Double-robot collaborative vision-guided processing system and method in a narrow space

Abstract

The present invention discloses a dual-robot co-integrated collaborative vision-guided processing system and method in a narrow space, wherein a hole-making collaborative robot and a vision-guided robot constitute a co-integrated collaborative robot; a tooling frame is used to fix and install a workpiece to be processed; the hole-making collaborative robot is fixed on an AGV carrier and arranged on one side of the workpiece to be processed, and a hole-making end effector is installed at its end, and the hole-making collaborative robot can move the hole-making end effector to a designated processing position inside the workpiece to be processed for hole-making processing; a vision-guided robot is fixed on a base and arranged on the other side of the workpiece to be processed, and a binocular vision camera is installed at its end, and the vision-guided robot can move the binocular vision camera to an appropriate position inside the workpiece to be processed to visually guide the hole-making end effector. The present invention uses a co-integrated robot collaborative method to replace manual operations in performing hole-making operations inside a narrow space, which greatly improves the accuracy and efficiency of hole-making in a narrow space.

Description

Technical Field

[0001] The present invention belongs to the technical field of robot processing and assembly, and specifically relates to a dual-robot collaborative vision-guided processing system and method in a narrow space. Background Art

[0002] In the field of narrow space processing, the existing processing method is mainly to carry handheld devices manually into narrow spaces for processing operations, which causes workers to maintain specific operating postures for a long time in cramped spaces, greatly consuming human resources and consuming a lot of processing time. In addition, manual operations using handheld devices have problems such as unstable processes, complex processes, and low efficiency, which are difficult to adapt to the high-precision processing requirements of narrow space parts. Therefore, the development of advanced intelligent processing and manufacturing technologies and equipment based on collaborative robots is one of the effective ways to solve this problem.

[0003] At present, the robots for narrow space operation mainly include snake-arm robots, such as the existing snake-shaped robots and snake-arm robots, which have the disadvantages of low absolute positioning accuracy, weak rigidity, and difficulty in achieving high-precision processing of workpieces. At the same time, the processing and assembly of narrow working spaces also pose new challenges to the high-precision control, task planning, and anti-collision strategies of robots. It is urgent to break through the core key technologies and develop collaborative robots for intelligent processing and manufacturing technology equipment for narrow space components. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a dual-robot co-operation vision-guided processing system and method in a narrow space in view of the deficiencies of the above-mentioned prior art. The system includes a tooling frame, a hole-making collaborative robot, a vision-guided robot, an AGV carrier, a hole-making end effector and a vision-guided robot. The tooling frame is used to fix and install the workpiece to be processed, and a hole-making collaborative robot and a vision-guided robot are respectively arranged on two sides of the tooling frame. The hole-making collaborative robot is fixed on an AGV carrier that can move in a plane, and a hole-making end effector is installed at its end for drilling and countersinking the workpiece to be processed; the vision-guided robot is fixed to the ground through a base, and a binocular vision camera is provided at its end for visually guiding and correcting the posture of the hole-making collaborative robot, so as to realize the co-operation between the dual robots. The system uses digital twin technology to monitor the co-operation process of the dual robots in real time, and uses the co-operation dual robot system to replace manual operations in hole-making and countersinking in a narrow space, thereby improving the hole-making accuracy and efficiency while ensuring the safety of the hole-making process.

[0005] In order to achieve the above technical objectives, the technical solution adopted by the present invention is:

[0006] The dual-robot collaborative vision-guided processing system in a narrow space includes an AGV carrier, a hole-making collaborative robot, a hole-making end effector, a tooling frame, a vision camera, a vision-guided robot and a base;

[0007] The hole-making collaborative robot and the vision-guided robot constitute a co-collaborative robot; the tooling frame is used to fix and install the workpiece to be processed; the hole-making collaborative robot is fixed on the AGV carrier and arranged on one side of the workpiece to be processed, and a hole-making end effector is installed at its end. The hole-making collaborative robot can move the hole-making end effector to the specified processing position inside the workpiece to be processed to perform hole-making processing; the vision-guided robot is fixed on the base and arranged on the other side of the workpiece to be processed, and a vision camera is installed at its end. The vision-guided robot can move the vision camera to the appropriate position inside the workpiece to be processed to visually guide the hole-making end effector.

[0008] To optimize the above technical solutions, the specific measures taken also include:

[0009] The above-mentioned hole-making end effector includes a ball-jointed suction cup, a suction cup support rod, a flange, a suction cup telescopic locking device, a spindle feeding device, a spindle, a pressure foot telescopic rod, a mounting plate and a contact pressure foot device; the flange is connected to the end of the hole-making collaborative robot; the mounting plate is located in the middle of the hole-making end effector, on which are fixed 4 suction cup support rods, a pressure foot telescopic rod, a spindle feeding device and a central opening for placing the spindle; the suction cup support rod is fixed by a hinge, and its support angle can be adjusted to adapt to the complex environment inside the narrow space to be processed; there is a ball-jointed suction cup under the suction cup support rod that can be movably adjusted The suction cup angle is connected to the air pipe at the upper end, and the air circuit is opened to control the suction of the ball-hinged suction cup, and a suction cup telescopic locking device is installed, and the locking and unlocking of the suction cup are used to control the locking and telescopic movement of the ball-hinged suction cup; the pressure foot telescopic rod is fixed on the mounting plate at the top and connected to the air circuit, and is connected to the contact pressure foot device at the bottom to control the extension and retraction of the pressure foot; three normal sensors are installed on the contact pressure foot device, which can detect the normal direction of the pressure foot; the spindle feeding device is connected to the spindle, and controls the up and down feeding of the spindle through the movement of the worm gear; a drill is installed at the bottom of the spindle for making holes and countersinking.

