A welding method based on an amphibious humanoid welding robot

Abstract

The application provides a welding method based on an amphibious humanoid welding robot, comprising the following steps: mapping a steel structure surface, a target welding area and an initial position of the amphibious humanoid welding robot to a world three-dimensional coordinate system; then mapping to a world two-dimensional coordinate system; dividing the steel structure surface into grids; calculating the cost value of each grid to form a cost map; planning a minimum cost route from the initial position of the amphibious humanoid welding robot to the target welding area; constructing a three-order Bezier curve to obtain a smooth planning route; moving to the target welding area according to the planning route; planning trajectories of two mechanical arm modules and moving according to the planning trajectories to complete the welding work. The method is closer to the working habits of human beings, simulates the actions of human beings to perform welding operations, has better flexibility and operability, can autonomously plan a welding path, realizes autonomous navigation and accurate positioning, and has a high intelligent level.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of welding, in particular to a welding method based on an amphibious humanoid welding robot. BACKGROUND

[0002] Underwater welding robots are an important innovation in the field of engineering technology in recent years, and have a very broad application prospect in the fields of marine engineering, underwater facility maintenance, shipbuilding, etc. Traditional underwater welding work usually relies on divers, however, the underwater environment is complex and variable, and there are many unknown factors, manual welding not only has low efficiency, but also has high safety risk for divers. Underwater welding robots can autonomously complete welding tasks without human intervention, greatly improving work efficiency and effectively avoiding personnel casualties. At the same time, the robot can work stably underwater for a long time, without being limited by human fatigue, thereby further improving work efficiency.

[0003] Currently, the research on underwater welding robots has made some progress, researchers have optimized the mechanical structure, control system and welding process of the welding robot, so that it can adapt to the underwater environment and realize underwater welding operation function [1-4] However, the complexity and uncertainty of the underwater environment pose high requirements on the perception, control and decision-making ability of the robot; at the same time, the particularity of the underwater welding process also requires the robot to have higher precision and stability; the diversity of the welding scene (such as underwater structure repair in marine engineering and steel pipe pile body welding tasks) poses new challenges to the cross-environment adaptability of the robot.

[0004] Therefore, developing an amphibious welding robot with excellent environmental adaptability, flexible operability and intelligent level is of great significance to the development of underwater engineering field. Among them, optimizing the structure and design of the robot and the sensor system can improve the environmental adaptability and flexible operability of the amphibious welding robot, while computer vision and robot path planning technology are the key to improving the intelligent level of the amphibious welding robot.

[0005] However, the existing amphibious welding robot path planning method still has many deficiencies: first, due to the influence of complex and variable factors in the underwater environment, the existing technology is difficult to accurately model the underwater environment, which limits the robot path planning. Secondly, the existing path planning algorithm takes a long time to calculate when dealing with complex environments, which is difficult to meet the real-time requirements. Finally, the flexibility and adaptability of the existing path planning algorithm are still insufficient, and it is difficult to face different underwater environments and welding tasks. SUMMARY

[0006] In order to overcome the defects and deficiencies in the prior art, the purpose of the present application is to provide a welding method based on an amphibious humanoid welding robot; the method is closer to human work habits, simulates human actions for welding operations, has better flexibility and operability; can autonomously plan a welding path, realize autonomous navigation and precise positioning, and has a high level of intelligence.

[0007] In order to achieve the above purpose, the present application is realized by the following technical scheme: a welding method based on an amphibious humanoid welding robot, the amphibious humanoid welding robot comprising a head module, a neck module, a trunk module, a chassis module and two arm modules;

[0008] The welding method based on the amphibious humanoid welding robot comprises the following steps:

[0009] S1, mapping the surface of the steel structure, the target welding area and the initial position of the amphibious humanoid welding robot to a world three-dimensional coordinate system;

[0010] S2, mapping the surface of the steel structure, the target welding area and the initial position of the amphibious humanoid welding robot in the world three-dimensional coordinate system to a world two-dimensional coordinate system;

[0011] S3, dividing the surface of the steel structure in the world two-dimensional coordinate system into grids, the size of each grid being the same as the length and width of the chassis module of the amphibious humanoid welding robot;

[0012] S4, calculating the cost value of each grid according to the three-dimensional structure information of the surface of the steel structure, the distance of each grid from the target welding area and the kinematic model of the robot, and forming a cost map;

[0013] S5, planning a route from the initial position of the amphibious humanoid welding robot to the target welding area through grids with the minimum sum of cost values according to the cost of each grid, and setting the route as the minimum cost route; the grids on the minimum cost route are defined as minimum cost grids; the minimum cost route is the line connecting the midpoints of the minimum cost grids;

[0014] S6, based on the minimum cost route, using the midpoints of the four minimum cost grids in succession as control points to construct a third-order Bezier curve, and obtaining a smooth planning route in the coordinate system of the chassis module of the amphibious humanoid welding robot;

[0015] S7, controlling the amphibious humanoid welding robot to move to the target welding area according to the planning route obtained in step S6; identifying the to-be-welded area using the head module binocular camera, and determining the relative position of the to-be-welded area and the two arm modules;

[0016] S8. Use the RRT-Connect algorithm in three-dimensional space to plan the trajectory of two robotic arm modules, control the two robotic arm modules to move along the planned trajectory and complete the welding work.

