A multi-welding robot cooperative control method and system

Abstract

A kind of multi-welding robot collaborative control method and system, the method includes: first step, input welding task and split, pre-allocate welding subtask in combination with robot initial state;Second step, simulate robot operation in virtual environment and detect operation conflict, no conflict then optimize parameters, conflict then adjust to no conflict;Third step, allocate subtask according to rehearsal result, plan dedicated path and coordinate timing;Fourth step, according to dedicated motion path and timing, drive welding robot actual operation in combination with optimized parameters;Fifth step, real-time tracking feedback path execution and weld condition, correct the motion path of robot and welding parameter, realize collaborative control;The system includes multiple welding robots, motion control module, man-machine interaction and monitoring module, task planning module, digital twin and optimization module, path planning and collision avoidance module, sensor and state perception module.Therefore, the collaborative control effect of the design multiple welding robots is good.

Description

Technical Field

[0001] This invention relates to a collaborative control method and system, belonging to the field of robot control system technology, and particularly to a collaborative control method and system for multiple welding robots. Background Technology

[0002] In modern manufacturing, welding, as a key processing technology, is widely used in many fields such as automobile manufacturing, shipbuilding, and machining. With the expansion of industrial production scale and the increase in product complexity, the requirements for welding efficiency and quality are becoming increasingly stringent. Traditional single welding robots, when faced with large and complex workpieces or batch production tasks, are no longer sufficient to meet the current development needs of the manufacturing industry due to their long processing cycles and low efficiency. To improve welding production efficiency, multiple welding robots often work collaboratively. However, existing multi-robot collaborative control methods and systems suffer from several drawbacks. First, they lack specific task decomposition, leading to poor compatibility between sub-tasks and robots, uneven loads on each robot, and difficulty in leveraging the efficiency advantages of multi-robot collaboration. Second, their reliance on preset rules can easily cause path interference and timing conflicts, requiring downtime for debugging and affecting welding continuity. Furthermore, the lack of real-time tracking feedback and dynamic correction can result in poor weld quality consistency, leading to poor collaborative control effects for multiple welding robots.

[0003] Chinese patent application No. 202111392307.7, filed on November 19, 2021, discloses a multi-robot control method, electronic device, and storage medium. It breaks down a task to be performed into multiple sub-tasks and generates task allocation data for each sub-task based on the attribute information of available robots. This includes: sub-task content and the first robot executing the sub-task; it generates movement path planning data for each first robot based on the task allocation data and a 3D semantic map of the environment; it simulates and evaluates the movement path planning data of each first robot; when the simulation results meet preset requirements, it sends the movement path execution data generated based on the movement path planning data to the corresponding first robot to control the first robot to execute the corresponding sub-task; it monitors the state data of the first robot during the execution of the sub-task and maintains or adjusts the movement path execution data of the first robot based on the state data. Although this patent can enable multi-robot collaboration to complete complex tasks, it still has the following shortcomings:

[0004] Because the design did not include task allocation, path planning, simulation evaluation, and real-time adjustment mechanisms specifically designed for the welding process, the collaborative control effect of multiple welding robots remained poor.

[0005] The information disclosed in this background section is intended only to enhance the understanding of the overall background of this patent application and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to overcome the defects and problems of poor collaborative control of multiple welding robots in the existing technology, and to provide a collaborative control method and system for multiple welding robots with better collaborative control effect.

[0007] To achieve the above objectives, the technical solution of the present invention is: a method and system for collaborative control of multiple welding robots, wherein the control method includes the following steps:

[0008] Step 1: First, input the welding task, break it down into multiple welding sub-tasks, and then pre-assign the welding sub-tasks to the corresponding welding robots based on the real-time status of each welding robot.

[0009] Step 2: Based on the pre-assigned welding sub-tasks of the welding robots, simulate the operation of multiple welding robots in a virtual environment to detect whether there are any operational conflicts among the multiple welding robots. If there are no operational conflicts, optimize the welding parameters of the multiple welding robots to form a conflict-free pre-simulation result. If there are operational conflicts, readjust the splitting and allocation of welding sub-tasks according to the conflict situation, and then simulate the operation of multiple welding robots in the virtual environment again until there are no operational conflicts. Then optimize the welding parameters of the multiple welding robots to form a conflict-free pre-simulation result.

[0010] Step 3: Optimize the welding task breakdown in Step 1 according to the conflict-free pre-simulation results described in Step 2, resulting in multiple optimized welding sub-tasks. Then, assign the optimized welding sub-tasks to the corresponding welding robots. Based on the optimized welding sub-tasks assigned to each welding robot, plan a dedicated motion path for each robot. At the same time, coordinate the path timing of multiple welding robots.

[0011] Step 4: Drive the welding robot to perform actual operation according to the above-mentioned dedicated motion path and path sequence, and in combination with the optimized welding parameters in the above-mentioned conflict-free pre-simulation results;

[0012] Step 5: During the actual operation of the welding robot, the execution of the dedicated motion path is tracked in real time, the actual completion of the weld is monitored, and feedback is provided. Then, based on the feedback of the execution of the dedicated motion path and the actual completion of the weld, the robot's motion path and welding parameters are corrected in real time, thereby realizing the collaborative control of multiple robots.

[0013] The control method further includes:

[0014] Step 6: During the actual operation of the welding robots, continuously monitor the operating status of each welding robot for any abnormalities. If an abnormality occurs, trigger a self-recovery operation to maintain the operation of each welding robot. If the self-recovery operation cannot be performed, trigger an alarm and wait for repair.

[0015] In the first step, the input welding task refers to inputting the actual welding task and welding object information;

[0016] In the first step, the step of splitting the welding task into multiple welding sub-tasks means: based on the geometric features in the input welding object information, using a graph segmentation algorithm, the input actual welding task is divided into multiple independent and continuous welding sub-tasks.

[0017] In the first step, the step of pre-allocating welding sub-tasks to the corresponding welding robots based on the real-time status of each welding robot means: based on the real-time status of each welding robot, solving the optimal allocation scheme through mixed integer programming, and pre-allocating the segmented welding sub-tasks to the corresponding welding robots.

[0018] In the second step, the simulation of the operation of multiple welding robots in a virtual environment based on the pre-assigned welding sub-tasks of the welding robots, and the detection of whether there are operational conflicts among the multiple welding robots, refers to: first, constructing a simulation environment consistent with the physical welding scene in the digital virtual space, importing the three-dimensional model of the welding object and the kinematic model of each welding robot, and integrating the finite element analysis module; then, assigning corresponding weld segment execution instructions to each robot according to the pre-assigned welding sub-tasks; then driving the virtual welding robot model to run according to the instructions, and simultaneously analyzing the real-time thermal deformation data of the welding object model through the finite element analysis module; at the same time, tracking the relative positional relationship between the virtual welding robot's motion trajectory and the thermally deformed welding object model in real time, and judging whether there is spatial interference between the robot and the welding object model, and between the robots themselves;

[0019] In the second step, the step of optimizing the welding parameters of multiple welding robots to form a conflict-free pre-simulation result if there is no operational conflict means that after confirming that there is no spatial interference between the robot and the welding object model, and between the robots, the welding parameters are optimized based on the simulated welding process data, and finally a conflict-free pre-simulation result is generated; the conflict-free pre-simulation result includes path suggestions, risk area markings, and optimized welding parameters;

[0020] In the second step, if there is a conflict, the welding sub-tasks are re-splitting and reassigned according to the conflict situation. Then, the operation of multiple welding robots is simulated again in the virtual environment until there is no conflict. This means that spatial interference is detected between the robot and the welding object model, or between the robot and the object. The conflict type is analyzed first, and the welding sub-tasks are re-splitting and reassigned according to the conflict type. After the adjustment is completed, the new sub-task allocation scheme is imported into the virtual environment, and the operation of multiple welding robots is simulated repeatedly until all conflicts are eliminated.