[0010] The above-mentioned hole-making end effector can replace the spindle bit to realize hole-making and countersinking of different hole diameters. The hole diameter accuracy of the hole-making is higher than H8, the normal accuracy is ≤0.5°, and the countersinking accuracy is ≤0.03mm.

[0011] The bottom of the above-mentioned tooling frame is provided with 6 pulleys, which can drive the tooling frame to move freely; each pulley can be locked by bolts to fix the tooling frame in a specified position; the upper support surface of the tooling frame can move up and down to adjust the vertical position of the workpiece to be processed thereon; there are 9 groups of movable workpiece support rods in the upper support surface, and their positions can be adaptively adjusted according to the shape of the workpiece to be processed to ensure that the workpiece to be processed is installed securely and stably.

[0012] The binocular vision camera is connected to the flange through a camera connecting rod and fixed at the end of the vision-guided robot.

[0013] The above-mentioned collaborative hole-making robot and vision-guided robot first plan their initial motion paths through offline programming, and use the dual-robot symbiotic collaborative hole-making software to control their movements. During the working process, the digital twin technology is used to monitor the motion posture of the dual robots in real time. The vision-guided robot provides real-time on-site situation data and guides the posture adjustment of the collaborative hole-making robot to reach the target posture; the position accuracy of the two robots is ≤0.2mm, the posture accuracy is ≤0.1°, and the trajectory accuracy is ≤0.5mm; the dual-robot symbiotic collaborative hole-making software uses the RRT algorithm to optimize the path, and obtains a collision-free trajectory for the collaborative hole-making robot and the vision-guided robot to move inside the workpiece to be processed.

[0014] The method of dual-robot collaborative vision-guided processing in a narrow space includes:

[0015] Step 1: Build and adjust the work platform according to the different parts to be processed, and build the corresponding working environment in the offline programming software;

[0016] Step 2: Analyze the positions of the holes to be drilled, determine the drilling order, and use the RRT algorithm based on multiple safety constraints to plan the motion path of the drilling collaborative robot to complete the drilling of all holes in a narrow space, with the inner wall of the workpiece to be processed, another fusion robot, and the visual observable range as the constraint targets, and plan the motion path of the visual guidance robot driving the binocular vision camera for visual guidance;

[0017] Step 3: The control software connects the collaborative hole-making robot and the vision-guided robot, turns on the digital twin monitoring function, and completes the preparation work;

[0018] Step 4: Control the collaborative robot to start the workflow. The vision guides the robot to the point where the camera can see the posture of the hole-making end effector making the first hole. The collaborative robot drives the hole-making end effector to the vicinity of the hole to be made.

[0019] Step 5: When the hole-making collaborative robot and the vision-guided robot reach their respective predetermined positions, the point matching algorithm based on visual position coupling and the robot multi-level posture adjustment algorithm are used to perform visual guidance posture adjustment on the hole-making collaborative robot to ensure that it accurately reaches the predetermined position;

[0020] Step 6: Open the air path of the suction cup of the hole-making end effector to make it adsorb on the inner surface of the workpiece to be processed;

[0021] Step 7: For the hole-making collaborative robot whose posture has changed after the hole-making end effector completes adsorption, a multi-level posture adjustment algorithm is used to guide and fine-tune the posture of the hole-making end effector again. At this point, the collaborative posture adjustment of the hole-making collaborative robot and the vision-guided robot is completed;

[0022] Step 8: Start the hole making process: Open the suction cup retractable locking device of the hole making end effector to lock the suction cup so that it will not move up and down;

[0023] Step 9: Open the air path above the telescopic rod of the pressure foot to push out the contact pressure foot device to the position where the hole is to be made;

[0024] Step 10: Control the spindle feed device to start drilling holes. During this process, the drill feed depth D, speed V and other parameters can be controlled;

[0025] Step 11: After the hole making is completed, the spindle feeding device is controlled to retract the spindle (306), the pressure foot telescopic rod is controlled to retract the contact pressure foot device, the suction cup telescopic locking device is controlled to unlock the suction cup, and the air is turned off. At this point, the single hole position hole making task is completed;

[0026] Step 12: After completing a hole drilling, the control software continues to control the collaborative drilling robot and the visual guidance robot to move to the ready posture of the next hole to be drilled, and re-executes the above-mentioned hole drilling process from step 5 to step 11;

[0027] Step 13: After all the holes are made, the control software controls the hole collaborative robot and the vision-guided robot to move out the workpiece to be processed, and the system processing task is completed.

[0028] The error handling strategy during the processing is as follows: when the collaborative robot is drilling a hole, if the drill bit is damaged or the end fails, first press the emergency stop button of the robot system and interrupt the program, then move the robot a short safe distance, record its current posture, and use the offline programming planning algorithm to plan the path from the current posture back to the initial position outside the workpiece to be processed, and then use the control software to control the robot to move out of the narrow space, and then replace or repair it.

[0029] The RRT algorithm based on multiple safety constraints in step 2 above is specifically as follows:

[0030] Based on the RRT algorithm, first determine the starting and ending postures of the robot path planning, set the robot workspace range and the number of algorithm iterations k;

[0031] In the robot axis space, take the starting point as the first parent node N0, and randomly select the robot axis angles to obtain the robot random point N r ; For the semi-enclosed environment where the co-integrated robot is in a narrow space inside the workpiece to be processed, a certain probability is set to take a random point N r is the end pose node N goal , and this probability is 15% to 30%;

[0032] For a random point N in the robot axis space rand , using the Euclidean distance formula Calculate the distance between it and each stored node, and find the closest node N as the parent node of the random point near ;

[0033] At this random point N rand With its parent node N near Take a certain step length on the direction vector as the new node N new ;

[0034] Using the collision detection function in the offline programming software, the inner wall of the workpiece to be processed and another co-integrated robot are regarded as multi-target obstacles to check whether the robot can near To the new node N new Will there be a collision during the movement? If a collision occurs, a random point is reselected and a new iteration is entered. If no collision occurs, the node is stored and recorded as N. m ;

[0035] Each node N m After finding it, we need to determine whether we have reached the end point. If the distance between the new node position and the end point position is less than one step, we consider that we have reached the end point and found the path.