[0017] Preferably, in step S4, the target welding area is defined as the planned endpoint area, and the initial position of the amphibious humanoid welding robot is defined as the planned starting point; for each grid obtained in step S3, the cost F of each grid is calculated. i,j This forms a cost map;

[0018] The cost F of each grid i,j The calculation is as follows:

[0019] F i,j =η g G i,j +η h H i,j +η z Z i,j +η ρ Ρ i,j +η κ K i,j

[0020] Among them, F i,j G represents the value of the i-th grid cell in the x-direction and the j-th grid cell in the y-direction, calculated starting from the origin of the world two-dimensional coordinate system; i,j H represents the cost of reaching the current grid cell from the planning starting point along the generated path; i,j Z represents the estimated cost from the current grid to the planned endpoint area; i,j P represents the weighted cost of the surface height of the general steel structure corresponding to the current grid. i,j K represents the surface curvature weighted cost of the current grid corresponding to the general steel structure surface; i,j η represents the weighted cost of the robot's kinematics model along the generated path from the previous grid cell to the current grid cell; g ,η h ,η z ,η ρ ,η κ These represent the weighting coefficients corresponding to each type of cost;

[0021] The cost G of reaching the current grid from the planning starting point along the generated path. i,j for:

[0022]

[0023] Where χ represents the set of all grids from the planning start point to the current grid in the generated path, with an initial cost of 0 at the planning start point;

[0024] The estimated cost H from the current grid to the planned endpoint areai,j is:

[0025]

[0026] wherein i0,j0 represents the center of the planning end region is the i0th in the x direction and the j0th in the y direction grid, H i,j represents the Euclidean distance from the current grid to the planning end region;

[0027] The surface height weighted cost Z of the current grid corresponding to the steel structure is: i,j is:

[0028] Z i，j =z i，j -z m，n

[0029] wherein z i，j represents the surface height of the current grid corresponding to the steel structure; z m，n represents the surface height of the previous grid corresponding to the steel structure; m, n represent the mth in the x direction and the nth in the y direction grid of the previous grid of the generated path;

[0030] The surface curvature weighted cost P of the current grid corresponding to the steel structure is: i，j is:

[0031]

[0032] wherein p i，j represents the curvature cost of the current grid; g i，j represents the torsion cost of the current grid; c represents a parameterized spatial curve formed by the current grid and all grids of the generated path in the world three-dimensional coordinate system, c', c'', c''' respectively represent the first, second, and third derivatives of the spatial curve c; det represents the determinant value of the matrix;

[0033] The kinematics model weighted cost K of the robot along the previous grid to the current grid of the generated path is: i，j is:

[0034]

[0035] wherein L is the wheelbase of the amphibious humanoid welding robot.

[0036] Preferably, the step S6 refers to: defining X={x1, x2, x3, x4, x5,...} represents the set of all grids of the minimum cost route, taking the center points of four consecutive grids as control points, and grouping:

[0037] X = {(x1, x2, x3, x4), (x4, x5, x6, x7), (x7, x8, x9, x 10 )...}

[0038] Respectively, each group of control points is constructed by a third order Bezier curve to form a smooth curve.

[0039] The third order Bezier curve construction method is:

[0040] y(t) = (1-t) 3 p0+3t(1-t) 2 p1+3t 2 (1-t)p2+t 3 p3

[0041] Wherein, y(t) is the generated third order Bezier curve; t ∈ (0, 1) is a parameter for controlling the curvature of the third order Bezier curve; p0, p1, p2, p3 are respectively four control points of the third order Bezier curve.

[0042] The smooth curves are connected in sequence, and the smooth planning route under the coordinate system of the amphibious humanoid welding robot chassis module.

[0043] Preferably, the step S1 refers to: in the land welding operation, the unmanned aerial vehicle is used to carry the laser radar and the camera module to realize the three-dimensional structure construction of the steel structure surface, the target welding area position identification and the initial position identification of the amphibious humanoid welding robot; the initial position of the unmanned aerial vehicle is taken as the world three-dimensional coordinate system.

[0044] In the underwater welding operation, the underwater floating robot is used to carry the underwater camera module to realize the three-dimensional structure construction of the steel structure surface, the target welding area position identification and the initial position identification of the amphibious humanoid welding robot; the initial position of the underwater floating robot is taken as the world three-dimensional coordinate system.