[0021] In the third step, optimizing the welding task breakdown of the first step according to the conflict-free pre-simulation results described in the second step to obtain multiple optimized welding sub-tasks means: adjusting the breakdown of the welding task in the first step according to the path suggestions and optimized welding parameters in the conflict-free pre-simulation results described in the second step to obtain multiple optimized welding sub-tasks.

[0022] In the third step, planning a dedicated motion path for each welding robot based on the optimized welding sub-task assigned to each welding robot means: planning a dedicated motion path for each robot using a fast expanding random tree algorithm based on the optimized welding sub-task assigned to each welding robot; the dedicated motion path includes a main motion path and several preliminary paths.

[0023] In the third step, coordinating the path sequence of multiple welding robots refers to: based on the dedicated motion path of each robot, synchronizing the position and running status of each welding robot in real time, and planning the order of path execution for each welding robot.

[0024] In the fourth step, driving the welding robot to actually run means: converting the dedicated motion path and optimized welding parameters planned in the third step into control instructions that the robot can execute, transmitting them to the corresponding welding robot, and then having the robot perform the actual welding operation according to the dedicated motion path and optimized welding parameters after receiving the control instructions.

[0025] In the fifth step, the real-time tracking of the execution of the dedicated motion path, monitoring the actual completion of the weld, and providing feedback refers to: capturing the actual motion path of the welding robot in real time, comparing it with the dedicated motion path in the third step, calculating the deviation between the actual motion path and the planned dedicated motion path, and obtaining the path execution deviation; at the same time, collecting the position information of the completed weld, the temperature distribution of the welding area, and the contact force data between the welding torch and the workpiece in real time, analyzing the forming quality of the completed weld, identifying the defects of the completed weld, and then providing feedback on the path execution deviation, weld quality data, and defect status;

[0026] In the fifth step, the real-time correction of the robot's motion path and welding parameters based on the feedback of the execution status of the dedicated motion path and the actual completion status of the weld refers to: dynamically adjusting the robot's motion trajectory through a sliding mode control algorithm based on the feedback of path execution deviation, so that the actual motion path of the welding robot returns to the dedicated motion path; at the same time, combining weld quality data and defect conditions, optimizing welding parameters through fuzzy PID control, and adjusting the welding torch's working posture to ensure that the welding process meets quality requirements; during the correction process, the path timing of multiple robots is synchronously associated to avoid new collaborative conflicts caused by the adjustment of a single robot, thereby achieving dynamic adaptation of motion path and welding parameters.

[0027] In the sixth step, continuously monitoring whether the operating status of each welding robot is abnormal means continuously monitoring whether the operating status of each welding robot, including motor current, temperature, and vibration signals, is abnormal.

[0028] The system includes multiple welding robots, a motion control module, a human-computer interaction and monitoring module, a task planning module, a digital twin and optimization module, a path planning and collision avoidance module, and a sensor and state perception module.

[0029] The human-computer interaction and monitoring module, task planning module, digital twin and optimization module, path planning and collision avoidance module, and sensor and state perception module interact with the motion control module, and the motion control module interacts with multiple welding robots.

[0030] The system also includes a fault diagnosis and self-recovery module; the fault diagnosis and self-recovery module interacts with the motion control module.

[0031] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0032] 1. This invention discloses a method and system for collaborative control of multiple welding robots. The method includes: First, inputting a welding task, breaking it down into sub-tasks, and pre-allocating them based on the real-time status of the robots; Second, simulating the operation in a virtual environment based on the pre-allocated sub-tasks to detect conflicts. If there are no conflicts, the parameters are optimized to form a result; if there are conflicts, the parameters are adjusted and re-simulated until there are no conflicts before optimization; Third, optimizing the sub-tasks according to the results to obtain optimized sub-tasks, and allocating the optimized sub-tasks, planning dedicated paths, and coordinating timing; Fourth, driving the robots to run according to the dedicated paths, timing, and optimized parameters; Fifth, real-time tracking, monitoring, and feedback to correct the paths and parameters, thereby achieving collaborative control. In application, a specific welding task is first input, the task is broken down into multiple welding sub-tasks, and then the welding sub-tasks are pre-allocated to the corresponding robots based on the real-time status of each welding robot. Based on the pre-allocation results, the robots are then controlled in a virtual environment. The invention simulates the operation of multiple robots in a simulated environment, detecting any operational conflicts. If no conflicts occur, welding parameters are optimized to generate a conflict-free pre-simulation result. If conflicts exist, the sub-tasks are re-divided and reassigned based on the conflict situation, and the simulation is repeated until no conflicts occur. Welding parameters are then optimized again to obtain a pre-simulation result. The welding task is further optimized based on the conflict-free pre-simulation result, resulting in optimized welding sub-tasks. These optimized sub-tasks are assigned, and a dedicated motion path is planned for each robot. Simultaneously, the path timing of multiple robots is coordinated. Subsequently, based on the dedicated motion path, timing, and optimized welding parameters, the robots are driven to actually run. During operation, the path execution is tracked in real time, the actual weld completion status is monitored, and this information is fed back. Based on the feedback, the robot's motion path and welding parameters are corrected in real time, thereby achieving collaborative control of multiple welding robots. The advantages of this invention also include:

[0033] First, by pre-allocating sub-tasks based on the robot's real-time status, the task is matched with the robot's capabilities, which solves the problems of lack of specificity and poor adaptability in task decomposition in existing technologies, and improves the basic efficiency of collaborative operations and the accuracy of welding task execution.

[0034] Secondly, by simulating and detecting conflicts in a virtual environment and optimizing parameters, and adjusting the allocation when conflicts occur, the simulation is repeated until there are no conflicts. Using virtual pre-simulation to optimize task allocation in advance reduces actual welding conflicts and on-site debugging time, ensuring welding continuity.

[0035] Thirdly, by planning a dedicated path and coordinating the timing for each robot, we ensure that multiple robots do not interfere with each other and execute precisely along a collision-free path. This solves the problems of existing technologies relying on preset rules, which are prone to path interference and timing conflicts. It improves the overall coordination of welding rhythm and effectively shortens the welding cycle.

