Robot control method and system for industrial pipeline

Abstract

The invention discloses a robot control method and system for an industrial pipeline, and relates to the technical field of robot control. The method comprises the steps that a flange plate is installed on an end effector of the robot, a binocular vision camera is installed in the center of the front end of the flange plate, and four point laser displacement sensors are symmetrically arranged on the flange plate at different heights; the binocular vision camera collects a pipeline image; the upper computer determines an initial pose control result of the robot according to the pipeline image so as to control the robot to move preliminarily; the four point laser displacement sensors are used for collecting spatial displacement data of the pipeline at different height points after the robot moves preliminarily; and the upper computer determines a target pose control result of the end effector of the robot through the spatial displacement data so as to control the end effector of the robot to carry out pose adjustment. According to the method, accurate control over the axial parallelism of the robot end effector and the pipeline can be achieved, and therefore the pipeline assembling accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the field of robot control technology, and in particular to a robot control method and system for industrial pipelines. Background Art

[0002] The use of robots to realize automated operation scenarios in industrial pipelines has become a widely concerned and important issue, and is gradually developing from emerging technology applications to widespread popularization.

[0003] In curved pipe applications, a robot must align its end effector (such as a gripper) with the cylindrical pipe end face for assembly, docking, or welding. Existing techniques typically measure the robot's end effector's pose to achieve path control and position calibration.

[0004] However, during the docking process of curved objects such as pipes, existing technologies have failed to effectively solve the problem of precise control of the axial parallelism between the robot end effector and the pipe, resulting in errors in pipe assembly. Summary of the Invention

[0005] Based on this, it is necessary to provide a robot control method and system for industrial pipelines to address the above technical problems. This method can achieve precise control of the axial parallelism between the robot end effector and the pipeline, thereby improving the accuracy of pipeline assembly.

[0006] The present invention adopts the following technical solutions: The present invention provides a robot control method for industrial pipelines. A flange is installed on the end effector of the robot, a binocular vision camera is installed at the front center of the flange, and four point laser displacement sensors are symmetrically arranged on the flange with different heights. The method includes: Collect pipeline images of industrial pipelines through binocular vision cameras; Determine the pipeline position and the angle between the robot and the industrial pipeline based on the pipeline image, and determine the initial position control result of the robot based on the pipeline position and the angle between the robot and the industrial pipeline to control the robot to perform preliminary movement; After the robot performs preliminary movements, the spatial displacement data of the industrial pipeline at different heights are collected by four point laser displacement sensors; The angular deviation and position offset between the robot end effector and the axis of the industrial pipeline are determined through spatial displacement data. Based on the angular deviation and position offset between the robot end effector and the axis of the industrial pipeline, the target posture control result of the robot end effector is determined to control the robot end effector to adjust its posture.

[0007] Optionally, the pipeline image includes a left image frame and a right image frame, and determining the pipeline position and the angle between the robot and the industrial pipeline according to the pipeline image includes: The left and right image frames are input into the end-to-end deep network PSMNet to obtain a pixel-level disparity map; Obtain a depth map based on the pixel-level disparity map, the focal length of the binocular vision camera, and the baseline length; The depth map is fed into the image semantic segmentation model to obtain the pipeline target bounding box; the image semantic segmentation model is built based on the lightweight network YOLOv8-Seg; Determine the pipeline location based on the pipeline target bounding box; Calculate the three-dimensional coordinates of the center point of the pipeline based on the intrinsic parameter matrix of the binocular vision camera, the pipeline position and its corresponding parallax value; The angle between the robot and the industrial pipeline is determined based on the three-dimensional coordinates of the pipeline center point.

[0008] Optionally, the pipeline position includes the three-dimensional coordinates of the pipeline center point; and determining the initial posture control result of the robot according to the pipeline position and the angle between the robot and the industrial pipeline includes: Based on the 3D coordinates of the pipeline center point and the angle between the robot and the industrial pipeline, the robot base coordinate system is aligned with the pipeline coordinate system to generate the linear motion path parameters from the robot's current position to the industrial pipeline. The path parameters include the translation and rotation of the end effector's target position. The translation and rotation of the target posture are converted into angle instructions of each joint of the end effector to form the initial posture control result.