[0045] Preferably, the step S2 refers to: using the steel structure surface obtained in step S1, a three-dimensional map is established in the world three-dimensional coordinate system; the Z axis of the world three-dimensional coordinate system is compressed to form a world two-dimensional coordinate system; the three-dimensional map, the initial position of the amphibious humanoid welding robot and the target welding area are mapped to the world two-dimensional coordinate system.

[0046] Preferably, the step S7 refers to: converting the planning route under the coordinate system of the amphibious humanoid welding robot chassis module obtained in step S6 to the IMU coordinate system of the trunk module.

[0047] Taking the trunk module IMU coordinate system of the initial position of the amphibious humanoid welding robot as the origin, the amphibious humanoid welding robot is controlled to track the planned route and the real-time pose is obtained by integral of the trunk module IMU data, so that error correction in the motion route tracking process is realized.

[0048] After the amphibious humanoid welding robot moves to the target welding area, the neck module is driven to move the head module, and the real-time pose of the head is obtained by integral of the head module IMU data; during the movement of the head module, the to-be-welded area is identified by the head module binocular camera data.

[0049] The to-be-welded area in the head module binocular camera coordinate system is converted to the base coordinate system of the two mechanical arm modules, and the relative position of the to-be-welded area and the two mechanical arm modules is determined.

[0050] The real-time pose conversion relationship between the base coordinate system of the two mechanical arm modules and the end coordinate system of the two mechanical arm modules is obtained in real time by using the pose calculation method of the two mechanical arm modules, and the to-be-welded area is converted to the end coordinate system of the two mechanical arm modules.

[0051] Preferably, the step S8 refers to: according to the real-time pose conversion relationship between the to-be-welded area and the end coordinate system of the two mechanical arm modules obtained in step S7, the end trajectories of the two mechanical arm modules are planned respectively by using the RRT-Connect algorithm in three-dimensional space, and the ends of the two mechanical arm modules are controlled to move according to the planned trajectories; and the welding speed and angle are taken as constraint conditions to complete the welding work.

[0052] Preferably, the amphibious humanoid welding robot, the head module comprises a head shell and a binocular camera and a head module IMU fixed on the head shell; the neck module comprises a neck shell and left and right rotation motors and front and rear rotation motors fixed on the neck shell;

[0053] The trunk module comprises a trunk shell and a waterproof cover fixed inside the trunk shell; the waterproof cover is internally fixed with a trunk module IMU and a communication cable; the chassis module comprises an integrated chassis and a magnetic adsorption crawler device driven by a direct current motor;

[0054] The two mechanical arm modules comprise a left mechanical arm module and a right mechanical arm module; the left mechanical arm module comprises a left mechanical arm and a left welding gun fixed at the end of the left mechanical arm; the right mechanical arm module comprises a right mechanical arm and a right welding gun fixed at the end of the right mechanical arm.

[0055] Compared with the prior art, the present application has the following advantages and beneficial effects:

[0056] 1、The present application uses a humanoid working mode, which is closer to the working habits of humans, can simulate the actions of an operator with rich experience in land welding and underwater welding to perform welding operations, has better flexibility and operability, and can work stably in two completely different environments, underwater and on land, and has very high environmental adaptability.

[0057] 2、The present application proposes a new amphibious humanoid welding robot path planning method; the three-dimensional structure of the general steel structure surface is divided into a series of unit cells, and is mapped into a two-dimensional grid map, so that the environmental layout is clearly understood, and accurate path planning is performed; the three-dimensional structure (curvature, height, etc.) of the general steel structure surface, the distance between the grid and the target welding area, and the kinematic model of the robot are used as constraints, a heuristic search algorithm is used, the cost of each grid is evaluated, and the best position is selected for search, so that a large number of unnecessary search paths are avoided, and the calculation efficiency is improved; the two-dimensional chassis module motion path planning and the three-dimensional mechanical arm module trajectory planning are used to cooperatively complete the welding operation, the grid size and the search strategy can be flexibly adjusted according to the changes of the environment and the requirements of the task, so as to improve the environmental adaptability. Compared with existing underwater three-dimensional path planning methods, such as the Chinese patent "Underwater Robot Path Planning Method Based on Improved A* Algorithm for Marine Environment" (Publication No. CN117093012A), "Underwater Robot Path Planning Method Considering Ocean Current Influence" (Publication No. CN116734856A), etc., the present application has the advantages of low dimensionality, high efficiency and fast response. Compared with existing underwater two-dimensional path planning methods, such as the Chinese patent "Underwater Robot Path Planning Method Based on Hybrid Motion Sparrow Search Algorithm" (Publication No. CN117555353A), "Underwater Robot Path Planning Method and Equipment Based on Directed Particle Ant Colony" (Publication No. CN117387628A), "Multi-Wall Surface Operation Path Planning Method and System for Underwater Wall-Climbing Cleaning Robot" (Publication No. CN116974279A), etc., the present application integrates the robot motion model and three-dimensional structure information in the path search process, so that the optimal motion path of the robot from the starting point to the target welding area can be obtained.