[0036] Fourthly, real-time tracking and feedback of path execution and weld completion status, and dynamic correction of path and welding parameters can promptly handle sudden deviations, avoiding the poor weld quality consistency caused by the lack of real-time tracking and feedback and dynamic correction in existing technologies. This ensures weld quality consistency, reduces rework rate, and improves collaborative effect.

[0037] Therefore, this invention can not only coordinate the control of multiple welding robots, but also achieve good results in the coordinated control of multiple welding robots.

[0038] 2. In the collaborative control method and system for multiple welding robots of the present invention, the control method further includes a sixth step: during the actual operation of the welding robots, the operating status of each welding robot is continuously monitored for any abnormalities. If an abnormality occurs, a self-recovery operation is triggered to maintain the operation of each welding robot. If the self-recovery operation cannot be performed, an alarm is triggered and the system waits for repair. In application, by continuously monitoring the operating status of each welding robot in real time, including motor current, temperature, and vibration signals, when an abnormality is detected, a self-recovery operation is triggered to maintain the continuity of operation of multiple welding robots. If the self-recovery operation cannot be performed, an alarm is triggered to remind personnel to perform repairs, thereby reducing downtime caused by malfunctions and further ensuring the stability and efficiency of collaborative operation of multiple welding robots. Therefore, the present invention not only has a good collaborative control effect for multiple welding robots but also ensures the stability of collaborative operation of multiple welding robots.

[0039] 3. This invention provides a collaborative control method and system for multiple welding robots. The system includes multiple welding robots, a motion control module, a human-machine interaction and monitoring module, a task planning module, a digital twin and optimization module, a path planning and collision avoidance module, and a sensor and state perception module. In application, the multi-sensor fusion unit in the sensor and state perception module collects weld position, temperature, and force data in real time, providing comprehensive state information for the welding process and providing a basis for subsequent control and decision-making. The AI ​​weld tracking unit in the sensor and state perception module tracks and corrects deviations, and the collected information is transmitted to the motion control module to ensure weld accuracy and improve welding quality. Therefore, this invention not only ensures the stability of collaborative operation of multiple welding robots but also improves welding quality. Attached Figure Description

[0040] Figure 1 This is a flowchart of the method of the present invention.

[0041] Figure 2 This is a schematic diagram of the system of the present invention. Detailed Implementation

[0042] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0043] See Figure 1 — Figure 2 A method and system for collaborative control of multiple welding robots, the control method comprising the following steps:

[0044] Step 1: First, input the welding task, break it down into multiple welding sub-tasks, and then pre-assign the welding sub-tasks to the corresponding welding robots based on the real-time status of each welding robot.

[0045] Step 2: Based on the pre-assigned welding sub-tasks of the welding robots, simulate the operation of multiple welding robots in a virtual environment to detect whether there are any operational conflicts among the multiple welding robots. If there are no operational conflicts, optimize the welding parameters of the multiple welding robots to form a conflict-free pre-simulation result. If there are operational conflicts, readjust the splitting and allocation of welding sub-tasks according to the conflict situation, and then simulate the operation of multiple welding robots in the virtual environment again until there are no operational conflicts. Then optimize the welding parameters of the multiple welding robots to form a conflict-free pre-simulation result.

[0046] Step 3: Optimize the welding task breakdown in Step 1 according to the conflict-free pre-simulation results described in Step 2, resulting in multiple optimized welding sub-tasks. Then, assign the optimized welding sub-tasks to the corresponding welding robots. Based on the optimized welding sub-tasks assigned to each welding robot, plan a dedicated motion path for each robot. At the same time, coordinate the path timing of multiple welding robots.

[0047] Step 4: Drive the welding robot to perform actual operation according to the above-mentioned dedicated motion path and path sequence, and in combination with the optimized welding parameters in the above-mentioned conflict-free pre-simulation results;

[0048] Step 5: During the actual operation of the welding robot, the execution of the dedicated motion path is tracked in real time, the actual completion of the weld is monitored, and feedback is provided. Then, based on the feedback of the execution of the dedicated motion path and the actual completion of the weld, the robot's motion path and welding parameters are corrected in real time, thereby realizing the collaborative control of multiple robots.

[0049] The control method further includes:

[0050] Step 6: During the actual operation of the welding robots, continuously monitor the operating status of each welding robot for any abnormalities. If an abnormality occurs, trigger a self-recovery operation to maintain the operation of each welding robot. If the self-recovery operation cannot be performed, trigger an alarm and wait for repair.

[0051] In the first step, the input welding task refers to inputting the actual welding task and welding object information;

[0052] In the first step, the step of splitting the welding task into multiple welding sub-tasks means: based on the geometric features in the input welding object information, using a graph segmentation algorithm, the input actual welding task is divided into multiple independent and continuous welding sub-tasks.

[0053] In the first step, the step of pre-allocating welding sub-tasks to the corresponding welding robots based on the real-time status of each welding robot means: based on the real-time status of each welding robot, solving the optimal allocation scheme through mixed integer programming, and pre-allocating the segmented welding sub-tasks to the corresponding welding robots.

[0054] In the second step, the simulation of the operation of multiple welding robots in a virtual environment based on the pre-assigned welding sub-tasks of the welding robots, and the detection of whether there are operational conflicts among the multiple welding robots, refers to: first, constructing a simulation environment consistent with the physical welding scene in the digital virtual space, importing the three-dimensional model of the welding object and the kinematic model of each welding robot, and integrating the finite element analysis module; then, assigning corresponding weld segment execution instructions to each robot according to the pre-assigned welding sub-tasks; then driving the virtual welding robot model to run according to the instructions, and simultaneously analyzing the real-time thermal deformation data of the welding object model through the finite element analysis module; at the same time, tracking the relative positional relationship between the virtual welding robot's motion trajectory and the thermally deformed welding object model in real time, and judging whether there is spatial interference between the robot and the welding object model, and between the robots themselves;

[0055] In the second step, the step of optimizing the welding parameters of multiple welding robots to form a conflict-free pre-simulation result if there is no operational conflict means that after confirming that there is no spatial interference between the robot and the welding object model, and between the robots, the welding parameters are optimized based on the simulated welding process data, and finally a conflict-free pre-simulation result is generated; the conflict-free pre-simulation result includes path suggestions, risk area markings, and optimized welding parameters;

[0056] In the second step, if there is a conflict, the welding sub-tasks are re-splitting and reassigned according to the conflict situation. Then, the operation of multiple welding robots is simulated again in the virtual environment until there is no conflict. This means that spatial interference is detected between the robot and the welding object model, or between the robot and the object. The conflict type is analyzed first, and the welding sub-tasks are re-splitting and reassigned according to the conflict type. After the adjustment is completed, the new sub-task allocation scheme is imported into the virtual environment, and the operation of multiple welding robots is simulated repeatedly until all conflicts are eliminated.

[0057] In the third step, optimizing the welding task breakdown of the first step according to the conflict-free pre-simulation results described in the second step to obtain multiple optimized welding sub-tasks means: adjusting the breakdown of the welding task in the first step according to the path suggestions and optimized welding parameters in the conflict-free pre-simulation results described in the second step to obtain multiple optimized welding sub-tasks.

