Automotive Rear Longitudinal Beam Assembly Welding System Based on 3D Vision Guidance and Multi-Figment Collaboration

Abstract

This invention relates to the field of automated welding technology, and specifically discloses a welding system for automotive rear longitudinal beam assemblies based on 3D vision guidance and multi-fixture collaboration. The system includes: an incoming material transport module that uses an automated guided vehicle to transport the automotive rear longitudinal beam material from the loading area to the welding production line's bench position; a gripping trajectory correction module that automatically corrects the robot's gripping trajectory based on the spatial position and attitude data of the automotive rear longitudinal beam material, obtaining a high-precision corrected gripping trajectory; an incoming material gripping module that controls the robot to grip the automotive rear longitudinal beam material based on the high-precision corrected gripping trajectory and place it on a spot welding fixture; a spot welding operation movement module that uses a servo slide to move the spot welding fixture and the automotive rear longitudinal beam material to the welding position and complete the spot welding operation; and an arc welding operation movement module that, after spot welding is completed, controls the robot to grip the spot-welded part and place it on an arc welding fixture located on a positioner to complete the arc welding operation. This achieves full-process automation from incoming material transport to welding.

Description

Technical Field

[0001] This invention relates to the field of automated welding technology, and in particular to a welding system for automotive rear longitudinal beam assemblies based on 3D vision guidance and multi-fixture collaboration. Background Technology

[0002] In the automotive manufacturing industry, welding is a crucial step in car body production, directly impacting the overall quality and safety of the vehicle. As a vital component of the body structure, the welding quality of the rear longitudinal beam assembly plays a critical role in energy absorption and occupant protection during a collision. With the development of the automotive manufacturing industry, higher demands are being placed on the automation, precision, and efficiency of rear longitudinal beam assembly welding. Traditional welding systems largely rely on manual experience and basic positioning techniques when welding automotive rear longitudinal beams.

[0003] In terms of incoming material transportation, the level of automation is low, and the transportation route planning is not flexible enough to adapt to the needs of different production rhythms. In the gripping stage, incoming material handling and clamping require manual labor, resulting in low production efficiency and high labor intensity. Manual placement of parts may lead to positional deviations, affecting the consistency of welding quality and subsequent welding accuracy. During spot welding and arc welding operations, fixture switching and adjustment require manual operation, impacting the production rhythm. The coordination between fixtures and welding equipment is poor, and the connection between various processes is not smooth, reducing production efficiency. Furthermore, the lack of real-time monitoring and intelligent adjustment of the welding process makes it difficult to guarantee the stability of welding quality. Although the industry has been exploring improvements, existing welding systems still face many challenges in coping with complex and ever-changing production tasks and improving welding accuracy and efficiency, urgently requiring the introduction of new technologies and methods for optimization and upgrading.

[0004] Therefore, this invention proposes a welding system for automotive rear longitudinal beam assemblies based on 3D vision guidance and multi-fixture collaboration. Summary of the Invention

[0005] This invention provides a welding system for automotive rear longitudinal beam assemblies based on 3D vision guidance and multi-fixture collaboration. The system includes: a material transport module utilizing an automated guided vehicle (AGV) to automate the transport of automotive rear longitudinal beam materials from a specific loading area to the welding production line's workstation, efficiently supplying materials, reducing manpower, and avoiding transport deviations; a gripping trajectory correction module automatically correcting the robot's gripping trajectory based on the incoming material's spatial position and posture data, improving gripping accuracy and ensuring accurate welding positions; an incoming material gripping module placing the material onto the spot welding fixture according to the corrected trajectory, providing a precise starting position for spot welding operations and ensuring welding quality; a spot welding operation movement module using a servo slide to precisely control the spot welding position, achieving an automated spot welding process and improving the stability and efficiency of spot welding quality; and an arc welding operation movement module precisely transferring the parts to the arc welding fixture after spot welding to complete arc welding, achieving seamless connection between spot welding and arc welding, ensuring welding continuity, improving overall welding quality and production efficiency, reducing the uncertainty of manual intervention, and promoting the automation and intelligence of the welding process.

[0006] This invention provides a welding system for automotive rear longitudinal beam assemblies based on 3D vision guidance and multi-fixture collaboration, comprising:

[0007] The incoming material transportation module is used to transport the automotive rear longitudinal beam from the loading area to the welding production line's bench garage position using an automated guided vehicle.

[0008] The grasping trajectory correction module is used to automatically correct the robot's grasping trajectory based on the spatial position and attitude data of the incoming automotive rear longitudinal beam, thereby obtaining a high-precision corrected grasping trajectory for the robot.

[0009] The incoming material gripping module is used to control the robot based on high-precision correction gripping trajectory to grip the incoming automotive rear longitudinal beam and place it on the spot welding fixture;

[0010] The spot welding operation moving module is used to move the spot welding fixture and the automotive rear longitudinal beam to the welding position based on the servo slide and complete the spot welding operation.

[0011] The arc welding operation moving module is used to control the robot to grab the spot-welded parts after spot welding and place them on the arc welding fixture located on the positioner to complete the arc welding operation.

[0012] Preferably, the incoming material transport module includes:

[0013] The initial path generation submodule is used to generate the initial transport path between the loading area and the welding production line's kiosks;

[0014] The obstacle avoidance mechanism triggering submodule is used to control the automated guided vehicle to pick up the automotive rear longitudinal beam material from the loading area and transport it to the welding production line's bench station according to the initial transport path. At the same time, based on the millimeter-wave radar on the automated guided vehicle, obstacles within a preset range centered on the automated guided vehicle are detected in real time. When the millimeter-wave radar detects an obstacle, the obstacle avoidance strategy is triggered to generate a new obstacle avoidance transport path, and the automotive rear longitudinal beam material is continued to be transported to the welding production line's bench station according to the new obstacle avoidance transport path.

[0015] The obstacle avoidance strategies include:

[0016] Multiple initial obstacle avoidance transport paths are generated based on the location of obstacles detected by millimeter-wave radar;

[0017] The path cost of each initial obstacle avoidance transportation path is calculated based on the path cost function:

[0018]

[0019] In the formula, The path cost of a single initial obstacle avoidance transport path. This represents the total number of locations in a single initial obstacle avoidance transport path. For the first obstacle avoidance transport path in a single initial obstacle avoidance transport path A location point, For the first obstacle avoidance transport path in a single initial obstacle avoidance transport path A location point, For the first obstacle avoidance transport path in a single initial obstacle avoidance transport path The location point and the first Euclidean distance between points This represents the total number of obstacles detected by the millimeter-wave radar. For the first obstacle avoidance transport path in a single initial obstacle avoidance transport path The x-coordinate of each location point For the first obstacle avoidance transport path in a single initial obstacle avoidance transport path The y-coordinate values ​​of each location point For the first The x-coordinate value of each obstacle For the first The vertical coordinate value of each obstacle;

[0020] The initial obstacle avoidance transport path with the minimum path cost among all initial obstacle avoidance transport paths is taken as the latest obstacle avoidance transport path.

[0021] Preferably, the grasping trajectory correction module includes:

[0022] The incoming material pose scanning submodule is used to scan and identify the spatial position and pose data of the incoming automotive rear longitudinal beam using a depth camera.

[0023] The spatial pose matching submodule is used to match the spatial position and attitude data of the incoming automotive rear longitudinal beam with the preset 3D model of the incoming material based on the point cloud registration algorithm, and generate a spatial pose deviation matrix.

[0024] The grasping trajectory correction submodule is used to automatically correct the robot's grasping trajectory based on the spatial pose deviation matrix, thereby obtaining a high-precision corrected grasping trajectory for the robot.

[0025] Preferably, the capture trajectory correction submodule includes:

[0026] The initial correction unit is used to automatically correct the robot's grasping trajectory based on the spatial pose deviation matrix, and obtain the robot's initial corrected grasping trajectory.

[0027] The dynamic model generation unit is used to generate the robot's state vector at each time step based on the robot's dynamic vibration offset and the rate of change of the dynamic vibration offset at each time step, and to generate the robot's dynamic model by combining the state transition matrix, the control input matrix, and the process noise vector.

[0028] The measurement model generation unit is used to generate a measurement model of the robot based on the robot's dynamic model, measurement matrix, and measurement noise vector;

[0029] The vibration offset estimation unit is used to estimate the dynamic vibration offset of the robot based on the robot's dynamic model, measurement model, and Kalman filter.

[0030] The reverse compensation unit is used to perform reverse displacement compensation on the robot's initial corrected grasping trajectory based on the robot's dynamic vibration offset, so as to obtain the robot's high-precision corrected grasping trajectory.

[0031] Preferably, the initial correction unit includes:

[0032] The position correction subunit is used to correct the position of each position point in the grasping trajectory of the robot based on the translation vector in the spatial pose deviation matrix, so as to obtain the position correction grasping trajectory of the robot.