[0058] 3、The present application uses a magnetic adsorption crawler chassis structure, which has the advantages of high positioning accuracy, good passability, simple motion model, etc., and has very high environmental adaptability. BRIEF DESCRIPTION OF DRAWINGS

[0059] Figure 1 is a flowchart of the welding method of the amphibious humanoid welding robot of the present application;

[0060] Figure 2 is a structural diagram of the amphibious humanoid welding robot of the present application;

[0061] Figure 3 is a schematic diagram of the coordinate system of the amphibious humanoid welding robot of the present application;

[0062] Figure 4 is a schematic diagram of the route planning process of the welding method based on the amphibious humanoid welding robot of the present application. DETAILED DESCRIPTION

[0063] The present application will be further described in detail below in combination with the accompanying drawings and specific embodiments.

[0064] EMBODIMENT

[0065] The present embodiment is a welding method based on an amphibious humanoid welding robot, as shown in the figure. The amphibious humanoid welding robot used, as shown in the figure, includes a head module 110, a neck module 120, a torso module 130, a chassis module 150, and two mechanical arm modules (a left mechanical arm module 140 and a right mechanical arm module 160). Figure 1 Figure 2 The head module 110 includes a head shell 111, a binocular camera (a right camera 112 and a left camera 113) fixed on the head shell 111, and a head module IMU 114. The neck module 120 includes a neck shell 121 and a left-right rotation motor 122 and a front-back rotation motor 123 fixed on the neck shell 121.

[0066] The torso module 130 includes a torso shell 131 and a waterproof cover 132 fixed inside the torso shell 131; the waterproof cover 132 has a torso module IMU 133 and a communication cable 134 fixed inside; the communication cable 134 has three kinds of communication lines integrated inside: a USB data line 135 for communication of the right camera 112, the left camera 113, the head module IMU 114, and the torso module IMU 133; a five-core aviation plug 136 for driving the left-right rotation motor 122, the front-back rotation motor 123, the left mechanical arm module 140, the chassis module 150, and the right mechanical arm module 160; and a nine-core aviation plug 137 for pose communication of the left mechanical arm module 140, the chassis module 150, and the right mechanical arm module 160.

[0067] The left mechanical arm module 140 includes a left mechanical arm 141 and a left welding torch 142 fixed at the end of the left mechanical arm 141. The chassis module 150 includes an integrated chassis 151 and a magnetic absorption crawler device 152 driven by a DC motor. The right mechanical arm module 160 includes a right mechanical arm 161 and a right welding torch 162 fixed at the end of the right mechanical arm 161. The welding torches at the ends of the left and right mechanical arms can also be replaced with an angle grinder, a cutting machine, a plasma cutting gun, etc. according to actual working condition requirements.

[0068] The left mechanical arm module 140 includes a left mechanical arm 141 and a left welding torch 142 fixed at the end of the left mechanical arm 141. The chassis module 150 includes an integrated chassis 151 and a magnetic absorption crawler device 152 driven by a DC motor. The right mechanical arm module 160 includes a right mechanical arm 161 and a right welding torch 162 fixed at the end of the right mechanical arm 161. The welding torches at the ends of the left and right mechanical arms can also be replaced with an angle grinder, a cutting machine, a plasma cutting gun, etc. according to actual working condition requirements.

[0069] ​The coordinate system definition of the amphibious humanoid welding robot device is shown in Figure 3 Fig. 1, including a world coordinate system, a right camera coordinate system, a left camera coordinate system, a head module IMU coordinate system, a trunk module IMU coordinate system, a left mechanical arm module base coordinate system, a left mechanical arm module end coordinate system, a chassis module coordinate system, a right mechanical arm module base coordinate system, and a right mechanical arm module end coordinate system.

[0070] The welding method based on the amphibious humanoid welding robot comprises the following steps:

[0071] S1. According to prior information, map the general steel structure surface, the target welding area, and the initial position of the amphibious humanoid welding robot to the world three-dimensional coordinate system.

[0072] The prior information refers to determining the three-dimensional topography of the general steel structure surface, the target welding area, and the initial position of the amphibious humanoid welding robot using existing technical means.

[0073] Step S1 refers to: in land welding operation, use a drone to carry a laser radar and a camera module to realize three-dimensional structure construction of the general steel structure surface, target welding area position identification, and initial position identification of the amphibious humanoid welding robot; take the initial position of the drone as the world three-dimensional coordinate system.