[0036] If the target point cannot be obtained in the workspace of the collaborative robot, it means that the collaborative robot needs to adjust its posture at a large angle when planning to the target point. In this case, the collaborative robot needs to be moved out of the workpiece to be processed, and the collaborative robot needs to complete the large-angle posture transformation outside the workpiece, and then enter the workpiece to complete the complete path planning;

[0037] If the target point is outside the workspace of the collaborative robot, it means that the collaborative robot cannot reach the point in the current deployment state. The tooling frame and the workpiece to be processed are turned around, so that the hole-making collaborative robot enters from another opening of the workpiece to be processed and continues to complete the path planning.

[0038] The found path is smoothed using spline interpolation, ultimately forming a collision-free path for the robot and completing the robot's path planning.

[0039] In the above step 5, the point matching algorithm based on viewpoint coupling is specifically as follows:

[0040] First, the TCP coordinate system {A} of the hole-making end effector is established, and several identification target points are attached to the hole-making end effector. The three-dimensional coordinates of these points in the TCP coordinate system {A} are derived, and the Euclidean distance between each point and all other points is calculated and stored in the data container.

[0041] The visual camera takes pictures of the points on the end effector of the hole, obtains the three-dimensional coordinates of these points in the camera coordinate system {B}, and also calculates the Euclidean distance between each point and all other points and stores them in the data container;

[0042] By comparing the distance data containers in the TCP and camera scene coordinate systems, the point with the highest matching degree between the two is the same point, and the one-to-one correspondence of points in different scenes can be completed.

[0043] The above-mentioned step 5 of multi-level posture adjustment of the robot based on visual position coupling is specifically as follows:

[0044] Step 1: First, prepare the data. Use the point matching algorithm based on view-position coupling to match the target points of the TCP coordinate system {A} of the hole-making end effector (3) with the target points of the camera coordinate system {B} one by one, and obtain the current accurate position of the TCP in the camera coordinate system {B}. B T;

[0045] Step 2: Set the current position of TCP in the camera coordinate system {B} B T is transformed to the base coordinate system {C} of the hole-making collaborative robot: C T;

[0046] Step 3: Read the current pose C After T, the target pose required by offline simulation is C T e Make comparisons;

[0047] Step 4: Calculate the current pose C T and target posture C T e The error Δ between them, where the error is calculated by comparing the corresponding values ​​of the two attitude data sets item by item. The position error is the difference between the two three-dimensional coordinates, and the attitude error is the difference between the two Euler angles;

[0048] Step 5: By comparing the calculated posture error Δ with the error threshold ψ, it is determined whether the hole-making collaborative robot has reached the target posture. If the errors of all six degrees of freedom are within the threshold range, it is considered that the hole-making collaborative robot has reached the target posture.

[0049] Step 6: If the error Δ exceeds the error threshold ψ, the error is input as feedback to the PID controller, and the error Δ is divided into multiple compensation amounts δ to improve the smoothness and accuracy of posture adjustment. The PID controller calculates the compensation amount δ that needs to be adjusted step by step, and adjusts the movement of the hole-making collaborative robot with the compensation amount δ through the control software. During the adjustment process, the hole-making collaborative robot moves its position and changes its posture according to the feedback error information to gradually reduce the error;

[0050] Step 7: Iterative posture adjustment process, the hole-making collaborative robot updates its current posture data after each adjustment C T′, and recalculate the error Δ′. Through continuous feedback and adjustment, the hole-making collaborative robot gradually approaches the target posture. C T e When the hole-making collaborative robot detects that the errors of all degrees of freedom are less than the set threshold ψ at a certain moment, it is considered that the hole-making collaborative robot has successfully reached the target posture C T e , the posture adjustment process is completed.

[0051] The present invention has the following beneficial effects:

[0052] The present invention can perform integrated collaborative control on the co-integrated dual robots in the system, use the dual robots to perform co-integrated collaborative processing operations, complete the processing of target holes in a narrow space, and ensure the safe operation of the robot system and the workpiece, replacing the previous manual processing tasks in a narrow space, reducing the fatigue of workers, and improving the efficiency of narrow space processing. It can be applied to narrow space workpiece processing in multiple fields, such as air intakes, wing sections, cabin sections, etc. in the aerospace field. The application of the present invention can reduce manual operation costs, improve the degree of intelligence, and improve product processing accuracy.

[0053] The dual-robot collaborative working mode of the present invention uses a binocular vision measurement system to guide the visually guided robot to complete high-precision guidance of the processing posture of the hole-making collaborative robot in a narrow space, so as to perform the processing work of the target hole position, thereby ensuring the accuracy of the robot's hole making.

[0054] The symbiotic dual robots of the present invention plan a collision-free trajectory for robot hole making through an offline programmed path planning algorithm, thereby ensuring the safety of the hole making process.