[0058] In the third step, planning a dedicated motion path for each welding robot based on the optimized welding sub-task assigned to each welding robot means: planning a dedicated motion path for each robot using a fast expanding random tree algorithm based on the optimized welding sub-task assigned to each welding robot; the dedicated motion path includes a main motion path and several preliminary paths.

[0059] In the third step, coordinating the path sequence of multiple welding robots refers to: based on the dedicated motion path of each robot, synchronizing the position and running status of each welding robot in real time, and planning the order of path execution for each welding robot.

[0060] In the fourth step, driving the welding robot to actually run means: converting the dedicated motion path and optimized welding parameters planned in the third step into control instructions that the robot can execute, transmitting them to the corresponding welding robot, and then having the robot perform the actual welding operation according to the dedicated motion path and optimized welding parameters after receiving the control instructions.

[0061] In the fifth step, the real-time tracking of the execution of the dedicated motion path, monitoring the actual completion of the weld, and providing feedback refers to: capturing the actual motion path of the welding robot in real time, comparing it with the dedicated motion path in the third step, calculating the deviation between the actual motion path and the planned dedicated motion path, and obtaining the path execution deviation; at the same time, collecting the position information of the completed weld, the temperature distribution of the welding area, and the contact force data between the welding torch and the workpiece in real time, analyzing the forming quality of the completed weld, identifying the defects of the completed weld, and then providing feedback on the path execution deviation, weld quality data, and defect status;

[0062] In the fifth step, the real-time correction of the robot's motion path and welding parameters based on the feedback of the execution status of the dedicated motion path and the actual completion status of the weld refers to: dynamically adjusting the robot's motion trajectory through a sliding mode control algorithm based on the feedback of path execution deviation, so that the actual motion path of the welding robot returns to the dedicated motion path; at the same time, combining weld quality data and defect conditions, optimizing welding parameters through fuzzy PID control, and adjusting the welding torch's working posture to ensure that the welding process meets quality requirements; during the correction process, the path timing of multiple robots is synchronously associated to avoid new collaborative conflicts caused by the adjustment of a single robot, thereby achieving dynamic adaptation of motion path and welding parameters.

[0063] In the sixth step, continuously monitoring whether the operating status of each welding robot is abnormal means continuously monitoring whether the operating status of each welding robot, including motor current, temperature, and vibration signals, is abnormal.

[0064] The system includes multiple welding robots, a motion control module, a human-computer interaction and monitoring module, a task planning module, a digital twin and optimization module, a path planning and collision avoidance module, and a sensor and state perception module.

[0065] The human-computer interaction and monitoring module, task planning module, digital twin and optimization module, path planning and collision avoidance module, and sensor and state perception module interact with the motion control module, and the motion control module interacts with multiple welding robots.

[0066] The system also includes a fault diagnosis and self-recovery module; the fault diagnosis and self-recovery module interacts with the motion control module.

[0067] The following are supplementary descriptions of the present invention:

[0068] The AI ​​weld seam tracking unit of the present invention preferably tracks and corrects deviations based on data collected by a multi-sensor fusion unit.

[0069] Example 1:

[0070] See Figure 1 — Figure 2 A method and system for collaborative control of multiple welding robots, the control method comprising the following steps:

[0071] Step 1: First, input the welding task, break it down into multiple welding sub-tasks, and then pre-assign the welding sub-tasks to the corresponding welding robots based on the real-time status of each welding robot.

[0072] Step 2: Based on the pre-assigned welding sub-tasks of the welding robots, simulate the operation of multiple welding robots in a virtual environment to detect whether there are any operational conflicts among the multiple welding robots. If there are no operational conflicts, optimize the welding parameters of the multiple welding robots to form a conflict-free pre-simulation result. If there are operational conflicts, readjust the splitting and allocation of welding sub-tasks according to the conflict situation, and then simulate the operation of multiple welding robots in the virtual environment again until there are no operational conflicts. Then optimize the welding parameters of the multiple welding robots to form a conflict-free pre-simulation result.

[0073] Step 3: Optimize the welding task breakdown in Step 1 according to the conflict-free pre-simulation results described in Step 2, resulting in multiple optimized welding sub-tasks. Then, assign the optimized welding sub-tasks to the corresponding welding robots. Based on the optimized welding sub-tasks assigned to each welding robot, plan a dedicated motion path for each robot. At the same time, coordinate the path timing of multiple welding robots.

[0074] Step 4: Drive the welding robot to perform actual operation according to the above-mentioned dedicated motion path and path sequence, and in combination with the optimized welding parameters in the above-mentioned conflict-free pre-simulation results;

[0075] Step 5: During the actual operation of the welding robot, the execution of the dedicated motion path is tracked in real time, the actual completion of the weld is monitored, and feedback is provided. Then, based on the feedback of the execution of the dedicated motion path and the actual completion of the weld, the robot's motion path and welding parameters are corrected in real time, thereby realizing the collaborative control of multiple robots.

[0076] Example 2:

[0077] The basic content is the same as in Embodiment 1, except that the control method further includes:

[0078] Step 6: During the actual operation of the welding robots, continuously monitor the operating status of each welding robot for any abnormalities. If an abnormality occurs, trigger a self-recovery operation to maintain the operation of each welding robot. If the self-recovery operation cannot be performed, trigger an alarm and wait for repair.

[0079] In application, in step six, the continuous monitoring of the operating status of each welding robot, triggering self-recovery operations, and triggering alarms are achieved by the fault diagnosis and self-recovery module in the system. The fault diagnosis and self-recovery module includes an anomaly detection unit and a self-recovery strategy unit.

[0080] The anomaly detection unit monitors operating status parameters such as motor current, temperature, and vibration signals in real time. When an anomaly is detected, the self-recovery strategy unit performs self-recovery operations based on the anomaly type: If the anomaly is a path conflict or a local fault, the multi-machine collaborative collision avoidance unit of the path planning and collision avoidance module is invoked to switch to a pre-generated path by the single-machine path planning unit. At the same time, the motion timing of the robot and associated robots is coordinated to avoid the fault area and ensure that the multi-machine path is collision-free. If the anomaly is related to welding parameters (such as abnormal contact force), the force control unit of the motion control module recalibrates the welding torch contact force based on an impedance control algorithm. Simultaneously, the dynamic parameter adjustment unit uses fuzzy PID control to fine-tune parameters such as welding current and speed. To maintain welding continuity, if a non-fatal failure occurs in the core component of a single robot, the system redundancy mechanism is triggered. The task allocation unit of the task planning module recalculates the sub-task allocation scheme using a mixed integer programming algorithm, dynamically allocating unfinished tasks to other idle or low-load robots to ensure the overall task progress. If the above self-recovery operations fail to resolve the anomaly (such as complete failure of the core component), an alarm is triggered through the fault diagnosis and self-recovery module. The motion control module transmits the anomaly information to the human-machine interaction and monitoring module. The visualization interface unit in the human-machine interaction and monitoring module displays the alarm information and fault location, awaiting manual repair, reducing downtime caused by faults and improving the reliability and stability of the system.