[0074] In underwater welding operation, use an underwater floating robot to carry an underwater camera module to realize three-dimensional structure construction of the general steel structure surface, target welding area position identification, and initial position identification of the amphibious humanoid welding robot; take the initial position of the underwater floating robot as the world three-dimensional coordinate system.

[0075] In the step S1, the initial position of the amphibious humanoid welding robot is identified to determine the pose conversion relationship between the world coordinate system and the chassis module coordinate system.

[0076] S2. Map the general steel structure surface, the target welding area, and the initial position of the amphibious humanoid welding robot in the world three-dimensional coordinate system to the world two-dimensional coordinate system.

[0077] As shown in Figure 4 Step S2 refers to: using the general steel structure surface obtained in step S1, establish a three-dimensional map in the world three-dimensional coordinate system; Figure 1 compress the Z-axis of the world three-dimensional coordinate system to form a world two-dimensional coordinate system; and map the three-dimensional map, the initial position of the amphibious humanoid welding robot, and the target welding area to the world two-dimensional coordinate system. Figure 1

[0078] S3. Divide the general steel structure surface in the world two-dimensional coordinate system into a grid to obtain a grid map. Figure 4 As shown in Figure 4 ​The size of each grid is the same as the length and width of the chassis module of the amphibious humanoid welding robot. The size of the grid helps to optimize the path of the search.

[0079] S4, according to the surface three-dimensional structure (curvature, height, etc.) information of the general steel structure, the distance of each grid to the target welding area and the robot kinematics model, the cost value of each grid is calculated to form a cost map 5, as shown in Figure 4 .

[0080] Step S4, the target welding area is defined as the planning end point area, and the initial position of the amphibious humanoid welding robot is defined as the planning start point. For each grid obtained in step S3, the cost value F i,j of each grid is calculated to form a cost map 5.

[0081] The calculation of the cost value F i,j of each grid is as follows:

[0082] F i,j = η g G i,j + η h H i,j + η z Z i,j + η ρ P i,j + η κ K i,j

[0083] Wherein, F i,j represents the cost value of the i-th grid in the x direction and the j-th grid in the y direction starting from the origin of the world two-dimensional coordinate system; G i,j represents the cost of the planning start point reaching the current grid along the generated path; H i,j represents the estimated cost of the current grid to the planning end point area; Z i,j represents the corresponding general steel structure surface height weighted cost of the current grid; P i,j represents the corresponding general steel structure surface curvature weighted cost of the current grid; K i,j represents the robot kinematics model weighted cost from the previous grid to the current grid along the generated path; η g , η h , η z , η ρ , η κ respectively represent the weighted coefficients corresponding to each cost.

[0084] The cost G i,j of the planning start point reaching the current grid along the generated path is as follows:

[0085]

[0086] wherein χ represents a set of all the grids in the generated path from the planning start point to the current grid, and the initial cost at the planning start point is 0;

[0087] The estimated cost H of the current grid to the planning end point region i,j is:

[0088]

[0089] wherein i0,j0 represents the center of the planning end point region as the ith0grid in the x direction and the jth0grid in the y direction, H i,j represents the Euclidean distance from the current grid to the planning end point region;

[0090] The surface height weighted cost Z of the current grid corresponding to the generic steel structure i,j is:

[0091] Z i，j = z i，j - z m，n

[0092] wherein z i，j represents the surface height of the current grid corresponding to the generic steel structure; z m，n represents the surface height of the previous grid in the generated path corresponding to the generic steel structure; m and n represent the previous grid in the generated path as the mthgrid in the x direction and the nthgrid in the y direction;

[0093] The surface curvature weighted cost P of the current grid corresponding to the generic steel structure i，j is:

[0094]

[0095] wherein p i，j represents the curvature cost of the current grid; g i，j represents the torsion cost of the current grid; c represents a parameterized spatial curve formed by the current grid and all the grids in the generated path in the world three-dimensional coordinate system, c', c", and c'" represent the first, second, and third derivatives of the spatial curve c respectively; and det represents the determinant value of a matrix;

[0096] The kinematics model weighted cost K of the robot along the previous grid in the generated path to the current grid i，j is:

[0097]

[0098] wherein L is the track of the amphibious humanoid welding robot, specifically the distance between the front wheel and the rear wheel on the same side track.

[0099] S5, according to the cost of each grid, a route from the amphibious humanoid welding robot initial position to the target welding area is planned, the sum of the cost of the grids passed through is the smallest, and is set as the minimum cost route 6, as shown in Figure 4 ; the grid on the minimum cost route 6 is defined as the minimum cost grid; the minimum cost route 6 is the line connecting the centers of the minimum cost grids.

[0100] S6, based on the minimum cost route, the centers of the four consecutive minimum cost grids are used as control points to construct a cubic Bezier curve to obtain a smooth planning route in the amphibious humanoid welding robot chassis module coordinate system.