[0033] The attitude correction subunit is used to adjust the attitude of each position point in the robot's position correction grasping trajectory based on the rotation matrix in the spatial pose deviation matrix, so as to obtain the robot's initial correction grasping trajectory.

[0034] Preferred options also include:

[0035] The fixture contact pressure detection module is used to detect the real-time contact pressure distribution data between the incoming automotive rear longitudinal beam and the fixture based on the pressure sensor array inside the spot welding fixture.

[0036] The fixture support posture adjustment module is used to adjust the support posture of the fixture based on the real-time contact pressure distribution data between the automotive rear longitudinal beam material and the fixture.

[0037] Preferably, the clamp support posture adjustment module includes:

[0038] The sliding trend prediction submodule is used to analyze the sliding trend of the automotive rear longitudinal beam material in the fixture based on the real-time contact pressure distribution data between the automotive rear longitudinal beam material and the fixture.

[0039] The support posture adjustment submodule is used to adjust the support posture of the fixture based on the sliding trend of the automotive rear longitudinal beam material in the fixture.

[0040] Preferably, the sliding trend prediction submodule includes:

[0041] The normal pressure calculation unit is used to calculate the total normal pressure in the contact area between the automotive rear longitudinal beam material and the fixture based on the real-time contact pressure distribution data between the automotive rear longitudinal beam material and the fixture.

[0042] The resultant force analysis unit is used to calculate the frictional force on the rear longitudinal beam material in the fixture based on the total normal pressure in the contact area between the rear longitudinal beam material and the fixture, and to determine the resultant force on the rear longitudinal beam material in the fixture in combination with other external forces on the rear longitudinal beam material.

[0043] The maximum static friction analysis unit is used to calculate the maximum static friction between the automotive rear longitudinal beam material and the fixture based on the total normal pressure of the contact area between the material and the fixture.

[0044] The sliding trend analysis unit is used to analyze the sliding trend of the automotive rear longitudinal beam material in the fixture based on the tangential component of the resultant force and the maximum static friction force acting on the material.

[0045] Preferably, the support pose adjustment submodule includes:

[0046] An optimization objective function construction unit is used to construct an objective function with the goal of minimizing the sliding trend when the sliding trend of the automotive rear longitudinal beam material in the fixture exceeds the sliding trend threshold.

[0047] The pose optimization unit is used to solve for the optimal support pose vector of the fixture based on the objective function and the upper and lower limits of the support pose vector of the fixture, and to control the support pose of the fixture based on the optimal support pose vector of the fixture.

[0048] Preferably, the arc welding fixture includes:

[0049] The clamping force is adjustable within a preset range via a pneumatic clamping mechanism and thermocouple array controlled by a proportional valve.

[0050] Among them, the thermocouple array monitors the real-time temperature of the welding area on the part after spot welding, and dynamically adjusts the welding current based on the real-time temperature of the welding area on the part after spot welding.

[0051] The beneficial effects of this invention compared to existing technologies are as follows: The incoming material transportation module utilizes an automated guided vehicle (AGV) to automate the transport of automotive rear longitudinal beams from a specific loading area to the welding production line's workstation, efficiently supplying materials, reducing manpower, and avoiding transport deviations. The gripping trajectory correction module automatically corrects the robot's gripping trajectory based on the incoming material's spatial position and posture data, improving gripping accuracy and ensuring accurate welding positions. The incoming material gripping module places the incoming material onto the spot welding fixture according to the corrected trajectory, providing a precise starting position for spot welding operations and ensuring welding quality. The spot welding operation movement module uses a servo slide to precisely control the spot welding position, achieving an automated spot welding process and improving the stability and efficiency of spot welding quality. After spot welding, the arc welding operation movement module precisely transfers the parts to the arc welding fixture to complete arc welding, achieving seamless connection between spot welding and arc welding, ensuring welding continuity, improving overall welding quality and production efficiency, reducing the uncertainty of manual intervention, and promoting the automation and intelligence of the welding process.

[0052] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in this application.

[0053] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0054] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0055] Figure 1 This is a schematic diagram of the internal functional modules of the automotive rear longitudinal beam assembly welding system based on 3D vision guidance and multi-fixture collaboration in an embodiment of the present invention.

[0056] Figure 2 This is a schematic diagram of the overall layout of a welding system for a rear longitudinal beam assembly of an automobile based on 3D vision guidance and multi-fixture collaboration, according to an embodiment of the present invention.

[0057] Figure 3 This is a schematic diagram of the overall layout of another automotive rear longitudinal beam assembly welding system based on 3D vision guidance and multi-fixture collaboration in an embodiment of the present invention.

[0058] Figure 4This is a structural diagram of the 3D vision system in an embodiment of the present invention;

[0059] Figure 5 This is a structural diagram of the spot welding fixture and servo slide in an embodiment of the present invention;

[0060] Figure 6 This is a structural diagram of the arc welding fixture and positioner in an embodiment of the present invention;

[0061] Figure 7 This is a flowchart of the assembly welding process in an embodiment of the present invention. Detailed Implementation

[0062] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0063] Example 1:

[0064] This invention provides a welding system for automotive rear longitudinal beam assemblies based on 3D vision guidance and multi-fixture collaboration, referenced Figures 1 to 7 ,include:

[0065] The incoming material transportation module is used to transport the automotive rear longitudinal beam from the loading area to the welding production line's bench garage position using an automated guided vehicle.

[0066] The grasping trajectory correction module is used to automatically correct the robot's grasping trajectory based on the spatial position and attitude data of the incoming automotive rear longitudinal beam, thereby obtaining a high-precision corrected grasping trajectory for the robot.

[0067] The incoming material gripping module is used to control the robot based on high-precision correction gripping trajectory to grip the incoming automotive rear longitudinal beam and place it on the spot welding fixture;

[0068] The spot welding operation moving module is used to move the spot welding fixture and the automotive rear longitudinal beam to the welding position based on the servo slide and complete the spot welding operation.

[0069] The arc welding operation moving module is used to control the robot to grab the spot-welded parts after spot welding and place them on the arc welding fixture located on the positioner to complete the arc welding operation.

[0070] In this embodiment, the spatial position and orientation data of the incoming rear longitudinal beam of the automobile are obtained by scanning and identification using a depth camera. This data reflects the specific position of the incoming rear longitudinal beam in three-dimensional space (such as its coordinates in the workshop coordinate system) and its own orientation (such as tilt, rotation angle, etc.), which is the key basic data for achieving precise gripping and welding. For example, the depth camera scans the coordinates of one end of the rear longitudinal beam as (x1, y1, z1) and the other end as (x2, y2, z2), and it rotates around each axis by a certain angle. These coordinates and angles are the spatial position and orientation data.

[0071] In this embodiment, the robot's grasping trajectory is the path of the robotic arm's movement when the robot grasps the incoming material from the rear longitudinal beam of the car. The initial trajectory is planned according to a preset standard model, but because the actual position and posture of the incoming material may deviate, subsequent correction is often required. For example, if the robot moves from the initial position A to point B according to the planned curve to grasp the incoming material, this curve from A to B is the grasping trajectory. The actual position of the incoming material may not be at point B, and the trajectory may need to be adjusted.

[0072] In this embodiment, the robot's high-precision calibrated grasping trajectory is based on the spatial position and attitude data of the incoming rear longitudinal beam of the automobile, and is a precise path obtained after correcting the initial grasping trajectory. The calibration comprehensively considers factors such as the deviation between the actual incoming material and the preset model, as well as the robot's own dynamic vibration offset, to ensure accurate grasping and placement and guarantee precise welding position. For example, after obtaining the actual incoming material data, a series of calibrations are performed to determine the precise motion path from point A to the actual incoming material position C, which is the high-precision calibrated grasping trajectory.

[0073] In this embodiment, the servo slide's ability to precisely control the position of the spot welding fixture and the incoming automotive rear longitudinal beam to the welding position and complete the spot welding operation refers to the servo slide's precise control over the positional movement. After the spot welding fixture clamps the incoming automotive rear longitudinal beam, the servo slide moves them precisely to a specific welding position according to a preset program. Once in place, the spot welding equipment is activated to complete the spot welding, achieving the initial connection and fixation of the parts. Similar to a production line, the servo slide acts like a precision transport track, carrying the spot welding fixture and the incoming material, smoothly and precisely moving them from the initial position to the designated welding position below the welding robot, where the welding robot then completes the spot welding.

[0074] In this embodiment, after spot welding is completed, the robot is controlled to grasp the spot-welded parts and place them onto the arc welding fixture located on the positioner, completing the arc welding operation. Specifically, after the initial connection of the rear longitudinal beam of the automobile is completed by spot welding, the robot is controlled to grasp the parts according to a high-precision calibrated grasping trajectory and fix them in the arc welding fixture on the positioner. The positioner adjusts the position and angle of the fixture and the parts to provide suitable welding conditions. Subsequently, the robot completes the arc welding, reinforcing the welded area and improving quality. In actual production, after spot welding is completed, the robot accurately grasps the spot-welded parts and places them onto the arc welding fixture on the positioner. The positioner adjusts the angle according to the program, and the arc welding robot performs arc welding to make the weld more robust.

