calculated according to the planned path points, and the joints of the robot arm (4) are driven so that the end of the robot arm (4) approaches the target object; Step 3.4: using the inertial measurement unit of the base of the robot arm (4) to estimate the offset of the base, and estimating the position of the target object relative to the base of the robot arm (4) after the base of the robot arm (4) is disturbed according to the offset; Step 3.5: Use a single neuron adaptive PID algorithm to drive each joint of the robotic arm (4) and compensate the end of the robotic arm (4); Step 3.6: Detect the error at the end of the robot arm (4). When the error exceeds the preset accuracy, jump to step 3.4.

Images