[0081] Example 3:

[0082] The basic content is the same as in Example 1, except that: in the first step, the input welding task refers to the input of the actual welding task and welding object information; in the first step, the step of splitting the welding task into multiple welding sub-tasks refers to: based on the geometric features in the input welding object information, using a graph segmentation algorithm, dividing the input actual welding task into multiple independent and continuous welding sub-tasks; in the first step, the step of pre-allocating the welding sub-tasks to the corresponding welding robots based on the real-time status of each welding robot refers to: based on the real-time status of each welding robot, solving the optimal allocation scheme through mixed integer programming, and pre-allocating the segmented welding sub-tasks to the corresponding welding robots.

[0083] In application, the first step is achieved collaboratively by the task planning module and the human-computer interaction and monitoring module. The operator manually uploads the actual welding task and welding object information through the remote control unit of the human-computer interaction and monitoring module, completing the input of task instructions. This task and information are first transmitted to the motion control module, and then synchronized to the task planning module. The task decomposition unit of the task planning module, according to the graph segmentation algorithm, divides the received actual welding task into multiple sub-tasks based on the correlation weights between weld segments (calculated according to curvature and length), breaking down the geometric features (such as straight line segments and arc segments) in the input welding object information into multiple sub-tasks, providing a basis for subsequent task allocation. The graph segmentation algorithm calculation formula is as follows:

[0084] ;

[0085] Among them, w ij Indicates the correlation weight between weld segments (calculated based on curvature and length);

[0086] Next, the task allocation unit of the task planning module combines the real-time status of multiple robots (including the current position, idle status, load, assigned tasks, and equipment operating parameters of each robot) and uses Mixed Integer Programming (MIP) to pre-allocate welding sub-tasks to the corresponding welding robots, taking into account execution costs (including path length and energy consumption) and the robot capability matrix, in order to balance the load and efficiency of each robot. The MIP calculation formula is as follows:

[0087] ;

[0088] Where, x i Indicates whether robot i is assigned a task (0 / 1 variable), c i Indicates the execution cost (including path length and energy consumption), a ij Represents the robot's capability matrix;

[0089] During this process, the visualization interface unit of the human-computer interaction and monitoring module displays the robot's real-time status (including pose, welding progress and alarm information) in 3D form, which makes it easy for operators to intuitively grasp the working status. If intervention is required, manual intervention can be carried out through the remote control unit to ensure the safety of welding operations and avoid welding quality problems or equipment damage due to unexpected situations.

[0090] Example 4:

[0091] The basic content is the same as in Example 1, except that in the second step, simulating the operation of multiple welding robots in a virtual environment based on the pre-assigned welding sub-tasks of the welding robots and detecting whether there is an operational conflict among the multiple welding robots refers to: first, constructing a simulation environment consistent with the physical welding scene in the digital virtual space, importing the three-dimensional model of the welding object and the kinematic model of each welding robot, and integrating the finite element analysis module; then, assigning corresponding weld segment execution instructions to each robot according to the pre-assigned welding sub-tasks; then driving the virtual welding robot model to run according to the instructions, and simultaneously analyzing the real-time thermal deformation data of the welding object model through the finite element analysis module; at the same time, tracking the relative positional relationship between the virtual welding robot's motion trajectory and the thermally deformed welding object model in real time, and judging whether there is spatial interference between the robot and the welding object model, and between the robots; in the second step, if there is no If there is a conflict, the welding parameters of multiple welding robots are optimized to form a conflict-free pre-simulation result. This means that after confirming that there is no spatial interference between the robot and the welding object model, and between the robots themselves, the welding parameters are optimized based on the simulated welding process data to generate a conflict-free pre-simulation result. The conflict-free pre-simulation result includes path suggestions, risk area markings, and optimized welding parameters. In the second step, if there is a conflict, the welding sub-tasks are re-splitting and reassigned according to the conflict situation, and then the operation of multiple welding robots is simulated again in the virtual environment until there is no conflict. This means that if spatial interference is detected between the robot and the welding object model, and between the robots themselves, the conflict type is analyzed first, and the welding sub-tasks are re-splitting and reassigned according to the conflict type. After the adjustment is completed, the new sub-task allocation scheme is imported into the virtual environment, and the operation of multiple welding robots is simulated repeatedly until all conflicts are eliminated.

[0092] In application, after the task planning module completes task decomposition and allocation, it synchronizes the allocation results to the digital twin and optimization module via the motion control module. The digital twin and optimization module includes a virtual welding pre-simulation unit and a parameter optimization unit. The virtual welding pre-simulation unit calls the finite element analysis (FEA) function and optimizes the parameters according to the formula... Perform calculations and analysis, among which, Represents the stiffness matrix. Indicates nodal displacement. Indicates thermal load (calculated using laser power and moving speed);

[0093] The virtual welding pre-simulation unit constructs a simulation environment consistent with the physical scene in a digital virtual space, imports the 3D model of the welding object and the kinematic model of the robot, and assigns the robot the execution instructions for the weld segment according to the pre-assigned welding sub-tasks. While driving the virtual robot to run, the finite element analysis module obtains the real-time thermal deformation data of the welding object, tracks the relative position of the virtual robot's motion trajectory and the weld object after thermal deformation, and judges the spatial interference. If a conflict is detected (such as the motion trajectory of a robot's welding torch intersecting with the arm motion range of another robot), the conflict type (spatial occupancy conflict or temporal conflict) is automatically analyzed, and the sub-task splitting method (such as assigning the entire weld segment of the conflict area to one robot) or allocation scheme (such as swapping the working areas of the two robots) is readjusted. The new scheme is then imported and the simulation is repeated until all conflicts are eliminated.

[0094] After confirming that there are no conflicts, the parameter optimization unit optimizes the welding parameters (including welding current, voltage, welding speed, and the relative angle between the welding torch and the workpiece) based on the data collected during the virtual welding process. Finally, it generates a conflict-free pre-simulation result that includes path suggestions, risk area markings, and optimized welding parameters, providing guidance for subsequent actual welding. By simulating the welding process in the virtual environment in advance, potential conflicts are detected and warnings are issued, reducing the risks and errors in actual welding. This is especially suitable for multi-robot collaborative operation scenarios of large and complex workpieces, and can significantly reduce on-site debugging costs and collision risks.

[0095] Example 5:

[0096] The basic content is the same as in Example 1, except that: in the third step, optimizing the welding task breakdown of the first step according to the conflict-free pre-simulation results described in the second step to obtain multiple optimized welding sub-tasks means adjusting the breakdown of the welding task in the first step according to the path suggestions and optimized welding parameters in the conflict-free pre-simulation results described in the second step to obtain multiple optimized welding sub-tasks; in the third step, planning a dedicated motion path for each welding robot based on the optimized welding sub-tasks assigned to each welding robot means planning a dedicated motion path for each robot using a fast expanding random tree algorithm based on the optimized welding sub-tasks assigned to each welding robot; the dedicated motion path includes a main motion path and several preliminary paths; in the third step, coordinating the path sequence of multiple welding robots means synchronizing the position and running status of each welding robot in real time based on the dedicated motion path of each robot, and planning the execution order of the path for each welding robot.