[0009] Optionally, the spatial displacement data includes position coordinates of four spatial points; determining the angular deviation and position offset between the robot end effector and the axis of the industrial pipeline through the spatial displacement data includes: The position coordinates of the four spatial points are fitted using the least squares method to obtain the normal vector of the pipeline axis; Determine the angular deviation between the robot end effector and the industrial pipeline axis according to the direction vector of the robot end effector and the normal vector of the industrial pipeline axis; According to the center point coordinates of the pipeline and the coordinate points of the robot's end effector, the position offset between the robot's end effector and the axis of the industrial pipeline is determined.

[0010] Optionally, the position coordinates of the four spatial points are fitted using the least squares method to obtain the normal vector of the industrial pipeline axis, including: Construct the objective function based on the three-dimensional space plane equation; According to the position coordinates of the four spatial points, the least square method is used to solve the objective function and determine the values ​​of the plane parameters in the objective function; The value of the plane parameter in the objective function is determined as the normal vector of the industrial pipeline axis.

[0011] Optionally, the objective function is: ; in,( , , ) indicates the i The position coordinates of a spatial point, 、 、 and All are variables.

[0012] Optionally, the angular deviation between the robot end effector and the pipeline axis The calculation formula is: ; in, represents the normal vector of the industrial pipeline axis, Represents the orientation vector of the robot's end effector.

[0013] Optionally, determining a target posture control result of the robot end effector according to an angular deviation and a position offset between the robot end effector and the axis of the industrial pipeline includes: According to the angular deviation and position offset between the robot's end effector and the axis of the industrial pipeline, the joint angle that needs to be adjusted of the robot's end effector is determined through the inverse kinematics algorithm.

[0014] The present invention provides a robot control system for industrial pipelines, the system comprising: a robot and a host computer; a flange is mounted on the robot end effector, a binocular vision camera is mounted at the front center of the flange, and four point laser displacement sensors are symmetrically arranged on the flange with different heights; Binocular vision camera, used to collect pipeline images of industrial pipelines and send the pipeline images to the host computer; The host computer is used to determine the pipeline position and the angle between the robot and the industrial pipeline based on the pipeline image, and determine the initial posture control result of the robot based on the pipeline position and the angle between the robot and the industrial pipeline to control the robot to perform preliminary movement; Four point laser displacement sensors are used to collect spatial displacement data of industrial pipelines at different heights after the robot performs initial movement, and send the spatial displacement data to the host computer; The host computer is used to determine the angular deviation and position offset between the robot end effector and the axis of the industrial pipeline through spatial displacement data, and determine the target posture control result of the robot end effector based on the angular deviation and position offset between the robot end effector and the axis of the industrial pipeline, so as to control the robot end effector to adjust its posture.

[0015] Optionally, the system further comprises a programmable logic controller, the host computer communicates with the programmable logic controller, and the robot's actuator communicates with the programmable logic controller at high speed and in two directions; The programmable logic controller is used to receive the pipeline image sent by the binocular vision camera and upload it to the host computer; receive the initial posture control result sent by the host computer and send it to the robot's actuator; receive the spatial displacement data sent by the four-point laser displacement sensors and upload it to the host computer; receive the target posture control result sent by the host computer and send it to the robot's actuator.

[0016] The present invention provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned XX method is implemented.

[0017] The present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned XX method when executing the program.