[0055] The hole-making collaborative robot of the present invention can be moved by an AGV carrier, and the tooling frame can adaptably fix any workpiece to be processed and be movable, so that the processing platform can be adjusted for workpieces of various shapes and sizes to perform different processing tasks, thereby improving the system's adaptability to the tasks of different workpieces. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 Schematic diagram of the dual-robot collaborative vision-guided processing system in a narrow space of the present invention;

[0057] Figure 2 This is a schematic diagram of the structure of the hole-making end effector of the collaborative robot of the present invention;

[0058] Figure 3 It is a schematic diagram of the structure of the binocular vision measurement system of the present invention;

[0059] Figure 4 It is a flow chart of the working method of the double-robot collaborative vision-guided processing system in a narrow space of the present invention;

[0060] Figure 5 The flowchart of the RRT algorithm based on multiple safety constraints of the present invention;

[0061] Figure 6 This is a flow chart of the robot multi-level posture adjustment algorithm based on view-position coupling of the present invention;

[0062] Figure 1-3 The figure numbers are as follows: 1-AGV carrier; 2-hole-making collaborative robot; 3-hole-making end effector; 4-tooling frame; 5-vision camera; 6-vision-guided robot; 7-base; 301-ball joint suction cup; 302-suction cup support rod; 303-flange; 304-suction cup telescopic locking device; 305-spindle feeding device; 306-spindle; 307-pressure foot telescopic rod; 308-mounting plate; 309-contact pressure foot device; 501-binocular vision camera; 502-camera connecting rod; 503-flange. DETAILED DESCRIPTION

[0063] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0064] Although the steps in the present invention are arranged with numbers, they are not used to limit the order of the steps. Unless the order of the steps is clearly stated or the execution of a certain step requires other steps as a basis, the relative order of the steps can be adjusted. It is understood that the term "and / or" used in this article involves and covers any and all possible combinations of one or more of the associated listed items.

[0065] The invention provides a dual-robot collaborative vision-guided processing system in a narrow space. The system can build a system platform for various shapes of workpieces to be processed. The dual robots are respectively arranged on both sides of the workpiece to be processed. One hole-making collaborative robot carries a hole-making end effector to enter the workpiece to make holes, and a vision-guided robot carries a binocular vision camera to enter the other side of the workpiece to visually guide the hole-making collaborative robot and complete the processing work. Figure 1 As shown, the system includes an AGV carrier 1, a hole-making collaborative robot 2, a hole-making end effector 3, a tooling frame 4, a visual camera 5, a visual guidance robot 6, and a base 7;

[0066] The tooling frame 4 is used to fix and install the workpiece to be processed; the hole-making collaborative robot 2 is fixed on the AGV carrier 1 and is arranged on one side of the workpiece to be processed, and a hole-making end effector 3 is installed at its end. The hole-making collaborative robot can move the hole-making end effector 3 to a specified processing position inside the workpiece to be processed to perform hole-making processing; the visual guidance robot 6 is fixed on the base 7 and is arranged on the other side of the workpiece to be processed, and a visual camera 5 is installed at its end. The visual guidance robot can move the visual camera 5 to an appropriate position inside the workpiece to be processed to visually guide the hole-making end effector 3;

[0067] like Figure 2 As shown, the hole-making end effector 3 of the present invention is composed of a ball joint suction cup 301, a suction cup support rod 302, a flange 303, a suction cup telescopic locking device 304, a spindle feeding device 305, a spindle 306, a pressure foot telescopic rod 307, a mounting plate 308 and a contact pressure foot device 309;

[0068] The hole-making end effector 3 is connected to the end of the hole-making collaborative robot 2 through a flange 303; the mounting plate 308 is located in the middle of the hole-making end effector 3, on which are fixed four suction cup support rods 302, a pressure foot telescopic rod 307, a spindle feeding device 305 and a central opening for placing the spindle 306; the suction cup support rod 302 is fixed by a hinge, and its support angle can be adjusted to adapt to the complex environment inside the workpiece to be processed in a narrow space; there is a ball-jointed suction cup 301 under the suction cup support rod 302, which can flexibly adjust the suction cup angle, and the upper end is connected to the air pipe, which is opened and controlled by the air circuit The suction cup inhales air and is equipped with a suction cup telescopic locking device 304, which controls the locking and telescoping of the suction cup by locking and unlocking it; the pressure foot telescopic rod 307 is fixed on the mounting plate 306 at the top and connected to the air circuit, and is connected to the contact pressure foot device 309 at the bottom to control the push-out and retraction of the pressure foot; three normal sensors are installed on the contact pressure foot device 309 to detect the normal direction of the pressure foot; the spindle feeding device 305 is connected to the spindle 306, and controls the up and down feeding of the spindle through the movement of the worm gear; a drill is installed at the bottom of the spindle 306 for making holes and countersinking;

[0069] The hole making end effector 3 can replace the spindle bit, so as to realize the hole making and countersinking of different hole diameters; the hole diameter accuracy of the hole making is higher than H8, the normal accuracy is ≤0.5°, and the countersinking accuracy is ≤0.03mm.

[0070] There are 6 pulleys at the bottom of the tooling frame 4, which can drive the tooling frame 4 to move freely; each pulley can be locked by a bolt to fix the tooling frame 4 at a specified position; the upper support surface of the tooling frame 4 can move up and down to adjust the vertical position of the workpiece to be processed thereon; there are 9 groups of movable workpiece support rods on the upper support surface, and their positions can be adaptively adjusted according to the shape of the workpiece to be processed to ensure that the workpiece to be processed is installed securely and stably.

[0071] like Figure 3 As shown, the visual camera 5 of the present invention is a binocular visual camera 501 , which is connected to a flange 503 via a camera connecting rod 502 and fixed at the end of a visual guidance robot 6 .

[0072] The system is controlled by dual-robot collaborative hole-making software, and the motion postures of the dual robots are monitored in real time through digital twin technology throughout the entire work process.

[0073] The hole-making collaborative robot 2 and the vision-guided robot 6 first plan their initial motion paths through offline programming, and use the dual-robot collaborative hole-making software to control their movements. During the working process, the vision-guided robot 6 provides real-time on-site situation data to guide the hole-making collaborative robot 2 to reach the target position and posture; the position accuracy of the two robots is ≤0.2mm, the posture accuracy is ≤0.1°, and the trajectory accuracy is ≤0.5mm.