[0101] Since the minimum cost route planned in step S5 is a line connecting the centers of all minimum cost grids, it has the characteristics of minimum cost but not smoothness, and cannot be directly used for robot motion control, so the centers of the minimum cost grids are used as control points to construct a cubic Bezier curve to generate a smooth planning route in the amphibious humanoid welding robot chassis module coordinate system.

[0102] A cubic Bezier curve requires four control points, and X = {x1, x2, x3, x4, x5,...} represents the set of all grids of the minimum cost route, and the centers of the four consecutive grids are grouped as control points:

[0103] X = {(x1, x2, x3, x4), (x4, x5, x6, x7), (x7, x8, x9, x 10 )...}

[0104] A cubic Bezier curve is constructed for each group of control points to form a smooth curve.

[0105] The cubic Bezier curve construction method is:

[0106] y(t) = (1-t) 3 p0+3t(1-t) 2 p1+3t 2 (1-t)p2+t 3 p3

[0107] Where y(t) is the generated cubic Bezier curve; t ∈ (0, 1) is a parameter that controls the curvature of the cubic Bezier curve; p0, p1, p2, p3 are the four control points of the cubic Bezier curve, respectively.

[0108] The smooth curves are connected in sequence to obtain the smooth planning route in the amphibious humanoid welding robot chassis module coordinate system.

[0109] S7, controlling the amphibious humanoid welding robot to move to the target welding area according to the planned route obtained in step S6; and identifying the area to be welded by using the binocular camera of the head module to determine the relative position of the area to be welded and the two arm modules.

[0110] Step S7 refers to converting the planned route of the amphibious humanoid welding robot chassis module in the coordinate system of the chassis module to the coordinate system of the IMU of the trunk module; the pose conversion relationship between the coordinate system of the chassis module and the coordinate system of the IMU of the trunk module can be obtained by calibrating the motion trajectory of the chassis module and the motion trajectory of the IMU of the trunk module;

[0111] Taking the coordinate system of the IMU of the trunk module at the initial position of the amphibious humanoid welding robot as the origin, the amphibious humanoid welding robot is controlled to track the planned route, and the real-time pose is obtained by integrating the IMU data of the trunk module to realize error correction in the motion route tracking process.

[0112] After the amphibious humanoid welding robot moves to the target welding area, the head module is moved by driving the left and right rotation motors and the front and rear rotation motors of the neck module, and the real-time pose of the head is obtained by integrating the IMU data of the head module; in the motion process of the head module, the area to be welded is identified by the binocular camera data of the head module; the pose conversion relationship between the coordinate system of the IMU of the head module and the coordinate system of the binocular camera can be calibrated by using existing tools such as Matlab and Kalibr; the area to be welded refers to the area where the two arm modules need to perform welding work.

[0113] The identified area to be welded in the coordinate system of the binocular camera of the head module is converted to the coordinate system of the two arm modules to determine the relative position of the area to be welded and the two arm modules.

[0114] The coordinate system of the binocular camera of the head module and the coordinate system of the two arm modules are obtained by hand-eye calibration and pose conversion of the coordinate system of the IMU of the head module: when the head module is stationary, the initial pose conversion relationship between the coordinate system of the binocular camera of the head module and the coordinate system of the two arm modules is calibrated by hand-eye calibration; when the amphibious humanoid welding robot is moving, the real-time pose of the head module is obtained by integrating the IMU data of the head module, based on the pose conversion relationship between the coordinate system of the binocular camera of the head module and the coordinate system of the IMU of the head module, and the initial pose conversion relationship between the coordinate system of the binocular camera of the head module and the coordinate system of the two arm modules; the real-time pose relationship between the coordinate system of the binocular camera of the head module and the coordinate system of the two arm modules is obtained by combining the three.

[0115] The real-time pose conversion relationship between the coordinate system of the two arm modules and the coordinate system of the two arm modules is obtained by using existing methods for calculating the pose of the two arm modules, and the area to be welded is converted to the coordinate system of the two arm modules.

[0116] S8, planning trajectories of the two robot modules in the three-dimensional space by using RRT-Connect algorithm, controlling the two robot modules to move according to the planned trajectories and completing the welding work.

[0117] Step S8 refers to: according to the real-time pose conversion relationship between the welding area to be welded and the end coordinate system of the two robot modules obtained in step S7, planning the end motion trajectories of the two robot modules by using RRT-Connect algorithm in the three-dimensional space, controlling the end of the two robot modules to move according to the planned motion trajectories; and taking the welding speed and angle as constraint conditions, completing the welding work.