[0097] In application, the task decomposition unit of the task planning module, based on the conflict-free pre-simulation results, breaks down and readjusts the welding task of the first step, resulting in multiple optimized welding sub-tasks. The task allocation unit of the task planning module then formally assigns the optimized welding sub-tasks to the corresponding robots. The allocation results are synchronized to each robot and the path planning and collision avoidance module via the motion control module. The single-machine path planning unit of the path planning and collision avoidance module, combined with the sub-tasks of each robot, uses the RRT (Rapid Expanding Random Tree) algorithm to plan a unique motion path for each robot. During algorithm execution, the following formula is used as the node expansion cost formula:

[0098] ;

[0099] Where Cost(x) new ) represents the new node x new Cost(x) is used to measure the path cost from the starting point to the new node; near ) indicates proximity to the new node x new Existing node x near The cost of existing nodes is an important reference for calculating the cost of new nodes. The cost of a new node is determined based on the costs of existing nodes and the relationship between the new node and existing nodes. new This represents a new node generated during the algorithm's expansion process; its position and attributes are determined randomly by the algorithm or according to specific rules. near This represents the node in the current tree structure that is closest to the new node (new), and is used to guide the expansion direction of the new node and calculate the cost of the new node;

[0100] New nodes are randomly sampled and generated in the configuration space, and priority is given to expanding towards low-cost paths. A main motion path and several reserve paths are planned for each robot (which can be quickly switched in case of emergencies). The cost information of nodes in the path is marked synchronously to ensure that the solution converges to a collision-free solution with the optimal path cost.

[0101] After completing the single-robot path planning, the multi-robot collaborative collision avoidance unit collects real-time status data such as the position coordinates, running speed, and remaining workload of each robot. Based on the main motion path, it analyzes the spatiotemporal intersection of the multi-robot motion trajectories. If a potential collision risk is detected (such as the paths of two robots overlapping in a certain time period or spatial area), the execution order of the paths is dynamically adjusted to generate a multi-robot collaborative path sequence. Finally, the path planning and collision avoidance module transmits the planning results, including the main / preparatory paths and the collaborative sequence, to the motion control module.

[0102] Example 6:

[0103] The basic content is the same as in Example 1, except that in the fourth step, driving the welding robot to actually run means: converting the dedicated motion path and optimized welding parameters planned in the third step into control instructions that the robot can execute, transmitting them to the corresponding welding robot, and then the robot performs the actual welding operation according to the dedicated motion path and optimized welding parameters after receiving the control instructions.

[0104] In application, the motion control module controls the welding robot to perform actual operation. The motion control module converts the dedicated motion path into joint control commands, and loads the initial welding parameters provided by the digital twin and optimization module. The motion control module includes an adaptive trajectory tracking unit, a dynamic parameter adjustment unit, and a force control unit.

[0105] During the welding robot's operation phase, the adaptive trajectory tracking unit initiates sliding mode control (SMC) according to the formula. The trajectory tracking error e(t) (e(t) represents the trajectory tracking error, i.e., the deviation between the actual position of the welding torch and the planned path) and the integral term are calculated in real time; then, the formula is used... (Where K represents the control gain and λ represents the integral coefficient) Output control quantity to dynamically compensate for disturbances such as robot arm vibration and workpiece load changes;

[0106] Simultaneously, the dynamic parameter adjustment unit, based on the fuzzy PID control algorithm, collects real-time data such as weld temperature and molten pool stress, converting temperature deviation and force deviation into e(t), which is then substituted into the following formula:

[0107] ;

[0108] Among them, K p K represents the proportionality coefficient. i K represents the integral coefficient. d It represents the differential coefficient; automatically optimizes parameters such as welding speed and current; if the weld temperature is detected to be too high (exceeding the preset range), it reduces the welding speed and fine-tunes the current to suppress the risk of spatter and burn-through; if the molten pool is subjected to abnormal stress, it adjusts the welding torch angle in time to compensate for the welding needs of different materials such as high-strength steel and aluminum alloy.

[0109] Force control unit in conjunction with impedance control, according to formula To maintain a stable contact force between the welding torch and the workpiece in real time; among which, The desired contact force is represented by M, virtual mass, B, and K (specific stiffness parameter determined based on the material being welded). x represents the current position, which in the welding scenario represents the actual position of the welding torch and changes in real-time during the welding process. x0 represents the reference position, the ideal position the welding torch should be in. The positional deviation is calculated by comparing x with the current position x, and then the desired contact force is calculated. During welding, the actual position x of the welding torch dynamically changes with workpiece deformation and path switching. The system calculates the deviation by comparing it with the reference position x0 and outputs the desired contact force. Adjust the force applied by the robotic arm.

[0110] Example 7:

[0111] The basic content is the same as in Example 1, except that in the fifth step, the real-time tracking of the execution of the dedicated motion path, monitoring the actual completion of the weld, and providing feedback means: capturing the actual motion path of the welding robot in real time, comparing it with the dedicated motion path in the third step, calculating the deviation between the actual motion path and the planned dedicated motion path, and obtaining the path execution deviation; simultaneously, collecting the position information of the completed weld, the temperature distribution of the welding area, and the contact force data between the welding torch and the workpiece in real time, analyzing the forming quality of the completed weld, identifying the defects of the completed weld, and then providing feedback on the path execution deviation, weld quality data, and defect status; the fifth step... In this step, the real-time correction of the robot's motion path and welding parameters based on the feedback of the execution status of the dedicated motion path and the actual completion status of the weld refers to: dynamically adjusting the robot's motion trajectory through a sliding mode control algorithm based on the feedback of path execution deviation, so that the actual motion path of the welding robot returns to the dedicated motion path; at the same time, combining weld quality data and defect conditions, optimizing welding parameters through fuzzy PID control, and adjusting the welding torch's working posture to ensure that the welding process meets quality requirements; during the correction process, the path timing of multiple robots is synchronously associated to avoid new collaborative conflicts caused by the adjustment of a single robot, thereby achieving dynamic adaptation of motion path and welding parameters.

[0112] In application, the fifth step is achieved collaboratively by the sensor and state perception module, motion control module, and path planning and collision avoidance module. The multi-sensor fusion unit of the sensor and state perception module acquires images of the robot's motion trajectory through a visual sensor and scans the weld position through a laser contour sensor. It captures the robot's actual motion path in real time and compares it with the dedicated motion path pre-stored by the path planning module to generate a path execution deviation. The path execution deviation is then transmitted to the motion control module. The adaptive trajectory tracking unit of the motion control module uses sliding mode control to track the planned path in real time, compensate for dynamic disturbances (such as robotic arm vibration and load changes), and drive the robot joints to adjust motion parameters to ensure that the actual path reverts to the planned path. This enables the welding robot to accurately weld according to the planned path, improving welding accuracy.