[0075] In this embodiment, the system consists of the following subsystems or structures:

[0076] AGV transportation system:

[0077] Automated guided vehicles (AGVs) are responsible for transporting incoming materials from the loading area to the welding production line's designated parking spaces, enabling unmanned logistics and distribution.

[0078] 3D visual guidance system:

[0079] Installed above the robot's gripping position, it uses a depth camera to scan and identify the spatial position and posture data of the incoming material, guiding the robot to automatically correct its gripping trajectory.

[0080] Spot welding fixture and servo slide mechanism:

[0081] The spot welding fixture is moved to the welding station via a servo slide mechanism to ensure high precision and consistency in fixture positioning.

[0082] Positioner:

[0083] The arc welding fixture is installed on the positioner, which can quickly switch work stations and ensure high consistency.

[0084] Robot gripping and welding module:

[0085] Industrial robots complete the entire process of material handling, fixture switching, spot welding, and arc welding.

[0086] Unloading platform:

[0087] The final product is placed on the unloading platform by a robot for easy subsequent transfer.

[0088] Its working principle:

[0089] Material transportation and loading: After the AGV transports the incoming materials to the machine parking space, the 3D vision system identifies and guides the robot to grab the materials and place them on the spot welding fixture.

[0090] Spot welding operation: The spot welding fixture is moved to the welding position by the servo slide to complete the spot welding operation.

[0091] Arc welding operation: After spot welding is completed, the robot picks up the parts and places them on the arc welding fixture, which is located on the positioner to complete the welding.

[0092] Unloading and Transfer: The welded assembly is picked up by the robot and placed on the unloading table, waiting for transfer.

[0093] The beneficial effects of the above technical solutions are as follows: The incoming material transportation module utilizes an automated guided vehicle (AGV) to automate the transport of automotive rear longitudinal beams from a specific loading area to the welding production line's workstation, efficiently supplying materials, reducing manpower, and avoiding transportation deviations. The gripping trajectory correction module automatically corrects the robot's gripping trajectory based on the incoming material's spatial position and posture data, improving gripping accuracy and ensuring accurate welding positions. The incoming material gripping module places the incoming material onto the spot welding fixture according to the corrected trajectory, providing a precise starting position for spot welding operations and ensuring welding quality. The spot welding operation movement module uses a servo slide to precisely control the spot welding position, achieving an automated spot welding process and improving the stability and efficiency of spot welding quality. After spot welding, the arc welding operation movement module precisely transfers the parts to the arc welding fixture to complete arc welding, achieving seamless connection between spot welding and arc welding, ensuring welding continuity, improving overall welding quality and production efficiency, reducing the uncertainty of manual intervention, and promoting the automation and intelligence of the welding process.

[0094] Example 2:

[0095] Based on Example 1, the incoming material transportation module includes:

[0096] The initial path generation submodule is used to generate the initial transport path between the loading area and the welding production line's kiosks;

[0097] The obstacle avoidance mechanism triggering submodule is used to control the automated guided vehicle to pick up the automotive rear longitudinal beam material from the loading area and transport it to the welding production line's bench station according to the initial transport path. At the same time, based on the millimeter-wave radar on the automated guided vehicle, obstacles within a preset range centered on the automated guided vehicle are detected in real time. When the millimeter-wave radar detects an obstacle, the obstacle avoidance strategy is triggered to generate a new obstacle avoidance transport path, and the automotive rear longitudinal beam material is continued to be transported to the welding production line's bench station according to the new obstacle avoidance transport path.

[0098] The obstacle avoidance strategies include:

[0099] Multiple initial obstacle avoidance transport paths are generated based on the location of obstacles detected by millimeter-wave radar;

[0100] The path cost of each initial obstacle avoidance transportation path is calculated based on the path cost function:

[0101]

[0102] In the formula, The path cost of a single initial obstacle avoidance transport path. This represents the total number of locations in a single initial obstacle avoidance transport path. For the first obstacle avoidance transport path in a single initial obstacle avoidance transport path A location point, For the first obstacle avoidance transport path in a single initial obstacle avoidance transport path A location point, For the first obstacle avoidance transport path in a single initial obstacle avoidance transport path The location point and the first Euclidean distance between points This represents the total number of obstacles detected by the millimeter-wave radar. For the first obstacle avoidance transport path in a single initial obstacle avoidance transport path The x-coordinate of each location point For the first obstacle avoidance transport path in a single initial obstacle avoidance transport path The y-coordinate values ​​of each location point For the first The x-coordinate value of each obstacle For the first The vertical coordinate value of each obstacle;

[0103] The initial obstacle avoidance transport path with the minimum path cost among all initial obstacle avoidance transport paths is taken as the latest obstacle avoidance transport path.

[0104] In this embodiment, generating the initial transportation path between the loading area and the welding production line's trolley parking space refers to the system pre-planning a route from the loading area where the rear longitudinal beams of the vehicle are stored to the welding production line's trolley parking space before material transportation begins, providing directional guidance for the initial movement of the automated guided vehicle (AGV). For example, based on information such as workshop layout and equipment location, a relatively direct route that conforms to the production process is planned as the initial transportation path.

[0105] In this embodiment, the preset range centered on the automated guided vehicle (AGV) refers to a specific spatial area defined around the AGV to determine the detection range of the millimeter-wave radar. The setting of this range needs to comprehensively consider factors such as the AGV's driving speed, braking capability, and the possible distribution of obstacles within the workshop to ensure timely detection of obstacles that may affect its movement. For example, a circular area with a radius of 5 meters centered on the AGV can be set as the preset range.

[0106] In this embodiment, the millimeter-wave radar on the automated guided vehicle (AGV) detects obstacles within a preset range centered on the AGV in real time. Specifically, the millimeter-wave radar continuously monitors the designated area for obstacles. By transmitting and receiving millimeter-wave signals and analyzing the characteristics of the reflected waves, the millimeter-wave radar determines the obstacle's position, speed, and other information, providing real-time data support for the AGV's obstacle avoidance decisions. For example, while the AGV is moving, the millimeter-wave radar continuously scans the surrounding 5-meter range. Once an object is detected entering this range, the relevant information is immediately fed back to the control system.

[0107] In this embodiment, the latest obstacle avoidance transport path refers to a new path that the system replans based on certain algorithms and strategies after the millimeter-wave radar detects an obstacle. This new path avoids the current obstacle, allowing the automated guided vehicle (AGV) to continue smoothly transporting the automotive rear longitudinal beam to the welding production line's docking station. This path must bypass the obstacle while maximizing transport efficiency and avoiding excessive increases in transport time and distance. For example, if the AGV was originally traveling in a straight line, and upon detecting an obstacle ahead, the system plans a curve that detours to the left as the latest obstacle avoidance transport path.

[0108] In this embodiment, generating multiple initial obstacle avoidance transport paths based on the obstacle positions detected by millimeter-wave radar means that after knowing the obstacle position information, the system uses a specific algorithm to generate multiple different candidate paths that can avoid the obstacle, starting from the current position of the automated guided vehicle and ending at the parking space in the garage. These paths take into account different detour directions, distances, and other factors, providing multiple possibilities for selecting the optimal path subsequently. For example, if the obstacle is slightly to the right in front of the automated guided vehicle, the system may generate multiple different initial obstacle avoidance transport paths, such as detouring to the left, detouring backward, and then moving forward.

[0109] In this embodiment, the path cost of the initial obstacle avoidance transportation path is calculated using a specific path cost function to measure the merits of each initial obstacle avoidance transportation path. The path cost value comprehensively considers factors such as path length and distance from obstacles. A lower path cost value indicates that the path is more advantageous in terms of safety and transportation efficiency, and the system will tend to select the path with the lowest path cost value as the final obstacle avoidance transportation path. For example, based on a given path cost function, if one initial obstacle avoidance transportation path has a cost value of 0.8 and another has a cost value of 0.6, the path with a cost value of 0.6 is relatively better.

[0110] In this embodiment, the preset adjustment factor is a parameter in the path cost function. It is used to adjust the calculation of the path cost value, balancing the relative importance of factors such as path length and obstacle distance in path evaluation. For example, when calculating the path cost value of the initial obstacle avoidance transportation path, if the preset adjustment factor is larger, the influence weight of the distance to the obstacle on the path cost value will increase in the path evaluation, making the system more inclined to choose a path farther from the obstacle; conversely, if the value is smaller, the influence weight of the path length factor will be relatively greater, and the system may focus more on choosing a shorter path, even if it is slightly closer to the obstacle. Its specific value needs to be preset according to different requirements for safety and transportation efficiency in the actual scenario.