[0018] At least one of the above technical solutions adopted by the present invention can achieve the following beneficial effects: In this invention, a binocular vision camera is first integrated into the flange. The robot's initial motion is then controlled based on the pipeline image captured by the binocular vision camera, enabling the robot to quickly approach the target pipeline. Four point laser displacement sensors are fixed to either side of the end effector in a symmetrical manner at different heights. When the robot approaches the pipeline, they collect spatial displacement data from the upper and lower points on each end of the pipeline. By calculating the angular deviation and positional offset between the robot end effector and the pipeline axis, the robot's end effector's posture is adjusted to achieve precise parallel alignment with the pipeline axis. In this way, the binocular vision camera and point laser displacement sensors form a closed-loop collaborative mechanism of "coarse positioning + fine positioning," enabling precise control of the robot end effector's parallelism with the pipeline axis, thereby improving the accuracy of pipeline assembly. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0020] Figure 1 A schematic structural diagram of a robot control system for industrial pipelines provided by the present invention; Figure 2 A schematic diagram of the overall architecture of a robot control system for industrial pipelines provided by the present invention; Figure 3 A schematic flow chart of a robot control method for industrial pipelines provided by the present invention; Figure 4 This is the network structure diagram of YOLOv8-Seg. DETAILED DESCRIPTION

[0021] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present invention and corresponding drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0022] In scenarios like automated assembly, welding, and handling of industrial pipelines, robots must achieve high-precision docking of end effectors (such as claws and grippers) with the pipe end faces, ensuring excellent axial parallelism to avoid deflection, assembly failure, and stress concentration. However, faced with challenges such as complex working conditions, limited space, and uncertain pipe posture, traditional technologies have significant shortcomings in positioning accuracy, response speed, and environmental adaptability.

[0023] 1. Limitations of existing technologies (1) Systems that rely solely on robot body posture compensation and ignore the relative posture control of the target object.

[0024] Most existing methods achieve accuracy improvement solely by measuring and compensating the robot's end-effector pose, ignoring the spatial variations of external targets (such as pipes), making it difficult to meet the requirements for relative pose control in actual assembly. Specifically, most existing technologies focus on using sensors to calibrate the robot's pose and compensate for errors, addressing the robot's inherent motion accuracy. However, these methods often only optimize the path control and position calibration of the robot's end-effector, neglecting the relative pose control between the robot and the external target (such as a pipe). Especially during the docking of curved objects such as pipes, traditional technologies fail to effectively control the axial parallelism between the robot's end-effector and the target pipe, resulting in assembly errors and insufficient execution accuracy.

[0025] (2) The visual positioning system relies on plane calibration and cannot adapt to complex curved surface targets.

[0026] Many existing positioning technologies use standard flat calibration plates or external vision systems for positioning and attitude measurement. This approach is effective for regular flat surfaces, but for axisymmetric and curved targets like pipelines, the calibration plates cannot provide an effective reference, resulting in a significant decrease in positioning accuracy. This is particularly true in complex industrial environments, where it cannot adapt to frequently changing dimensions or irregularly shaped targets, resulting in limited deployment flexibility and unstable accuracy.

[0027] (3) Single-point vision or laser ranging methods have insufficient information dimensions, that is, the sensor layout and error compensation mechanism are insufficient, making it difficult to cope with complex working conditions.

[0028] Single-point vision recognition or laser point measurement systems are subject to limitations such as ambient lighting, occlusion, and surface reflections. Furthermore, the measurement data lacks complete spatial structural information, making it difficult to support closed-loop adjustments between the robot and the target in dynamic environments. Furthermore, single-point or small-point measurement limits their positioning accuracy on complex curved surfaces, such as pipelines. Furthermore, error compensation algorithms are mostly based on simplified models and fail to fully utilize the spatial information from multi-point measurement data. This results in low error compensation accuracy and is unable to meet the requirements of high-precision dynamic positioning, especially in the precise docking tasks between robots and pipelines.

[0029] 2. Core needs of industrial sites Industrial sites urgently need an integrated, modular positioning system with the following functions: (1) Able to identify pipes of various sizes and postures, and complete rapid coarse positioning and precise docking; (2) Get rid of the dependence on calibration plates and stable geometric benchmarks and adapt to irregular surface targets; (3) Achieve closed-loop control of the entire process from initial perception to posture adjustment; (4) It has the performance characteristics of anti-interference, high robustness and strong real-time performance.