[0113] Simultaneously, the temperature distribution of the welding area is collected by an infrared thermal imager, and the contact force data of the welding torch is collected by a six-axis force sensor. Combined with the weld formation image obtained by the vision sensor, the AI ​​weld tracking unit in the sensor and state perception module completes the weld quality analysis: real-time identification of weld offset, uneven width and other shape deviations, comparison with the preset weld standard shape to calculate the correction amount, and extraction of defect features such as porosity and lack of fusion to obtain weld quality data and defect status; the AI ​​weld tracking unit transmits the weld quality data and defect status to the dynamic parameter adjustment unit of the motion control module, and optimizes welding parameters such as welding current and speed through fuzzy PID control algorithm to ensure that the weld formation quality meets the preset standard; during the trajectory correction and parameter adjustment process of a single robot, the multi-robot collaborative collision avoidance unit receives the position and running status data of each robot in real time, reallocates the running interval of associated robots through the timing calculation model, and updates the global path timing table to ensure that the local adjustment of a single robot does not cause spatial interference or timing conflict of multiple robots.

[0114] Example 8:

[0115] The basic content is the same as in Embodiment 1, except that: the system includes multiple welding robots, a motion control module, a human-machine interaction and monitoring module, a task planning module, a digital twin and optimization module, a path planning and collision avoidance module, and a sensor and state perception module; the human-machine interaction and monitoring module, the task planning module, the digital twin and optimization module, the path planning and collision avoidance module, and the sensor and state perception module interact with the motion control module, and the motion control module interacts with the multiple welding robots.

[0116] In application, the human-machine interaction and monitoring module is connected to the motion control module to input welding tasks, support manual intervention, and display robot status and welding progress. The task planning module is connected to the motion control module to decompose the welding task into welding sub-tasks and dynamically allocate the welding sub-tasks to each welding robot. The digital twin and optimization module is connected to the motion control module to simulate the welding process in advance in a virtual environment based on the welding task, detect potential conflicts and provide early warnings. The path planning and collision avoidance module is connected to the motion control module to dynamically adjust the welding paths of multiple welding robots. The sensor and status perception module is connected to the motion control module to integrate vision, LiDAR, and force sensor data.

[0117] The human-computer interaction and monitoring module includes a visualization interface unit and a remote control unit, with the visualization interface unit interacting with the remote control unit via data. The task planning module includes a task decomposition unit and a task allocation unit, with the task decomposition unit interacting with the task allocation unit via data. The digital twin and optimization module includes a virtual welding simulation unit and a parameter optimization unit, with the virtual welding simulation unit interacting with the parameter optimization unit via data. The path planning and collision avoidance module includes a single-machine path planning unit and a multi-machine collaborative collision avoidance unit, with the single-machine path planning unit interacting with the multi-machine collaborative collision avoidance unit via data. The sensor and state perception module includes a multi-sensor fusion unit and an AI weld seam tracking unit, with the multi-sensor fusion unit interacting with the AI ​​weld seam tracking unit via data. The motion control module includes an adaptive trajectory tracking unit, a dynamic parameter adjustment unit, and a force control unit, with the adaptive trajectory tracking unit, dynamic parameter adjustment unit, and force control unit interacting with each other via data.

[0118] Example 9:

[0119] The basic content is the same as in Example 8, except that: the adaptive trajectory tracking unit, dynamic parameter adjustment unit, and force control unit cooperate to convert the path provided by the path planning and collision avoidance module into joint control commands, and use sliding mode control to track the planned path in real time, compensate for dynamic disturbances, and adjust parameters such as welding speed and current according to feedback information to maintain stable contact force between the welding torch and the workpiece; the visualization interface unit works in conjunction with the remote control unit to support manual uploading of welding tasks and related information, and displays the robot's real-time pose, welding progress, and alarm information in 3D form, facilitating operator intervention and fault handling; the task decomposition unit and the task allocation unit work together to split the welding task according to geometric features and The workload is distributed to each robot to balance load and efficiency; the virtual welding pre-simulation unit and the parameter optimization unit collaborate to simulate the welding process in a virtual environment, detect conflicts, and optimize welding parameters, generating pre-simulation results to guide task planning; the single-robot path planning unit and the multi-robot collaborative collision avoidance unit work together to generate collision-free paths for a single robot and coordinate the motion sequence of multiple robots to avoid collisions and deadlocks; the multi-sensor fusion unit and the AI ​​weld seam tracking unit jointly collect data such as weld seam position and temperature, track and correct deviations, and feed them back to the motion control module to ensure welding accuracy; the data interaction between the units ensures efficient collaboration of the task from planning, simulation to execution and correction, improving the stability and efficiency of multi-robot collaborative operation.

[0120] Example 10:

[0121] The basic content is the same as in Embodiment 8, except that the system further includes a fault diagnosis and self-recovery module; the fault diagnosis and self-recovery module interacts with the motion control module.

[0122] In application, the fault diagnosis and self-recovery module includes an anomaly detection unit and a self-recovery strategy unit. The anomaly detection unit and the self-recovery strategy unit interact with each other, and the self-recovery strategy unit interacts with the motion control module. During the welding process, the anomaly detection unit continuously monitors the operating status of each robot in real time, such as motor current, temperature, and vibration signals. Once an anomaly signal is detected, it is transmitted to the self-recovery strategy unit. After receiving the anomaly signal, the self-recovery strategy unit triggers self-recovery operations such as switching to a backup path to realize the system's self-recovery function and ensure the normal progress of the welding process. At the same time, the self-recovery strategy unit feeds back the anomaly information to the motion control module. In addition, if the self-recovery operation cannot be performed, the self-recovery strategy unit will issue an alarm. The motion control module will then synchronously transmit the anomaly information or alarm information to the human-machine interaction and monitoring module. The human-machine interaction and monitoring module's visualization interface unit presents the anomaly information (such as alarm information) to the operator, allowing the operator to be aware of it in a timely manner and take measures. This supports manual intervention to handle abnormal situations, ensuring the safety and continuity of the welding process, reducing downtime caused by faults, and improving the reliability and stability of the system.

[0123] The above description is only a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. Any equivalent modifications or changes made by those skilled in the art based on the content disclosed in the present invention should be included within the scope of protection set forth in the claims.

Claims

1. A method for collaborative control of multiple welding robots, characterized in that: The control method includes the following steps: Step 1: First, input the welding task, break it down into multiple welding sub-tasks, and then pre-assign the welding sub-tasks to the corresponding welding robots based on the real-time status of each welding robot. Step 2: Based on the pre-assigned welding sub-tasks of the welding robots, simulate the operation of multiple welding robots in a virtual environment to detect whether there are any operational conflicts among the multiple welding robots. If there are no operational conflicts, optimize the welding parameters of the multiple welding robots to form a conflict-free pre-simulation result. If there are operational conflicts, readjust the splitting and allocation of welding sub-tasks according to the conflict situation, and then simulate the operation of multiple welding robots in the virtual environment again until there are no operational conflicts. Then optimize the welding parameters of the multiple welding robots to form a conflict-free pre-simulation result. Step 3: Optimize the welding task breakdown in Step 1 according to the conflict-free pre-simulation results described in Step 2, resulting in multiple optimized welding sub-tasks. Then, assign the optimized welding sub-tasks to the corresponding welding robots. Based on the optimized welding sub-tasks assigned to each welding robot, plan a dedicated motion path for each robot. At the same time, coordinate the path timing of multiple welding robots. Step 4: Drive the welding robot to perform actual operation according to the above-mentioned dedicated motion path and path sequence, and in combination with the optimized welding parameters in the above-mentioned conflict-free pre-simulation results; Step 5: During the actual operation of the welding robot, the execution of the dedicated motion path is tracked in real time, the actual completion of the weld is monitored, and feedback is provided. Then, based on the feedback of the execution of the dedicated motion path and the actual completion of the weld, the robot's motion path and welding parameters are corrected in real time, thereby realizing the collaborative control of multiple robots.