[0111] The beneficial effects of the above technical solution are as follows: The initial path generation submodule can autonomously generate the initial transportation path between the loading area and the welding production line's workstation, clearly defining the starting direction for material transportation and ensuring orderly commencement of transport. The obstacle avoidance mechanism triggering submodule enables the automated guided vehicle (AGV) to monitor obstacles within a preset range in real time using millimeter-wave radar while transporting along the initial path. Once an obstacle is detected, the obstacle avoidance strategy is immediately triggered, quickly generating a new obstacle-avoidance transportation path to avoid collisions, ensuring safe and continuous material transportation and reducing production delays. The obstacle avoidance strategy generates multiple initial obstacle-avoidance transportation paths and uses a path cost function to calculate the path cost value. It comprehensively considers the relationship between path length and obstacle distance, ultimately selecting the path with the minimum path cost value as the new obstacle-avoidance transportation path. This achieves safe obstacle avoidance while also considering path economy and efficiency, reducing path growth and time consumption caused by obstacle avoidance, ensuring transportation efficiency, and timely supplying materials to the welding production line.

[0112] Example 3:

[0113] Based on Example 1, the capture trajectory correction module includes:

[0114] The incoming material pose scanning submodule is used to scan and identify the spatial position and pose data of the incoming automotive rear longitudinal beam using a depth camera.

[0115] The spatial pose matching submodule is used to match the spatial position and attitude data of the incoming automotive rear longitudinal beam with the preset 3D model of the incoming material based on the point cloud registration algorithm, and generate a spatial pose deviation matrix.

[0116] The grasping trajectory correction submodule is used to automatically correct the robot's grasping trajectory based on the spatial pose deviation matrix, thereby obtaining a high-precision corrected grasping trajectory for the robot.

[0117] In this embodiment, scanning and identifying the spatial position and orientation data of the incoming rear longitudinal beam of the automobile using a depth camera involves scanning the incoming material with a depth camera. The depth camera can acquire three-dimensional spatial information of the object. By emitting and receiving light signals and calculating the round-trip time of the light rays, it determines the distance between each point on the object's surface and the camera, thereby identifying the specific position of the incoming rear longitudinal beam in space, such as its coordinates in a certain coordinate system, as well as its orientation information, such as tilt angle and rotation direction. This provides basic data for subsequent precise handling of the incoming material. For example, the depth camera can scan the coordinates (x, y, z) of a certain endpoint of the rear longitudinal beam in a specific coordinate system and obtain the angle of rotation of the material around a certain axis. These data constitute the spatial position and orientation data of the incoming rear longitudinal beam.

[0118] In this embodiment, a point cloud registration algorithm is used to match the spatial position and orientation data of the incoming rear longitudinal beam of the automobile with a preset 3D model of the incoming material, generating a spatial pose deviation matrix. This involves presenting the spatial position and orientation data of the incoming rear longitudinal beam of the automobile, acquired by a depth camera, in the form of a point cloud. The point cloud contains coordinate information of numerous points representing feature points on the object's surface. Then, the point cloud registration algorithm is used to match these actually measured point cloud data with the point cloud data of the preset 3D model of the automobile rear longitudinal beam. Through algorithm calculation, the differences in spatial position and orientation between the actual incoming material and the preset model are identified, and these differences are quantified into a matrix form, namely the spatial pose deviation matrix. This matrix clearly reflects the deviations of the actual incoming material from the standard model in terms of translation and rotation, providing crucial information for subsequent correction of the robot's grasping trajectory. For example, after calculation by the point cloud registration algorithm, the resulting spatial pose deviation matrix can indicate information such as the actual incoming material translating a certain distance in the x-direction and rotating a certain angle around the y-axis.

[0119] The beneficial effects of the above technical solutions are as follows: The incoming material pose scanning submodule uses a depth camera to accurately scan and identify the spatial position and posture data of the incoming automotive rear longitudinal beam, providing accurate raw information for subsequent correction and making the acquisition of the incoming material's state more accurate. The spatial pose matching submodule uses a point cloud registration algorithm to match the scanned data with a preset 3D model of the incoming material and generate a spatial pose deviation matrix. This comparative analysis clearly shows the difference between the actual incoming material and the standard model, providing a quantitative basis for trajectory correction. The grasping trajectory correction submodule automatically corrects the robot's grasping trajectory based on the spatial pose deviation matrix, obtaining a high-precision corrected grasping trajectory, greatly improving the accuracy of the robot's grasping, enabling the robot to accurately grasp the incoming material, reducing problems such as inaccurate welding positions caused by grasping deviations, effectively ensuring welding quality, improving the reliability and stability of production, and helping the automotive rear longitudinal beam assembly welding system operate more efficiently and accurately.

[0120] Example 4:

[0121] Based on Example 3, the capture trajectory correction submodule includes:

[0122] The initial correction unit is used to automatically correct the robot's grasping trajectory based on the spatial pose deviation matrix, and obtain the robot's initial corrected grasping trajectory.

[0123] The dynamic model generation unit is used to generate the robot's state vector at each time step based on the robot's dynamic vibration offset and the rate of change of the dynamic vibration offset at each time step, and to generate the robot's dynamic model by combining the state transition matrix, the control input matrix, and the process noise vector.

[0124] The measurement model generation unit is used to generate a measurement model of the robot based on the robot's dynamic model, measurement matrix, and measurement noise vector;

[0125] The vibration offset estimation unit is used to estimate the dynamic vibration offset of the robot based on the robot's dynamic model, measurement model, and Kalman filter.

[0126] The reverse compensation unit is used to perform reverse displacement compensation on the robot's initial corrected grasping trajectory based on the robot's dynamic vibration offset, so as to obtain the robot's high-precision corrected grasping trajectory.

[0127] In this embodiment, the robot's initial corrected grasping trajectory is obtained by making preliminary adjustments to the robot's original grasping trajectory based on the spatial pose deviation matrix. It makes preliminary corrections to the robot's grasping path based on the deviation between the actual spatial position and posture of the incoming material and the preset model, but does not yet consider the impact of dynamic vibration during the robot's operation, thus laying the foundation for further precise correction.

[0128] In this embodiment, based on the robot's dynamic vibration offset at each moment... and the rate of change of dynamic vibration offset Generating the robot's state vector at each moment, and combining it with the state transition matrix, control input matrix, and process noise vector to generate the robot's dynamic model, refers to the continuous change of the robot's dynamic vibration offset (the deviation between the robot's actual position and its ideal position) and the rate of change of dynamic vibration offset over time during the robot's operation. These time-varying quantities are combined to form the state vector, which describes the robot's state at each moment. The state transition matrix describes the natural evolution of the robot's state from one moment to the next, the control input matrix reflects the influence of external control commands on the robot's state, and the process noise vector covers various unavoidable random disturbances during robot operation. The dynamic model constructed by integrating these elements can accurately characterize the robot's dynamic behavior under the influence of various factors, providing a basis for more accurate analysis and correction of the robot's grasping trajectory.

[0129] in, State vector at time step ,in, It is an n-dimensional dynamic vibration offset vector. It is its rate of change vector;

[0130] The dynamic model of the robot can be represented as ;

[0131] In the formula, It is a 2n×2n state transition matrix, describing the system state from time 1 to 2n. At the time The evolution of. For example, for simple linear systems, It can be represented as ,in It is an n-dimensional identity matrix. It is a time interval. It is an n-dimensional zero matrix; yes The state vector at any given time; It is a 2n×m control input matrix; It is an m-dimensional control input vector. In this scenario, if there is no external control input, The item can be ignored; It is a 2n-dimensional process noise vector with a mean of zero and a covariance matrix of... Gaussian distribution, Let be the variance of the Gaussian distribution.

[0132] In this embodiment, the measurement model of the robot, generated based on the robot's dynamic model, measurement matrix, and measurement noise vector, takes into account the actual measurement process while building upon the established robot dynamic model. The measurement matrix links the robot's internal state with the actual measurable physical quantities, and the measurement noise vector represents the unavoidable errors and interferences during the measurement process. By combining the dynamic model, measurement matrix, and measurement noise vector to generate the measurement model, the relationship between the actual measured values ​​and the robot's true state can be reflected more accurately, thus providing an effective way to estimate the robot's true state and helping to more accurately correct the robot's grasping trajectory.

[0133] The measurement model can be represented as:

[0134]

[0135] In the formula, It is a p-dimensional measurement vector, which can be obtained, for example, by measuring an accelerometer or displacement sensor mounted on a robot; It is a p×2n measurement matrix that describes the relationship between the system state and the measured values; It is a p-dimensional measurement noise vector with a mean of zero and a covariance matrix of... Gaussian distribution, This is the covariance matrix for measuring noise.