[0074] The dual-robot path planning uses the RRT algorithm and path optimization to obtain a collision-free trajectory for the robots to move inside the workpiece.

[0075] The working method of the dual-robot co-integrated collaborative vision-guided processing system in a narrow space, that is, the dual-robot co-integrated collaborative vision-guided processing method in a narrow space, such as Figure 4 As shown, including:

[0076] Step 1: Build and adjust the work platform according to the different parts to be processed, and build the corresponding working environment in the offline programming software;

[0077] Step 2: Analyze the positions of the holes to be drilled, determine the drilling order, and use the RRT algorithm based on multiple safety constraints to plan the motion path of the drilling collaborative robot 2 to complete the drilling of all holes in a narrow space, and the motion path of the visual guidance robot 6 driving the visual camera 5 for visual guidance.

[0078] Step 3: Control the software to connect the two robots in collaboration, enable the digital twin monitoring function, and complete the preparation work;

[0079] Step 4: Control the cooperating robots to start the workflow, and the two robots reach their respective positions, that is, the visual guidance robot 6 reaches the position where the camera can see the hole-making end effector 3 making the first hole, and the hole-making collaborative robot 2 drives the hole-making end effector 3 to the vicinity of the hole to be made;

[0080] Step 5: When the two robots reach their respective predetermined positions, the point matching algorithm based on visual position coupling and the robot multi-level posture adjustment algorithm are used to visually guide the hole-making collaborative robot (2) to adjust its posture to ensure that it accurately reaches the predetermined position; the visual camera 5 takes a photo to collect the actual posture information of the hole-making end effector 3 The ideal position to reach with the collaborative robot By comparison, the algorithm obtains the posture deviation Δ1 of the hole-making collaborative robot 2;

[0081] According to the posture deviation Δ1, the hole-making collaborative robot 2 is controlled to perform posture fine-tuning so that it reaches a predetermined posture in reality;

[0082] Step 6: Open the air path of the suction cup of the hole-making end effector 3 so that it is adsorbed on the inner surface of the workpiece to be processed;

[0083] Step 7: Since the posture of the hole-making end effector 3 will change during the adsorption process, it is necessary to use a multi-level posture adjustment algorithm to guide and fine-tune the posture of the hole-making end effector 3 again after the hole-making end effector (3) has completed adsorption. At this point, the collaborative posture of the two robots has been adjusted;

[0084] Step 8: Start the hole making process. Open the suction cup extension and locking device 304 of the hole making end effector 3 to lock the suction cup so that it will not move up and down;

[0085] Step 9: Open the air path above the pressure foot telescopic rod 307 to push out the contact pressure foot device 309 to the position where the hole is to be made;

[0086] Step 10: Control the spindle feeding device 305 to start drilling holes. During this process, the feeding depth D, rotation speed V and other parameters of the drill bit can be controlled;

[0087] Step 11: After the hole making is completed, the spindle feeding device 305 is controlled to retract the spindle (306), the pressure foot telescopic rod 307 is controlled to retract the contact pressure foot device 309, and the suction cup telescopic locking device 304 is controlled to unlock the suction cup and turn off the air. At this point, the single hole position hole making task is completed;

[0088] Step 12: After completing a hole, the control software continues to control the two robots to move to the next hole to be made. Re-execute the hole making process from step 5 to step 11 above;

[0089] Step 13: After all holes are made, the control software controls the dual robots to move out the workpieces to be processed, and the system processing task is completed;

[0090] The planning and processing error handling strategies in the above method are:

[0091] During the operation of the collaborative robot, when drilling a hole, if the drill bit is damaged or the end fails, first press the emergency stop button of the robot system and interrupt the program, then move the robot a short safe distance, record its current posture, and use the offline programming planning algorithm to plan the path from the current posture back to the initial position outside the workpiece to be processed, and then use the control software to control the robot to move out of the narrow space, and then replace or repair it.

[0092] like Figure 5 As shown, the collaborative robot path planning algorithm of the present invention is implemented by combining the RRT algorithm with offline programming software.

[0093] First, a simulation platform identical to the actual system platform must be built in the offline programming software so that the required model can be called in the algorithm.

[0094] Using the RRT algorithm, first determine the starting and ending postures of the robot path planning, set the robot workspace range and the number of algorithm iterations k;

[0095] In the robot axis space, take the starting point as the first parent node N0, and randomly select the robot axis angles to obtain the robot random point N r ; For the semi-enclosed environment where the co-integrated robot is in a narrow space inside the workpiece to be processed, in order to improve the accuracy and speed of path search, a certain probability is set to take a random point N r is the end pose node N goal , and this probability is between 15% and 30%.

[0096] For a random point N in the robot axis space rand , using the Euclidean distance formula Calculate the distance between it and each stored node, and find the closest node N as the parent node of the random point near ;

[0097] At this random point N rand With its parent node N near Take a certain step length on the direction vector as the new node N new ;

[0098] Using the collision detection function in the offline programming software, the inner wall of the workpiece to be processed and another co-integrated robot are regarded as multi-target obstacles to check whether the robot can near To the new node N new Will there be a collision during the movement? If a collision occurs, a random point is reselected and a new algorithm iteration is entered. If no collision occurs, the node is stored and recorded as N. m ;

[0099] Each node N m After finding it, we need to determine whether we have reached the end point. If the distance between the new node position and the end point position is less than one step, we consider that we have reached the end point and found the path.

[0100] If the target point cannot be obtained in the workspace of the collaborative robot, it means that the collaborative robot needs to adjust its posture at a large angle when planning to the target point. In this case, the collaborative robot needs to be moved out of the workpiece to be processed, and the collaborative robot needs to complete the large-angle posture transformation outside the workpiece, and then enter the workpiece to complete the complete path planning;

[0101] If the target point is outside the workspace of the collaborative robot, it means that the collaborative robot cannot reach the point in the current deployment state. The tooling frame 4 and the workpiece to be processed are turned around, so that the hole-making collaborative robot 2 enters from another opening of the workpiece to be processed and continues to complete the path planning;

[0102] The found path is smoothed using spline interpolation, ultimately forming a collision-free path for the robot and completing the robot's path planning.