2. The method for collaborative control of multiple welding robots according to claim 1, characterized in that: The control method further includes: Step 6: During the actual operation of the welding robots, continuously monitor the operating status of each welding robot for any abnormalities. If an abnormality occurs, trigger a self-recovery operation to maintain the operation of each welding robot. If the self-recovery operation cannot be performed, trigger an alarm and wait for repair.

3. A method for collaborative control of multiple welding robots according to claim 1 or 2, characterized in that: In the first step, the input welding task refers to inputting the actual welding task and welding object information; In the first step, the step of splitting the welding task into multiple welding sub-tasks means: based on the geometric features in the input welding object information, using a graph segmentation algorithm, the input actual welding task is divided into multiple independent and continuous welding sub-tasks. In the first step, the step of pre-allocating welding sub-tasks to the corresponding welding robots based on the real-time status of each welding robot means: based on the real-time status of each welding robot, solving the optimal allocation scheme through mixed integer programming, and pre-allocating the segmented welding sub-tasks to the corresponding welding robots.

4. A method for collaborative control of multiple welding robots according to claim 1 or 2, characterized in that: In the second step, the simulation of the operation of multiple welding robots in a virtual environment based on the pre-assigned welding sub-tasks of the welding robots, and the detection of whether there are operational conflicts among the multiple welding robots, refers to: first, constructing a simulation environment consistent with the physical welding scene in the digital virtual space, importing the three-dimensional model of the welding object and the kinematic model of each welding robot, and integrating the finite element analysis module; then, assigning corresponding weld segment execution instructions to each robot according to the pre-assigned welding sub-tasks; then driving the virtual welding robot model to run according to the instructions, and simultaneously analyzing the real-time thermal deformation data of the welding object model through the finite element analysis module; at the same time, tracking the relative positional relationship between the virtual welding robot's motion trajectory and the thermally deformed welding object model in real time, and judging whether there is spatial interference between the robot and the welding object model, and between the robots themselves; In the second step, the step of optimizing the welding parameters of multiple welding robots to form a conflict-free pre-simulation result if there is no operational conflict means that after confirming that there is no spatial interference between the robot and the welding object model, and between the robots, the welding parameters are optimized based on the simulated welding process data, and finally a conflict-free pre-simulation result is generated; the conflict-free pre-simulation result includes path suggestions, risk area markings, and optimized welding parameters; In the second step, if there is a conflict, the welding sub-tasks are re-splitting and reassigned according to the conflict situation. Then, the operation of multiple welding robots is simulated again in the virtual environment until there is no conflict. This means that spatial interference is detected between the robot and the welding object model, or between the robot and the object. The conflict type is analyzed first, and the welding sub-tasks are re-splitting and reassigned according to the conflict type. After the adjustment is completed, the new sub-task allocation scheme is imported into the virtual environment, and the operation of multiple welding robots is simulated repeatedly until all conflicts are eliminated.

5. The method for collaborative control of multiple welding robots according to claim 4, characterized in that: In the third step, optimizing the welding task breakdown of the first step according to the conflict-free pre-simulation results described in the second step to obtain multiple optimized welding sub-tasks means: adjusting the breakdown of the welding task in the first step according to the path suggestions and optimized welding parameters in the conflict-free pre-simulation results described in the second step to obtain multiple optimized welding sub-tasks. In the third step, planning a dedicated motion path for each welding robot based on the optimized welding sub-task assigned to each welding robot means: planning a dedicated motion path for each robot using a fast expanding random tree algorithm based on the optimized welding sub-task assigned to each welding robot; the dedicated motion path includes a main motion path and several preliminary paths. In the third step, coordinating the path sequence of multiple welding robots refers to: based on the dedicated motion path of each robot, synchronizing the position and running status of each welding robot in real time, and planning the order of path execution for each welding robot.

6. A method for collaborative control of multiple welding robots according to claim 1 or 2, characterized in that: In the fourth step, driving the welding robot to actually run means: converting the dedicated motion path and optimized welding parameters planned in the third step into control instructions that the robot can execute, transmitting them to the corresponding welding robot, and then having the robot perform the actual welding operation according to the dedicated motion path and optimized welding parameters after receiving the control instructions.

7. A method for collaborative control of multiple welding robots according to claim 1 or 2, characterized in that: In the fifth step, the real-time tracking of the execution of the dedicated motion path, monitoring the actual completion of the weld, and providing feedback refers to: capturing the actual motion path of the welding robot in real time, comparing it with the dedicated motion path in the third step, calculating the deviation between the actual motion path and the planned dedicated motion path, and obtaining the path execution deviation; at the same time, collecting the position information of the completed weld, the temperature distribution of the welding area, and the contact force data between the welding torch and the workpiece in real time, analyzing the forming quality of the completed weld, identifying the defects of the completed weld, and then providing feedback on the path execution deviation, weld quality data, and defect status; In the fifth step, the real-time correction of the robot's motion path and welding parameters based on the feedback of the execution status of the dedicated motion path and the actual completion status of the weld refers to: dynamically adjusting the robot's motion trajectory through a sliding mode control algorithm based on the feedback of path execution deviation, so that the actual motion path of the welding robot returns to the dedicated motion path; at the same time, combining weld quality data and defect conditions, optimizing welding parameters through fuzzy PID control, and adjusting the welding torch's working posture to ensure that the welding process meets quality requirements; During the correction process, the path timing of multiple robots is synchronously linked to avoid new coordination conflicts caused by the adjustment of a single robot, thereby achieving dynamic adaptation of motion paths and welding parameters.

8. The method for collaborative control of multiple welding robots according to claim 2, characterized in that: In the sixth step, continuously monitoring whether the operating status of each welding robot is abnormal means continuously monitoring whether the operating status of each welding robot, including motor current, temperature, and vibration signals, is abnormal.

9. A system containing a collaborative control method for multiple welding robots according to claim 1 or 2, characterized in that: The system includes multiple welding robots, a motion control module, a human-computer interaction and monitoring module, a task planning module, a digital twin and optimization module, a path planning and collision avoidance module, and a sensor and state perception module. The human-computer interaction and monitoring module, task planning module, digital twin and optimization module, path planning and collision avoidance module, and sensor and state perception module interact with the motion control module, and the motion control module interacts with multiple welding robots.

10. A collaborative control system for multiple welding robots according to claim 9, characterized in that: The system also includes a fault diagnosis and self-recovery module; the fault diagnosis and self-recovery module interacts with the motion control module.

Images