[0136] In this embodiment, the dynamic vibration offset of the robot is estimated based on the robot's dynamic model, measurement model, and Kalman filter. This estimation utilizes the Kalman filter algorithm. The Kalman filter can predict the robot's state at the next moment based on the dynamic model and, combined with the actual measurement values ​​obtained from the measurement model, iteratively estimates the robot's dynamic vibration offset. It effectively handles uncertainties in the dynamic and measurement models, filters out noise interference, and thus obtains a relatively accurate estimate of the robot's dynamic vibration offset. This provides crucial data for subsequent precise correction of the grasping trajectory.

[0137] Prediction steps:

[0138]

[0139]

[0140] In the formula, At any moment Based on time The estimated value is a prediction of the system state;

[0141] It is the covariance matrix of the predicted state;

[0142] It is a moment The optimal estimate;

[0143] It is a moment The covariance matrix of the optimal estimate;

[0144] for The transpose of .

[0145] Update steps:

[0146]

[0147]

[0148]

[0149] In the formula, It is the Kalman gain matrix; for The transpose of the matrix; It is the optimal estimate of time k; It is a moment The covariance matrix of the optimal estimate;

[0150] By continuously performing prediction and update steps, the Kalman filter can estimate the robot's dynamic vibration offset in real time. It is The first n elements.

[0151] In this embodiment, the initial corrected grasping trajectory of the robot is compensated for by reverse displacement based on the robot's dynamic vibration offset to obtain a high-precision corrected grasping trajectory. This means that after obtaining the robot's dynamic vibration offset, this offset will cause deviations in the robot's grasping trajectory. To eliminate this deviation, the initial corrected grasping trajectory is compensated for by displacement in the opposite direction to the vibration offset direction. For example, if dynamic vibration causes the robot to deviate in a certain direction, it is moved the same distance in the opposite direction, thus obtaining a high-precision corrected grasping trajectory that takes into account the influence of dynamic vibration. This ensures that the robot can accurately grasp the incoming material of the rear longitudinal beam of the automobile, improving the accuracy and quality of welding.

[0152] Let the original trajectory planning position be... The dynamic vibration offset estimated by the Kalman filter is The corrected grasping trajectory position It can be represented as: ;

[0153] In the formula, and They are all n-dimensional position vectors, with units in meters (m); It is an n-dimensional dynamic vibration offset vector, and its unit is also meters (m).

[0154] The beneficial effects of the above technical solutions are as follows: The initial correction unit performs preliminary correction on the robot's grasping trajectory based on the spatial pose deviation matrix, quickly narrowing the gap between the grasping trajectory and the ideal state, laying the foundation for more accurate subsequent corrections. The dynamic model generation unit comprehensively considers the robot's dynamic vibration offset and rate of change at each moment, combining the state transition matrix, control input matrix, and process noise vector to generate a dynamic model, comprehensively and accurately describing the robot's dynamic characteristics, providing accurate model support for further optimization of the grasping trajectory. The measurement model generation unit generates a measurement model based on the robot's dynamic model, measurement matrix, and measurement noise vector, making the measurement of the robot's state more scientific and reasonable, improving the accuracy and reliability of the measurement. The vibration offset estimation unit uses a Kalman filter to accurately estimate the robot's dynamic vibration offset based on the dynamic model and measurement model, effectively removing noise interference and extracting true vibration offset information. The reverse compensation unit performs reverse displacement compensation on the initial correction gripping trajectory based on the estimated dynamic vibration offset. Through this targeted compensation method, the trajectory deviation caused by robot vibration is effectively eliminated, and a high-precision correction gripping trajectory is finally obtained. This greatly improves the accuracy of the robot's gripping action, thereby significantly improving the positional accuracy of the gripping and placement of the rear longitudinal beam of the automobile, further ensuring welding quality, improving the stability and reliability of the entire welding system, and reducing product defects and production delays caused by inaccurate gripping.

[0155] Example 5:

[0156] Based on Example 4, the initial correction unit includes:

[0157] The position correction subunit is used to correct the position of each position point in the grasping trajectory of the robot based on the translation vector in the spatial pose deviation matrix, so as to obtain the position correction grasping trajectory of the robot.

[0158] The attitude correction subunit is used to adjust the attitude of each position point in the robot's position correction grasping trajectory based on the rotation matrix in the spatial pose deviation matrix, so as to obtain the robot's initial correction grasping trajectory.

[0159] In this embodiment, the translation vector in the spatial pose deviation matrix describes the difference between the actual position of the incoming material of the rear longitudinal beam of the automobile and the preset model position in the spatial translation direction. It consists of three components, corresponding to the x, y, and z axes in three-dimensional space, respectively, intuitively reflecting the amount of translation of the incoming material relative to the standard position along each coordinate axis. For example, if the translation vector is (10, -5, 3), it means that the incoming material has moved 10 units more in the x-axis direction than the preset position, 5 units less in the y-axis direction, and 3 units more in the z-axis direction.

[0160] In this embodiment, the position of each point in the robot's grasping trajectory is corrected based on the translation vector in the spatial pose deviation matrix to obtain the robot's position-corrected grasping trajectory. This involves adjusting each point on the robot's grasping trajectory according to the material's position deviation reflected by the translation vector. The points on the grasping trajectory are displaced along the x, y, and z axes according to the corresponding components of the translation vector, ensuring the robot's grasping trajectory matches the actual position of the material, thus obtaining the position-corrected grasping trajectory. For example, if the initial coordinates of a point on the grasping trajectory are (x0, y0, z0) and the translation vector is (Δx, Δy, Δz), then after correction, the coordinates of that point become (x0+Δx, y0+Δy, z0+Δz). This process is repeated for all points to obtain the position-corrected grasping trajectory.

[0161] In this embodiment, the robot's position-corrected grasping trajectory is obtained by correcting the original grasping trajectory based on translation vectors. This trajectory takes into account the spatial deviation of the incoming rear longitudinal beam of the automobile relative to the preset model, making the robot's grasping path more closely match the actual position of the incoming material, providing a positional basis for subsequent precise grasping, but the grasping posture has not yet been adjusted.

[0162] In this embodiment, the rotation matrix in the spatial pose deviation matrix is ​​used to describe the rotational difference between the actual posture of the incoming material of the rear longitudinal beam of the automobile and the posture of the preset model. It is a 3×3 matrix, and the rotation transformation of the coordinate system can be realized through matrix operations, reflecting the rotation of the incoming material around the x, y, and z axes in three-dimensional space, and determining the rotation angle and direction of the incoming material relative to the standard posture.

[0163] In this embodiment, the posture of each point in the robot's position correction grasping trajectory is adjusted based on the rotation matrix in the spatial pose deviation matrix to obtain the robot's initial corrected grasping trajectory. This is achieved by using the rotation matrix to correct the posture of each point on the already position-corrected grasping trajectory. By performing matrix operations between the rotation matrix and the coordinates of each point on the position correction grasping trajectory, each point rotates according to the rotation of the actual posture of the incoming material relative to the posture of the preset model, thereby adjusting the robot's posture during grasping. Finally, an initial corrected grasping trajectory that considers both position deviation and posture deviation is obtained, further improving the accuracy and stability of the robot's grasping.

[0164] The beneficial effects of the above technical solutions are as follows: The position correction subunit uses the translation vector in the spatial pose deviation matrix to correct the position of each point in the robot's grasping trajectory, precisely adjusting the robot's grasping position and effectively compensating for the deviation between the incoming material's spatial position and the preset model. This makes the robot's grasping position more closely match the actual incoming material position, providing a foundation for subsequent precise grasping. The attitude correction subunit adjusts the attitude of each point in the position-corrected grasping trajectory based on the rotation matrix in the spatial pose deviation matrix, meticulously correcting the robot's grasping attitude. This ensures that the robot can grasp the incoming material at a suitable angle, avoiding problems such as unstable grasping or inaccurate placement due to improper attitude. Through the orderly execution of position and attitude corrections, not only is the spatial pose deviation handled step by step with precision, but the robot's grasping trajectory is also gradually optimized. The final initial corrected grasping trajectory is more accurate, further improving the accuracy and stability of the robot's grasping. This effectively guarantees the quality of the grasping and placement process before welding the automotive rear longitudinal beam, reduces welding defects caused by grasping problems, and improves the overall production efficiency and product quality of the welding system.

[0165] Example 6:

[0166] Based on Example 1, it also includes:

[0167] The fixture contact pressure detection module is used to detect the real-time contact pressure distribution data between the incoming automotive rear longitudinal beam and the fixture based on the pressure sensor array inside the spot welding fixture.

[0168] The fixture support posture adjustment module is used to adjust the support posture of the fixture based on the real-time contact pressure distribution data between the automotive rear longitudinal beam material and the fixture.

[0169] In this embodiment, the pressure sensor array inside the spot welding fixture consists of multiple pressure sensors installed on the inner part of the spot welding fixture where it contacts the incoming rear longitudinal beam of the automobile. These sensors work together to detect the pressure at the contact point between the fixture and the incoming material in real time, providing data support for determining the placement status of the material within the fixture. For example, after the incoming rear longitudinal beam of the automobile is placed onto the spot welding fixture, the pressure sensor array starts working, collecting pressure information from each contact point.