[0103] Based on this, the motion path required for the symbiotic robot to process all hole positions can be planned. This method realizes the collision-free motion of the symbiotic robot in a small space and improves the efficiency of offline programming of the symbiotic robot.

[0104] In step 5, the point matching algorithm based on viewpoint coupling is specifically:

[0105] First, a TCP coordinate system {A} of the hole-making end effector (3) is established, and a number of identification target points are attached to the hole-making end effector (3), and the three-dimensional coordinates of these points in the TCP coordinate system {A} are derived, and the Euclidean distance between each point and all other points is calculated and stored in a data container;

[0106] The visual camera (5) takes pictures of the points on the hole end effector (3) to obtain the three-dimensional coordinates of these points in the camera coordinate system {B}. The Euclidean distance between each point and all other points is also calculated and stored in the data container;

[0107] Since the relative distance between points remains unchanged in different scenes and different coordinate systems, by comparing the distance data containers in the TCP and camera scene coordinate systems, the point with the highest matching degree between the two is the same point, and the one-to-one correspondence between points in different scenes can be completed.

[0108] like Figure 6 As shown, the present invention guides the hole-making end effector 3 to adjust its posture through the visual camera 5, so that the hole-making end effector 3 reaches the specified posture and ensures the hole-making accuracy. The entire visual guidance process requires data preparation in the early stage, coordinate system conversion, obtaining the TCP posture of the end, and then fine-tuning according to the posture error, so that the hole-making collaborative robot 2 drives the hole-making end effector 3 to reach the required posture. The detailed process of the robot multi-level posture adjustment algorithm based on visual position coupling includes:

[0109] Step 1: First, data preparation is performed. The target points of the TCP coordinate system {A} of the hole-making end effector (3) and the target points of the camera coordinate system {B} are matched one by one using a point matching algorithm based on view-position coupling, and the current accurate position BT of the TCP in the camera coordinate system {B} is obtained;

[0110] Step 2: Set the current position of TCP in the camera coordinate system {B} B T is transformed to the base coordinate system {C} of the hole-making collaborative robot 2 as follows: C T.

[0111] Step 3: Read the current pose C After T, the target pose required by offline simulation is C T e Make a comparison.

[0112] Step 4: Calculate the current pose C T and target posture C T e The error is calculated by comparing the corresponding values ​​of the two attitude data sets item by item. The position error is the difference between the two 3D coordinates, and the attitude error is the difference between the two Euler angles.

[0113] Step 5: Set an error threshold ψ, and compare the calculated posture error Δ with the threshold ψ to determine whether the hole-making collaborative robot 2 has reached the target posture. If the errors of all six degrees of freedom are within the threshold range, it is considered that the hole-making collaborative robot 2 has reached the target posture.

[0114] Step 6: If the error Δ exceeds the threshold ψ, the error is input as feedback to the PID controller, which can divide the error Δ into multiple compensation amounts δ to improve the smoothness and accuracy of posture adjustment. The PID controller will calculate the compensation amount δ that needs to be adjusted step by step, and use the compensation amount δ to adjust the movement of the hole-making collaborative robot 2 through the control software. During the adjustment process, the hole-making collaborative robot 2 moves its position and changes its posture according to the feedback error information to gradually reduce the error.

[0115] Step 7: Iterative posture adjustment process, the hole-making collaborative robot 2 updates its current posture data after each adjustment C T', and recalculate the error Δ'. Through continuous feedback and adjustment, the hole-making collaborative robot 2 gradually approaches the target posture. C T e When the hole-making collaborative robot 2 detects that the errors of all degrees of freedom are less than the set threshold value ψ at a certain moment, it is considered that the hole-making collaborative robot 2 has successfully reached the target posture C T e , the posture adjustment process is completed.