[0118] Specifically, the welding area to be welded is divided into two parts (defined as left and right welding areas), and the welding work is performed by two robot modules (defined as left and right robot modules), and the starting points and ending points of the two parts of the welding area to be welded are defined (defined as left and right starting points and left and right ending points); the RRT-Connect algorithm is used to plan the motion trajectories of the left robot module end to the left starting point and the right robot module end to the right starting point in the three-dimensional space; the left and right robot module ends are controlled to move along the planned motion trajectories; after the left and right robot module ends move to the left and right starting points, the welding speed and welding angle constraints are used to plan the welding paths of the left and right robot module ends from the left and right starting points to the left and right ending points along the left and right welding areas to be welded; the welding guarantee equipment such as welding power supply, wire feeding system and gas cylinder is started, the left and right robot module ends are controlled to move along the welding paths, and the welding work is completed.

[0119] The above embodiments are the preferred embodiments of the present application, but the embodiments of the present application are not limited by the above embodiments, and any changes, modifications, substitutions, combinations, simplifications made without departing from the spirit and principles of the present application should be equivalent replacement methods, and all are included in the protection scope of the present application.

Claims

1. A welding method based on an amphibious humanoid welding robot, characterized in that: The amphibious humanoid welding robot includes a head module, a neck module, a torso module, a chassis module, and two robotic arm modules. The welding method based on an amphibious humanoid welding robot includes the following steps: S1. Map the surface of the steel structure, the target welding area, and the initial position of the amphibious humanoid welding robot to the world three-dimensional coordinate system; S2. Map the surface of the steel structure, the target welding area, and the initial position of the amphibious humanoid welding robot in the three-dimensional world coordinate system to the two-dimensional world coordinate system. S3. Divide the surface of the steel structure in the world two-dimensional coordinate system into grids, and the size of each grid is the same as the length and width of the chassis module of the amphibious humanoid welding robot. S4. Based on the three-dimensional structural information of the steel structure surface, the distance between each grid and the target welding area, and the robot kinematics model, calculate the cost value of each grid to form a cost map. S5. Based on the cost of each grid, plan a route from the initial position of the amphibious humanoid welding robot to the target welding area that minimizes the sum of the grid costs, and set it as the minimum cost route; the grids on the minimum cost route are defined as minimum cost grids; the minimum cost route is the line connecting the midpoints of the minimum cost grids. S6. Based on the minimum cost route, using the midpoints of four consecutive minimum cost grids as control points, a third-order Bézier curve is constructed to obtain a smooth planning route in the coordinate system of the chassis module of the amphibious humanoid welding robot. S7. Control the amphibious humanoid welding robot to move to the target welding area according to the planned route obtained in step S6; The head module uses a binocular camera to identify the area to be soldered and determine the relative position of the area to be soldered to the two robotic arm modules. S8. Use the RRT-Connect algorithm in three-dimensional space to plan the trajectory of two robotic arm modules, control the two robotic arm modules to move along the planned trajectory and complete the welding work; In step S4, the target welding area is defined as the planned endpoint area, and the initial position of the amphibious humanoid welding robot is defined as the planned starting point; for each grid obtained in step S3, the cost F of each grid is calculated. i,j This forms a cost map; The cost F of each grid i,j The calculation is as follows: F i,j =the g G i,j +n h H i,j +n z Z i,j +n ρ R i,j +n κ K i,j Among them, F i,j G represents the value of the i-th grid cell in the x-direction and the j-th grid cell in the y-direction, calculated starting from the origin of the world two-dimensional coordinate system; i,j H represents the cost of reaching the current grid cell from the planning starting point along the generated path; i,j Z represents the estimated cost from the current grid to the planned endpoint area; i,j P represents the weighted cost of the surface height of the general steel structure corresponding to the current grid. i,j K represents the surface curvature weighted cost of the current grid corresponding to the general steel structure surface; i,j η represents the weighted cost of the robot's kinematics model along the generated path from the previous grid cell to the current grid cell; g ,η h ,η z ,η ρ ,η κ These represent the weighting coefficients corresponding to each type of cost; The cost G of reaching the current grid from the planning starting point along the generated path. ij for: Where χ represents the set of all grids from the planning start point to the current grid in the generated path, with an initial cost of 0 at the planning start point; The estimated cost H from the current grid to the planned endpoint area i,j for: Where i0,j0 represents the planning endpoint area center being the i0th grid in the x-direction and the j0th grid in the y-direction, H i,j This represents the Euclidean distance from the current grid cell to the planned endpoint area; The weighted cost Z of the surface height of the steel structure corresponding to the current grid i,j for: WITH i,j =z i,j -With m,n Among them, z i,j This represents the height of the steel structure surface corresponding to the current grid; z m,n This represents the height of the steel structure surface corresponding to the previous grid cell of the generated path; m, n represent that the previous grid cell of the generated path is the m-th grid cell in the x-direction and the n-th grid cell in the y-direction. The surface curvature weighted cost P of the steel structure corresponding to the current grid i,j for: Where, ρ i,j Represents the curvature cost corresponding to the current grid; γ i,j represents the torsion cost corresponding to the current grid; c represents a parameterized space curve formed by the current grid and all grids of the generated path in the world three-dimensional coordinate system, c′, c″, c″′ represent the first, second, and third derivatives of the space curve c, respectively; det represents the determinant value of the matrix. The weighted cost K of the robot kinematics model along the generated path from the previous grid to the current grid ij for: Where L is the wheelbase of the amphibious humanoid welding robot.