[0170] In this embodiment, the real-time contact pressure distribution data between the incoming rear longitudinal beam of the automobile and the fixture is acquired in real time through a pressure sensor array inside the spot welding fixture. This data reflects the pressure magnitude and distribution at different locations on the contact surface between the incoming rear longitudinal beam and the fixture. By analyzing this data, it can be determined whether the incoming material is placed stably in the fixture and whether the force is uniform. For example, if the pressure in a certain area is too high or too low, it may indicate that the incoming material is placed tilted or there is a problem with that part of the fixture. This data provides a basis for subsequent adjustments to the fixture's support posture.

[0171] In this embodiment, the support posture of the fixture refers to the position and orientation of the spot welding fixture when supporting the incoming rear longitudinal beam of the automobile. It includes the spatial position of the fixture (such as its coordinates in the welding station) and its own orientation information, such as angle and tilt. A suitable support posture ensures that the incoming rear longitudinal beam remains stable during spot welding, preventing displacement or deformation due to external forces, thereby guaranteeing the quality of the spot weld. For example, the fixture needs to be adjusted to a specific position and angle according to the shape and size of the incoming material to ensure stable support within the fixture.

[0172] The beneficial effects of the above technical solution are as follows: The fixture contact pressure detection module, through the pressure sensor array inside the spot welding fixture, can accurately detect the contact pressure distribution data between the incoming automotive rear longitudinal beam and the fixture in real time. This data directly reflects the degree of fit between the incoming material and the fixture, as well as the stress situation, providing an accurate basis for subsequent adjustments. The fixture support posture adjustment module adjusts the support posture of the fixture in a targeted manner based on the real-time contact pressure distribution data. This helps ensure that the incoming automotive rear longitudinal beam receives uniform and stable support in the fixture, avoiding problems such as deformation and displacement caused by excessive or insufficient local pressure. By adjusting the fixture support posture in a timely manner, the incoming material is placed in the optimal position and posture before spot welding, greatly improving the accuracy and quality of spot welding, reducing welding defects caused by poor contact between the fixture and the incoming material, improving the overall quality and stability of the automotive rear longitudinal beam assembly welding, ensuring the smooth progress of the welding production process, improving production efficiency, and reducing the defect rate.

[0173] Example 7:

[0174] Based on Example 6, the clamp support pose adjustment module includes:

[0175] The sliding trend prediction submodule is used to analyze the sliding trend of the automotive rear longitudinal beam material in the fixture based on the real-time contact pressure distribution data between the automotive rear longitudinal beam material and the fixture.

[0176] The support posture adjustment submodule is used to adjust the support posture of the fixture based on the sliding trend of the automotive rear longitudinal beam material in the fixture.

[0177] In this embodiment, the sliding trend of the automotive rear longitudinal beam material in the fixture is derived from the analysis of real-time contact pressure distribution data between the material and the fixture, reflecting the possibility and direction of the material's sliding tendency within the fixture. By calculating the total normal pressure, friction force, resultant force, and maximum static friction force in the contact area, and comparing the tangential component of the resultant force with the maximum static friction force, it is determined whether the material has a risk of sliding and the possible direction of sliding. For example, when the tangential component of the resultant force is greater than the maximum static friction force, the material has a sliding tendency. Analyzing these data can also reveal the magnitude and approximate direction of the sliding tendency, providing a basis for adjusting the fixture support posture in advance and preventing the material from sliding during spot welding, thereby ensuring welding quality.

[0178] The beneficial effects of the above technical solution are as follows: The sliding trend prediction submodule analyzes the sliding trend of the automotive rear longitudinal beam material in the fixture based on real-time contact pressure distribution data. This allows the system to predict potential sliding situations in advance, transforming passive response into proactive prevention. By deeply analyzing pressure distribution data, the potential sliding direction and probability of the material are accurately captured, providing highly forward-looking information for subsequent adjustments. The support posture adjustment submodule adjusts the support posture of the fixture in a timely manner based on the predicted sliding trend, effectively avoiding positional offset problems caused by material sliding. This targeted adjustment can better constrain the material, keeping it stable during spot welding and further improving spot welding quality. By preventing welding defects that may be caused by sliding in advance, the production of defective products is reduced, production efficiency is improved, and the stability of the welding process and the consistency of product quality are also ensured, making the automotive rear longitudinal beam assembly welding system operate more reliably and efficiently in actual production.

[0179] Example 8:

[0180] Based on Example 7, the sliding trend prediction submodule includes:

[0181] The normal pressure calculation unit is used to calculate the total normal pressure in the contact area between the automotive rear longitudinal beam material and the fixture based on the real-time contact pressure distribution data between the automotive rear longitudinal beam material and the fixture.

[0182] The resultant force analysis unit is used to calculate the frictional force on the rear longitudinal beam material in the fixture based on the total normal pressure in the contact area between the rear longitudinal beam material and the fixture, and to determine the resultant force on the rear longitudinal beam material in the fixture in combination with other external forces on the rear longitudinal beam material.

[0183] The maximum static friction analysis unit is used to calculate the maximum static friction between the automotive rear longitudinal beam material and the fixture based on the total normal pressure of the contact area between the material and the fixture.

[0184] The sliding trend analysis unit is used to analyze the sliding trend of the automotive rear longitudinal beam material in the fixture based on the tangential component of the resultant force and the maximum static friction force acting on the material.

[0185] In this embodiment, the total normal pressure in the contact area between the automotive rear longitudinal beam material and the fixture is calculated based on real-time contact pressure distribution data. This involves integrating and calculating the pressure data from each contact point acquired by the pressure sensor array using a specific mathematical method. Since this pressure data reflects the normal pressure at different locations, the total normal pressure value for the entire contact area can be obtained through summation or other related calculations. For example, the total normal pressure is obtained by adding the pressure values ​​measured by each pressure sensor, reflecting the magnitude of the force acting vertically between the material and the fixture. Furthermore, within the contact area, the normal pressure can be obtained by integrating the contact pressure distribution data. For a tiny area element... Normal pressure It can be represented as: For the entire contact area Integrating, we obtain the total normal pressure. : ;

[0186] In the formula, This represents the coordinates of the points within the contact area. It is a two-dimensional matrix, where each element represents the pressure value at the corresponding coordinate point, in Pascals (Pa).

[0187] In this embodiment, the frictional force on the rear longitudinal beam material in the fixture is calculated based on the total normal pressure in the contact area between the incoming material and the fixture. This frictional force is then combined with other external forces acting on the material to determine the resultant force on the fixture. According to the frictional force calculation formula, frictional force is usually proportional to the normal pressure. The frictional force can be calculated using the total normal pressure and the coefficient of friction (which depends on factors such as the material and fixture materials). Besides frictional force, the material may also be subjected to other external forces on the fixture, such as gravity and forces generated during welding. These external forces are combined with the calculated frictional force according to the rules of force composition (such as the parallelogram rule or the triangle rule) to determine the resultant force on the rear longitudinal beam material in the fixture. For example, given that the total normal pressure is... The coefficient of friction is Calculate the friction force Then, by combining other external forces, the magnitude and direction of the resultant force can be obtained through vector addition.

[0188] In this embodiment, the maximum static friction force between the automotive rear longitudinal beam material and the fixture is calculated based on the total normal pressure in the contact area between the material and the fixture. According to the formula for calculating the maximum static friction force, it is directly proportional to the normal pressure. Similarly, it is calculated using the total normal pressure and the static friction coefficient (which depends on the material properties of the material and the fixture, etc.). For example, let the static friction coefficient be... The total normal pressure is Then the maximum static friction force F. This maximum static friction value represents the maximum tangential force that the material can withstand while remaining stationary on the fixture. Once the external force exceeds this value, the material may begin to slide.

[0189] In this embodiment, the sliding tendency of the automotive rear longitudinal beam material in the fixture is analyzed based on the tangential component of the resultant force and the maximum static friction force acting on the material. The resultant force can be decomposed into a normal component perpendicular to the contact surface and a tangential component along the contact surface. Comparing the tangential component of the resultant force with the maximum static friction force, if the tangential component is less than the maximum static friction force, it indicates that the force causing the material to slide is insufficient to overcome the maximum static friction force, and the material remains stationary relative to the fixture, with a small sliding tendency. Conversely, if the tangential component of the resultant force is greater than the maximum static friction force, the material has a sliding tendency, and the greater the tangential component exceeds the maximum static friction force, the greater the sliding tendency. Through this comparative analysis, the sliding tendency of the material in the fixture can be accurately determined, providing a key basis for timely adjustment of the fixture's support posture and preventing the material from affecting the welding quality during spot welding.