[0116] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

[0117] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. A method for co-operating visually guided processing by two robots in a narrow space is implemented based on a system for co-operating visually guided processing by two robots in a narrow space, and is characterized by: The system comprises an AGV carrier (1), a hole-making collaborative robot (2), a hole-making end effector (3), a tooling frame (4), a visual camera (5), a visually guided robot (6) and a base (7); The hole-making collaborative robot (2) and the visual guidance robot (6) constitute a co-collaborative robot; the tooling frame (4) is used to fix and install the workpiece to be processed; the hole-making collaborative robot (2) is fixed on the AGV carrier (1) and arranged on one side of the workpiece to be processed, and a hole-making end effector (3) is installed at its end. The hole-making collaborative robot (2) can move the hole-making end effector (3) to a specified processing position inside the workpiece to be processed to perform hole-making processing; the visual guidance robot (6) is fixed on the base (7) and arranged on the other side of the workpiece to be processed, and a visual camera (5) is installed at its end. The visual guidance robot (6) can move the visual camera (5) to an appropriate position inside the workpiece to be processed to visually guide the hole-making end effector (3); The method comprises: Step 1: Build and adjust the work platform according to the different parts to be processed, and build the corresponding working environment in the offline programming software; Step 2: Analyze the positions of the holes to be drilled, determine the drilling sequence, and use the RRT algorithm based on multiple safety constraints to plan the motion path of the drilling collaborative robot (2) to complete drilling of all holes and the motion path of the visual guidance robot (6) to drive the visual camera (5) for visual guidance in a narrow space with the inner wall of the workpiece to be processed, another fusion robot and the visual observable range as the constraint targets; Step 3: The control software connects the collaborative hole-making robot (2) and the vision-guided robot (6) in a coordinated manner, activates the digital twin monitoring function, and completes the preparation work; Step 4: Control the collaborative robot to start the workflow, the visual guidance robot (6) reaches the camera where the camera can see the posture of the hole-making end effector (3) making the first hole, and the hole-making collaborative robot (2) drives the hole-making end effector (3) to the vicinity of the hole to be made; Step 5: After the hole-making collaborative robot (2) and the visual guidance robot (6) have reached their respective predetermined positions, the hole-making collaborative robot (2) is visually guided to adjust its position using a point matching algorithm based on visual position coupling and a robot multi-level attitude adjustment algorithm to ensure that it accurately reaches the predetermined position; Step 6: Open the air path of the suction cup of the hole-making end effector (3) so that it is adsorbed on the inner surface of the workpiece to be processed; Step 7: For the hole-making collaborative robot (2) whose posture has changed after the hole-making end effector (3) completes adsorption, a multi-level posture adjustment algorithm is used to guide and fine-tune the posture of the hole-making end effector (3) again. At this point, the collaborative posture adjustment of the hole-making collaborative robot (2) and the vision-guided robot (6) is completed; Step 8: Start the hole making process: open the suction cup extension and locking device (304) of the hole making end effector (3) to lock the suction cup so that it does not move up and down; Step 9: Open the air path above the pressure foot telescopic rod (307) to push out the contact pressure foot device (309) to the position where the hole is to be made; Step 10: Control the spindle feeding device (305) to start drilling holes. During this process, the feed depth D, rotation speed V and other parameters of the drill bit can be controlled; Step 11: After the hole making is completed, the spindle feeding device (305) is controlled to retract the spindle (306), the pressure foot telescopic rod (307) is controlled to retract the contact pressure foot device (309), the suction cup telescopic locking device (304) is controlled to unlock the suction cup, and the air is turned off. At this point, the single hole position hole making task is completed; Step 12: After completing the drilling of a hole position, the control software continues to control the hole collaborative robot (2) and the visual guidance robot (6) to move to the ready posture of the next hole position to be drilled, and re-executes the above-mentioned drilling process from step 5 to step 11; Step 13: After all the holes are made, the control software controls the hole collaborative robot (2) and the visual guidance robot (6) to move out the workpiece to be processed, and the system processing task is completed; The robot multi-level posture adjustment algorithm based on view-position coupling in step 5 is specifically as follows: Step 1: First, prepare the data. Use the point matching algorithm based on view-position coupling to match the target points of the TCP coordinate system {A} of the hole-making end effector (3) with the target points of the camera coordinate system {B} one by one, and obtain the current accurate position of the TCP in the camera coordinate system {B}. B T; Step 2: Set the current position of TCP in the camera coordinate system {B} B T is transformed to the base coordinate system {C} of the hole-making collaborative robot (2) as follows: C T; Step 3: Read the current pose C After T, the target pose required by offline simulation is C T e Make comparisons; Step 4: Calculate the current pose C T and target posture C T e The error Δ between them, where the error is calculated by comparing the corresponding values ​​of the two attitude data sets item by item. The position error is the difference between the two three-dimensional coordinates, and the attitude error is the difference between the two Euler angles; Step 5: By comparing the calculated posture error Δ with the error threshold ψ, it is determined whether the hole-making collaborative robot (2) has reached the target posture. If the errors of all six degrees of freedom are within the threshold range, it is considered that the hole-making collaborative robot (2) has reached the target posture. Step 6: If the error Δ exceeds the error threshold ψ, the error is input as feedback to the PID controller, and the error Δ is divided into multiple compensation amounts δ to improve the smoothness and accuracy of the posture adjustment. The PID controller calculates the compensation amount δ that needs to be adjusted step by step, and adjusts the movement of the hole-making collaborative robot (2) with the compensation amount δ through the control software. During the adjustment process, the hole-making collaborative robot (2) moves its position and changes its posture according to the feedback error information to gradually reduce the error; Step 7: Iterative posture adjustment process, the hole-making collaborative robot (2) updates its current posture data after each adjustment C T′, and recalculate the error Δ′. Through continuous feedback and adjustment, the hole-making collaborative robot (2) gradually approaches the target posture. C T e When the hole-making collaborative robot (2) detects that the errors of all degrees of freedom are less than the set threshold value ψ at a certain moment, it is considered that the hole-making collaborative robot (2) has successfully reached the target posture C T e , the posture adjustment process is completed.

2. The method for dual-robot collaborative visual-guided processing in a narrow space according to claim 1 is characterized in that: The hole-making end effector (3) comprises a ball-jointed suction cup (301), a suction cup support rod (302), a flange, a suction cup telescopic locking device (304), a spindle feeding device (305), a spindle (306), a pressure foot telescopic rod (307), a mounting plate (308) and a contact pressure foot device (309); the flange is connected to the end of the hole-making collaborative robot (2); the mounting plate (308) is located in the middle of the hole-making end effector (3), and four suction cup support rods (302), a pressure foot telescopic rod (307), a spindle feeding device (305) and a central opening for placing the spindle (306) are fixed thereon; the suction cup support rod (302) is fixed by hinge connection, and its support angle can be adjusted to adapt to the complex environment inside the narrow space of the workpiece to be processed; there is a The ball-jointed suction cup (301) can movably adjust the suction cup angle, and the upper end is connected to an air pipe, and the air circuit is opened to control the suction of the ball-jointed suction cup (301), and a suction cup telescopic locking device (304) is installed, and the locking and unlocking of the ball-jointed suction cup (301) are controlled; the pressure foot telescopic rod (307) is fixed on the mounting plate (308) at the top and connected to the air circuit, and is connected to the contact pressure foot device (309) at the bottom to control the extension and retraction of the pressure foot; three normal sensors are installed on the contact pressure foot device (309), which can detect the normal direction of the pressure foot; the spindle feeding device (305) is connected to the spindle (306), and the up and down feeding of the spindle (306) is controlled by the movement of the worm gear; a drill is installed at the bottom of the spindle (306) for making holes and countersinking.