2. The welding method based on an amphibious humanoid welding robot according to claim 1, characterized in that: Step S6 refers to: defining X = {x1, x2, x3, x4, x5, ...} as the set of all grids for the minimum cost route, and grouping four consecutive grid center points as control points: X={(x1,x2,x3,x4),(x4,x5,x6,x7),(x7,x8,x9,x 10 )...} For each group of control points, construct a third-order Bézier curve to form a smooth curve; The method for constructing the third-order Bézier curve is as follows: y(t)=(1-t) 3 p0+3t(1-t) 2 p1+3t 2 (1-t)p2+t 3 p3 Where y(t) is the generated third-order Bézier curve; t∈(0,1) is the parameter controlling the curvature of the third-order Bézier curve; p0, p1, p2, p3 are the four control points of the third-order Bézier curve. Connect the smooth curves in sequence to plan the smooth route in the coordinate system of the chassis module of the amphibious humanoid welding robot.

3. The welding method based on an amphibious humanoid welding robot according to claim 1, characterized in that: Step S1 refers to: in land welding operations, using a drone equipped with a lidar and camera module to realize the three-dimensional structure construction of the surface of the steel structure, the location identification of the target welding area, and the initial position identification of the amphibious humanoid welding robot; using the initial position of the drone as the world three-dimensional coordinate system; In underwater welding operations, an underwater floating robot equipped with an underwater camera module is used to realize the three-dimensional structure construction of the surface of the steel structure, the location identification of the target welding area, and the initial position identification of the amphibious humanoid welding robot; the initial position of the underwater floating robot is used as the world three-dimensional coordinate system.

4. The welding method based on an amphibious humanoid welding robot according to claim 1, characterized in that: Step S2 refers to: using the surface of the steel structure obtained in step S1, establishing a three-dimensional map in the world three-dimensional coordinate system; compressing the Z-axis of the world three-dimensional coordinate system to form a world two-dimensional coordinate system; and mapping the three-dimensional map, the initial position of the amphibious humanoid welding robot, and the target welding area to the world two-dimensional coordinate system.

5. The welding method based on an amphibious humanoid welding robot according to claim 1, characterized in that: Step S7 refers to: transforming the planned route in the coordinate system of the chassis module of the amphibious humanoid welding robot obtained in step S6 to the coordinate system of the torso module IMU. Using the torso module IMU coordinate system as the origin, the amphibious humanoid welding robot is controlled to track the movement along the planned route. The torso module IMU data is integrated to obtain the real-time pose and realize error correction during the movement tracking process. After the amphibious humanoid welding robot moves to the target welding area, the neck module is driven to move the head module, and the real-time head pose is obtained by integrating the IMU data of the head module; during the movement of the head module, the area to be welded is identified by the binocular camera data of the head module. Transform the area to be welded in the binocular camera coordinate system of the head module to the base coordinate system of the two robotic arm modules to determine the relative position of the area to be welded to the two robotic arm modules. Using a pose calculation method for two robotic arm modules, the real-time pose transformation relationship between the base coordinate system and the end coordinate system of the two robotic arm modules is obtained, and the area to be welded is transformed to the end coordinate system of the two robotic arm modules.

6. The welding method based on an amphibious humanoid welding robot according to claim 5, characterized in that: Step S8 refers to: based on the real-time pose transformation relationship between the area to be welded and the coordinate systems of the two robotic arm modules obtained in step S7, the RRT-Connect algorithm in three-dimensional space is used to plan the motion trajectory of the two robotic arm modules, and the two robotic arm modules are controlled to move according to the planned motion trajectory; and the welding work is completed by using welding speed and angle as constraints.

7. The welding method based on an amphibious humanoid welding robot according to claim 1, characterized in that: The amphibious humanoid welding robot includes a head module comprising a head shell, a binocular camera fixed to the head shell, and a head module IMU; and a neck module comprising a neck shell, a left-right rotation motor, and a front-back rotation motor fixed to the neck shell. The torso module includes a torso shell and a waterproof cover fixed inside the torso shell; the torso module IMU and communication cables are fixed inside the waterproof cover; the chassis module includes an integrated chassis and a DC motor driven magnetic adsorption track device. The two robotic arm modules include a left robotic arm module and a right robotic arm module; the left robotic arm module includes a left robotic arm and a left welding torch fixed to the end of the left robotic arm; the right robotic arm module includes a right robotic arm and a right welding torch fixed to the end of the right robotic arm.

Images