[0190] The beneficial effects of the above technical solutions are as follows: The normal pressure calculation unit accurately calculates the total normal pressure in the contact area between the incoming material and the fixture by processing real-time contact pressure distribution data, providing basic data support for subsequent analysis. This data accurately reflects the vertical force between the incoming material and the fixture, and is a key starting point for studying their interaction. The resultant force analysis unit calculates the frictional force based on the total normal pressure and determines the resultant force on the incoming material in the fixture by integrating other external forces, comprehensively considering various force factors affecting the stability of the incoming material. This comprehensive force analysis method makes the understanding of the force situation of the incoming material more comprehensive and in-depth, and can more accurately grasp the actual state of the incoming material in the fixture. The maximum static friction force analysis unit calculates the maximum static friction force based on the total normal pressure, clarifying the critical condition for the incoming material to remain stationary in the fixture. This value is crucial for judging whether the incoming material will slide, providing a key reference for sliding trend analysis. The sliding trend analysis unit combines the tangential component of the resultant force with the maximum static friction force to accurately analyze the sliding trend of the incoming material in the fixture. By comparing these two key parameters, it is possible to accurately predict whether there is a risk of material slippage, as well as the direction and likelihood of slippage. This precise prediction of slippage trends allows the fixture support posture adjustment module to react in advance and accurately, adjusting the fixture support posture in a timely manner. This effectively prevents material slippage during spot welding, thereby further improving spot welding quality, ensuring product consistency and stability, reducing defects caused by slippage, and enhancing the production efficiency and reliability of the entire welding system.

[0191] Example 9:

[0192] Based on Example 7, a pose adjustment submodule is provided, including:

[0193] An optimization objective function construction unit is used to construct an objective function with the goal of minimizing the sliding trend when the sliding trend of the automotive rear longitudinal beam material in the fixture exceeds the sliding trend threshold.

[0194] The pose optimization unit is used to solve for the optimal support pose vector of the fixture based on the objective function and the upper and lower limits of the support pose vector of the fixture, and to control the support pose of the fixture based on the optimal support pose vector of the fixture.

[0195] In this embodiment, the sliding trend threshold is a pre-set standard value used to determine whether the sliding trend of the incoming material of the automotive rear longitudinal beam in the fixture is within an acceptable range. When the calculated sliding trend of the incoming material exceeds this threshold, it indicates that the sliding risk is relatively high, and the support posture of the fixture needs to be adjusted to prevent the incoming material from sliding during spot welding and affecting the welding quality; if it does not exceed the threshold, it is considered that the sliding risk of the incoming material is within a controllable range under the current support posture of the fixture.

[0196] In this embodiment, the objective function is constructed with the goal of minimizing the sliding tendency, aiming to find a method that can effectively reduce the sliding tendency of the automotive rear longitudinal beam material in the fixture. Through mathematical modeling, various factors affecting the sliding tendency, such as the support angle and position of the fixture, are incorporated as variables into the function. The form and specific content of the objective function will be determined according to the actual situation and the mathematical method used, but its core objective is to minimize the sliding tendency, providing a mathematical basis for subsequently solving the optimal support pose of the fixture.

[0197] In this embodiment, the upper and lower limits of the fixture's support pose vector refer to the restrictions on the value range of each component in the vector describing the fixture's support pose. The fixture's support pose can be represented by a vector, where each component corresponds to a position or orientation parameter of the fixture in space (such as x, y, z coordinates, rotation angle, etc.). The upper and lower limits are set based on factors such as the fixture's physical structural limitations, workspace requirements, and welding process requirements. For example, the fixture's movement distance in a certain direction cannot exceed a certain range, or the rotation angle must be within a specific interval; these restrictions constitute the upper and lower limits of each component of the support pose vector.

[0198] In this embodiment, the optimal support pose vector of the fixture is determined based on the objective function and the upper and lower limits of the fixture's support pose vector. This is achieved using mathematical optimization algorithms to find the fixture support pose vector that minimizes the objective function (i.e., minimizes the sliding tendency) under the constraints of the upper and lower limits of the support pose vector. These optimization algorithms can be linear programming, nonlinear programming, or other methods. Through iterative calculation and optimization, a set of fixture support pose parameters that minimizes the material sliding tendency under given conditions is finally obtained, i.e., the optimal support pose vector.

[0199] In this embodiment, the optimal support pose vector of the fixture is a set of parameter values ​​obtained through the above-described solution process. It represents a precise description of the fixture support pose that minimizes the sliding tendency of the automotive rear longitudinal beam material within the fixture under the current conditions. Each component of this set of vectors corresponds to the specific position or orientation information of the fixture in space. Adjusting the fixture according to the pose determined by this set of vectors can effectively improve the stability of the incoming material during the spot welding process and ensure welding quality.

[0200] In this embodiment, controlling the support posture of the fixture based on its optimal support posture vector involves converting the solved optimal support posture vector into actual control signals and sending them to the fixture's drive system or adjustment mechanism. The fixture automatically adjusts its position and orientation according to these control signals to achieve the state determined by the optimal support posture vector. This ensures stable support for the incoming automotive rear longitudinal beam during spot welding, minimizing the risk of slippage and improving welding accuracy and quality.

[0201] The beneficial effects of the above technical solution are as follows: When the sliding trend of the incoming material of the automotive rear longitudinal beam exceeds a threshold, the optimization objective function construction unit constructs an objective function with the goal of minimizing the sliding trend. This provides a clear and targeted optimization direction for fixture pose adjustment, namely, to minimize the sliding trend of the incoming material, ensure the stability of the incoming material in the fixture, and provide a reliable foundation for subsequent welding processes. Based on the constructed objective function and the upper and lower limits of the fixture support pose vector, the pose optimization unit solves for the optimal support pose vector of the fixture and controls the fixture support pose accordingly. This method of solving for the optimal solution fully considers the physical limitations of the fixture itself, finds the pose most conducive to reducing the sliding trend within the feasible range, and makes the adjustment process more scientific and precise. Through precise pose adjustment, the sliding of the incoming material is effectively suppressed, the spot welding quality is significantly improved, the product defects caused by sliding are reduced, and the product qualification rate is improved. At the same time, the optimized fixture support pose reduces unnecessary adjustments, improves production efficiency, ensures the stable and efficient operation of the welding system, and brings better quality control and economic benefits to the welding production of automotive rear longitudinal beam assemblies.

[0202] Example 10:

[0203] Based on Example 1, the arc welding fixture includes:

[0204] The clamping force is adjustable within a preset range via a pneumatic clamping mechanism and thermocouple array controlled by a proportional valve.

[0205] Among them, the thermocouple array monitors the real-time temperature of the welding area on the part after spot welding, and dynamically adjusts the welding current based on the real-time temperature of the welding area on the part after spot welding.

[0206] In this embodiment, the clamping force is adjustable within a preset range by a proportional valve, meaning that the clamping force of the pneumatic clamping mechanism on the arc welding fixture can be flexibly changed according to actual needs. The proportional valve, as a control element, can proportionally adjust the fluid flow rate according to the input signal, thereby controlling the pneumatic clamping mechanism to generate different clamping forces. The preset range is a clamping force interval pre-set based on the characteristics of the part and the welding process requirements, ensuring that the part is firmly clamped to prevent movement during arc welding without deforming due to excessive clamping force. For example, for parts of different thicknesses or materials after spot welding, the clamping force can be varied within a suitable range of Newton values ​​by adjusting the proportional valve.

[0207] In this embodiment, the thermocouple array, composed of multiple thermocouples, is mounted on the arc welding fixture to monitor the temperature of the welding area of ​​the part in real time after spot welding. A thermocouple is a temperature sensor that operates based on the thermoelectric effect, converting temperature signals into electrical signals. An array of multiple thermocouples can collect temperature data from multiple locations, providing a more comprehensive and accurate reflection of the temperature distribution in the welding area, and offering precise data support for subsequent adjustments to the welding current based on temperature.

[0208] In this embodiment, the real-time welding zone temperature refers to the actual temperature value of the area being welded on the part after spot welding is completed during the arc welding process. This temperature data is acquired in real time by a thermocouple array, and its value changes continuously, reflecting the real-time transfer and distribution of heat during the welding process, playing a crucial role in controlling welding quality. For example, after welding begins, the real-time welding zone temperature rises rapidly with the action of the welding arc, maintaining a certain fluctuation during the welding process.

[0209] In this embodiment, the welding current is dynamically adjusted based on the real-time temperature of the welding area on the part after spot welding. This is achieved by automatically adjusting the current during arc welding using the real-time temperature information of the welding area monitored by a thermocouple array. Because different welding temperatures require corresponding welding currents to ensure good welding results, when the real-time welding area temperature is too high, the welding current is appropriately reduced to avoid problems such as excessively wide welds and metal spatter caused by overheating; when the temperature is too low, the welding current is increased to ensure good weld fusion and prevent defects such as incomplete penetration. Through this dynamic adjustment, adaptive control of the welding process is achieved, improving the stability and reliability of welding quality.