3. The method for dual-robot collaborative visual-guided processing in a narrow space according to claim 1 is characterized in that: The hole making end effector (3) can replace the spindle (306) bit to achieve hole making and countersinking of different hole diameters. The hole diameter accuracy of the hole making is higher than H8, the normal accuracy is ≤0.5°, and the countersinking accuracy is ≤0.03mm.

4. The method for dual-robot collaborative visual-guided processing in a narrow space according to claim 1 is characterized in that: The bottom of the tooling frame (4) is provided with 6 pulleys, which can drive the tooling frame (4) to move freely; each pulley can be locked by a bolt to fix the tooling frame (4) at a specified position; the upper support surface of the tooling frame (4) can move up and down to adjust the vertical position of the workpiece to be processed thereon; there are 9 groups of movable workpiece support rods on the upper support surface, and their positions can be adaptively adjusted according to the shape of the workpiece to be processed to ensure that the workpiece to be processed is installed securely and stably.

5. The method for dual-robot collaborative visual-guided processing in a narrow space according to claim 1 is characterized in that: The visual camera (5) is a binocular visual camera (501), which is connected to a flange through a camera connecting rod (502) and fixed to the end of the vision-guided robot (6).

6. The method for dual-robot collaborative visual-guided processing in a narrow space according to claim 1 is characterized in that: The hole-making collaborative robot (2) and the visual guidance robot (6) first plan their initial motion paths through offline programming, and use the dual-robot symbiotic collaborative hole-making software to control their movements. During the working process, the digital twin technology is used to monitor the motion postures of the dual robots in real time. The visual guidance robot (6) provides real-time on-site situation data to guide the hole-making collaborative robot (2) to adjust its posture so that it reaches the target posture. The position accuracy of the two robots is ≤0.2mm, the posture accuracy is ≤0.1°, and the trajectory accuracy is ≤0.5mm. The dual-robot symbiotic collaborative hole-making software uses the RRT algorithm to optimize the path and obtain a collision-free trajectory for the hole-making collaborative robot (2) and the visual guidance robot (6) to move inside the workpiece to be processed.

7. The method for dual-robot collaborative visual-guided processing in a narrow space according to claim 1 is characterized in that: The error handling strategy for the machining process is as follows: during the operation of the collaborative robot, if the drill bit is damaged or the end fails when drilling a hole, first press the emergency stop button of the robot system and interrupt the program, then move the robot a short safe distance, record its current posture, and use the offline programming planning algorithm to plan the path from the current posture back to the initial position outside the workpiece to be processed, and then use the control software to control the robot to move out of the narrow space, and then replace or repair it.

8. The method for dual-robot collaborative visual-guided processing in a narrow space according to claim 1 is characterized in that: The RRT algorithm based on multiple safety constraints in step 2 is specifically as follows: Based on the RRT algorithm, first determine the starting and ending postures of the robot path planning, set the robot workspace range and the number of algorithm iterations k; In the robot axis space, take the starting point as the first parent node N0, and randomly select the robot axis angles to obtain the robot random point N r ; For the semi-enclosed environment where the co-integrated robot is in a narrow space inside the workpiece to be processed, a certain probability is set to take a random point N r is the end pose node N goal , and this probability is 15% to 30%; For a random point N in the robot axis space rand , using the Euclidean distance formula Calculate the distance between it and each stored node, and find the closest node N as the parent node of the random point near ; At this random point N rand With its parent node N near Take a certain step length on the direction vector as the new node N new ; Using the collision detection function in the offline programming software, the inner wall of the workpiece to be processed and another co-integrated robot are regarded as multi-target obstacles to check whether the robot can near To the new node N new Will there be a collision during the movement? If a collision occurs, a random point is reselected and a new iteration is entered. If no collision occurs, the node is stored and recorded as N. m ; Each node N m After finding it, we need to determine whether we have reached the end point. If the distance between the new node position and the end point position is less than one step, we consider that we have reached the end point and found the path. If the target point cannot be obtained in the workspace of the collaborative robot, it means that the collaborative robot needs to adjust its posture at a large angle when planning to the target point. In this case, the collaborative robot needs to be moved out of the workpiece to be processed, and the collaborative robot needs to complete the large-angle posture transformation outside the workpiece, and then enter the workpiece to complete the complete path planning; If the target point is outside the workspace of the collaborative robot, it means that the collaborative robot is unreachable under the current deployment state. The tooling frame (4) and the workpiece to be processed are turned around, so that the hole-making collaborative robot (2) enters from another opening of the workpiece to be processed and continues to complete the path planning. The found path is smoothed using spline interpolation, ultimately forming a collision-free path for the robot and completing the robot's path planning.

9. The method for dual-robot collaborative visual-guided processing in a narrow space according to claim 1 is characterized in that: In step 5, the point matching algorithm based on viewpoint coupling is specifically: First, a TCP coordinate system {A} of the hole-making end effector (3) is established, and a number of identification target points are attached to the hole-making end effector (3), and the three-dimensional coordinates of these points in the TCP coordinate system {A} are derived, and the Euclidean distance between each point and all other points is calculated and stored in a data container; The visual camera (5) takes pictures of the points on the hole end effector (3) to obtain the three-dimensional coordinates of these points in the camera coordinate system {B}. The Euclidean distance between each point and all other points is also calculated and stored in the data container; By comparing the distance data containers in the TCP and camera scene coordinate systems, the point with the highest matching degree between the two is the same point, and the one-to-one correspondence between points in different scenes can be completed.

Images