[0210] The beneficial effects of the above technical solutions are as follows: The pneumatic clamping mechanism in the arc welding fixture can flexibly adjust the clamping force within a preset range via a proportional valve. This design allows for precise control of the clamping force based on the specifications and materials of the spot-welded parts. This ensures that the parts are firmly clamped during the arc welding process, avoiding welding position deviations caused by unstable clamping, and also prevents deformation of the parts due to excessive clamping force. This effectively guarantees the stability and precision of the parts during arc welding, thereby improving welding quality. Simultaneously, the thermocouple array monitors the real-time temperature of the welding area on the spot-welded parts and dynamically adjusts the welding current based on this real-time temperature. This function achieves adaptive control during the welding process, as different welding temperatures require matching welding currents to ensure good welding results. When the temperature is too high or too low, the system can adjust the welding current in a timely manner, avoiding welding defects such as welds that are too wide or too narrow, or incomplete penetration. This further improves the stability and reliability of welding quality, reduces the defect rate caused by temperature and current mismatch, increases production efficiency, and provides more precise and stable processing conditions for the arc welding process of automotive rear longitudinal beam assemblies.

[0211] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A welding system for automotive rear longitudinal beam assemblies based on 3D vision guidance and multi-fixture collaboration, characterized in that, include: The incoming material transportation module is used to transport the automotive rear longitudinal beam from the loading area to the welding production line's bench garage position using an automated guided vehicle. The grasping trajectory correction module is used to automatically correct the robot's grasping trajectory based on the spatial position and attitude data of the incoming automotive rear longitudinal beam, thereby obtaining a high-precision corrected grasping trajectory for the robot. The incoming material gripping module is used to control the robot based on high-precision correction gripping trajectory to grip the incoming automotive rear longitudinal beam and place it on the spot welding fixture; The spot welding operation moving module is used to move the spot welding fixture and the automotive rear longitudinal beam to the welding position based on the servo slide and complete the spot welding operation. The arc welding operation moving module is used to control the robot to grab the spot-welded parts after spot welding and place them on the arc welding fixture located on the positioner to complete the arc welding operation. Also includes: The fixture contact pressure detection module is used to detect the real-time contact pressure distribution data between the incoming automotive rear longitudinal beam and the fixture based on the pressure sensor array inside the spot welding fixture. The fixture support posture adjustment module is used to adjust the support posture of the fixture based on real-time contact pressure distribution data between the incoming automotive rear longitudinal beam and the fixture, including: The sliding trend prediction submodule is used to analyze the sliding trend of the automotive rear longitudinal beam material in the fixture based on real-time contact pressure distribution data between the incoming material and the fixture. This includes: The normal pressure calculation unit is used to calculate the total normal pressure in the contact area between the automotive rear longitudinal beam material and the fixture based on the real-time contact pressure distribution data between the automotive rear longitudinal beam material and the fixture. The resultant force analysis unit is used to calculate the frictional force on the rear longitudinal beam material in the fixture based on the total normal pressure in the contact area between the rear longitudinal beam material and the fixture, and to determine the resultant force on the rear longitudinal beam material in the fixture in combination with other external forces on the rear longitudinal beam material. The maximum static friction analysis unit is used to calculate the maximum static friction between the automotive rear longitudinal beam material and the fixture based on the total normal pressure of the contact area between the material and the fixture. The sliding trend analysis unit is used to analyze the sliding trend of the automotive rear longitudinal beam material in the fixture based on the tangential component of the resultant force and the maximum static friction force on the material. The support posture adjustment submodule is used to adjust the support posture of the fixture based on the sliding trend of the automotive rear longitudinal beam material in the fixture.

2. The automotive rear longitudinal beam assembly welding system based on 3D vision guidance and multi-fixture collaboration as described in claim 1, characterized in that, The incoming material transportation module includes: The initial path generation submodule is used to generate the initial transport path between the loading area and the welding production line's machine parking space; The obstacle avoidance mechanism triggering submodule is used to control the automated guided vehicle to pick up the automotive rear longitudinal beam material from the loading area and transport it to the welding production line's bench station according to the initial transport path. At the same time, based on the millimeter-wave radar on the automated guided vehicle, obstacles within a preset range centered on the automated guided vehicle are detected in real time. When the millimeter-wave radar detects an obstacle, the obstacle avoidance strategy is triggered to generate a new obstacle avoidance transport path, and the automotive rear longitudinal beam material is continued to be transported to the welding production line's bench station according to the new obstacle avoidance transport path. The obstacle avoidance strategies include: Multiple initial obstacle avoidance transport paths are generated based on the location of obstacles detected by millimeter-wave radar; The path cost of each initial obstacle avoidance transportation path is calculated based on the path cost function: In the formula, The path cost of a single initial obstacle avoidance transport path. This represents the total number of locations in a single initial obstacle avoidance transport path. For the first obstacle avoidance transport path in a single initial obstacle avoidance transport path A location point, For the first obstacle avoidance transport path in a single initial obstacle avoidance transport path A location point, For the first obstacle avoidance transport path in a single initial obstacle avoidance transport path The location point and the first Euclidean distance between points This represents the total number of obstacles detected by the millimeter-wave radar. For the first obstacle avoidance transport path in a single initial obstacle avoidance transport path The x-coordinates of each location point For the first obstacle avoidance transport path in a single initial obstacle avoidance transport path The y-coordinate values ​​of each location point For the first The x-coordinate value of each obstacle For the first The vertical coordinate value of each obstacle; The initial obstacle avoidance transport path with the minimum path cost among all initial obstacle avoidance transport paths is taken as the latest obstacle avoidance transport path.

3. The automotive rear longitudinal beam assembly welding system based on 3D vision guidance and multi-fixture collaboration as described in claim 1, characterized in that, The capture trajectory correction module includes: The incoming material pose scanning submodule is used to scan and identify the spatial position and pose data of the incoming automotive rear longitudinal beam using a depth camera. The spatial pose matching submodule is used to match the spatial position and attitude data of the incoming automotive rear longitudinal beam with the preset 3D model of the incoming material based on the point cloud registration algorithm, and generate a spatial pose deviation matrix. The grasping trajectory correction submodule is used to automatically correct the robot's grasping trajectory based on the spatial pose deviation matrix, thereby obtaining a high-precision corrected grasping trajectory for the robot.

4. The automotive rear longitudinal beam assembly welding system based on 3D vision guidance and multi-fixture collaboration according to claim 3, characterized in that, The trajectory correction submodule includes: The initial correction unit is used to automatically correct the robot's grasping trajectory based on the spatial pose deviation matrix, and obtain the robot's initial corrected grasping trajectory. The dynamic model generation unit is used to generate the robot's state vector at each time step based on the robot's dynamic vibration offset and the rate of change of the dynamic vibration offset at each time step, and to generate the robot's dynamic model by combining the state transition matrix, the control input matrix, and the process noise vector. The measurement model generation unit is used to generate a measurement model of the robot based on the robot's dynamic model, measurement matrix, and measurement noise vector; The vibration offset estimation unit is used to estimate the dynamic vibration offset of the robot based on the robot's dynamic model, measurement model, and Kalman filter. The reverse compensation unit is used to perform reverse displacement compensation on the robot's initial corrected grasping trajectory based on the robot's dynamic vibration offset, so as to obtain the robot's high-precision corrected grasping trajectory.

5. The automotive rear longitudinal beam assembly welding system based on 3D vision guidance and multi-fixture collaboration according to claim 4, characterized in that, The initial calibration unit includes: The position correction subunit is used to correct the position of each position point in the grasping trajectory of the robot based on the translation vector in the spatial pose deviation matrix, so as to obtain the position correction grasping trajectory of the robot. The attitude correction subunit is used to adjust the attitude of each position point in the robot's position correction grasping trajectory based on the rotation matrix in the spatial pose deviation matrix, so as to obtain the robot's initial correction grasping trajectory.

6. The automotive rear longitudinal beam assembly welding system based on 3D vision guidance and multi-fixture collaboration according to claim 1, characterized in that, The support pose adjustment submodule includes: An optimization objective function construction unit is used to construct an objective function with the goal of minimizing the sliding trend when the sliding trend of the automotive rear longitudinal beam material in the fixture exceeds the sliding trend threshold. The pose optimization unit is used to solve for the optimal support pose vector of the fixture based on the objective function and the upper and lower limits of the support pose vector of the fixture, and to control the support pose of the fixture based on the optimal support pose vector of the fixture.

7. The automotive rear longitudinal beam assembly welding system based on 3D vision guidance and multi-fixture collaboration according to claim 1, characterized in that, Arc welding fixtures include: The clamping force is adjustable within a preset range via a pneumatic clamping mechanism and thermocouple array controlled by a proportional valve. Among them, the thermocouple array monitors the real-time temperature of the welding area on the part after spot welding, and dynamically adjusts the welding current based on the real-time temperature of the welding area on the part after spot welding.

Images