[0030] Based on this, the present invention provides a robot control method and system for industrial pipelines, which is suitable for automated operation scenarios of pipeline targets. By integrating binocular camera and laser sensor technology, it not only solves the problems of low positioning accuracy of robots in complex environments and low axial parallelism between the robot end effector and the pipeline in the existing technology, but also solves problems such as slow response speed and poor system adaptability. Specifically, it includes the following aspects: 1. The robot has poor positioning stability and insufficient accuracy in complex environments Traditional positioning methods often rely on artificial vision or a single sensor. These methods often suffer from recognition failures and large errors in real-world conditions, such as limited space, unstable pipe positions, or reflective surfaces. This makes it difficult to meet the dual requirements of stability and precision in industrial settings. This new approach integrates a visual camera on the robot flange to obtain the approximate spatial position of the pipe, achieving coarse positioning. This is then combined with four laser sensors for fine-tuning, creating a multimodal perception collaborative link combining coarse and fine positioning, thereby improving positioning robustness and accuracy in complex environments.

[0031] 2. Aiming at the problem of high-precision axial parallel positioning of curved pipelines, the robot end effector and the axial parallelism control problem of the pipeline are accurately solved.

[0032] In practical applications, the robot needs to align the end effector (such as a claw) with the end face of a cylindrical pipe for assembly, docking or welding operations. Since the outer surface of the pipe is a typical curved structure and its installation posture is uncertain, it is very challenging to achieve precise alignment of the robot end and the axial direction of the pipe. Different from the traditional approach of "compensating for the posture error of the robot body", the present invention focuses on the posture relationship control between the robot end and the external target (pipeline), rather than the traditional approach of only optimizing the robot body error compensation. It avoids the dependence on the robot kinematic modeling error and encoder accuracy at the system level, and improves the versatility and accuracy of axial parallel control. Specifically, by combining a binocular camera and a point laser displacement sensor, precise control of the axial parallelism of the pipeline and the robot end effector is achieved, enabling the robot to maintain high-precision parallel docking in the docking task of complex curved workpieces (such as pipes), solving the accuracy problem that the existing technology cannot effectively deal with.

[0033] 3. The problem of achieving spatial attitude solution without the help of external calibration plates, that is, avoiding the use of calibration plates and achieving calibration-free layout.

[0034] Most current surface target positioning methods rely on standard flat plates or calibration blocks to establish a spatial reference. This is limited by the ambient space and the frequency of target changes, resulting in complex field deployment and insufficient flexibility. This new method uses four symmetrically arranged point laser displacement sensors to directly measure the relative height differences at different locations on the pipeline's outer surface, thereby establishing a spatial reference plane. Through real-time measurement and data fusion from these point laser displacement sensors, the relative position and posture between the pipeline and the robot can be autonomously calculated without the need for external calibration objects or additional visual markers. This significantly improves the system's deployment flexibility and adaptability, enabling rapid deployment and adjustment, especially during field operations.

[0035] 4. Sensor structure adaptability and surface target recognition robustness issues Compared to the existing "coplanar" or "single-sided arrangement" sensor designs, this system installs four point laser displacement sensors on the end of the robot in a bilaterally symmetrical structure with different heights in the front and back, so that they can measure the spatial height information of the corresponding points above and below the end of the pipe respectively, and build a posture model of the target through precise multi-point measurement and advanced error compensation algorithms. This arrangement structure has good adaptability to pipes of different diameters, materials, surface characteristics and installation angles, significantly improving recognition accuracy and system robustness. This enables the present invention to perform high-precision positioning of curved targets such as pipes, and to adapt in real time to changes in the posture of the pipe under different working conditions. Through data fusion and dynamic error compensation, it is ensured that the robot's end effector always maintains precise positioning and correct posture under dynamic working conditions.

[0036] 5. Robot actuator control closed-loop response lag problem In traditional systems, sensor data often requires manual analysis or offline processing before operators issue control commands, resulting in significant feedback delays and low response efficiency. This invention establishes a Profinet high-speed communication link between a PLC controller and a robot controller, enabling real-time transmission and fusion processing of laser and visual sensor data. Dynamic correction and closed-loop control are then performed based on posture errors, effectively eliminating system response lag and improving dynamic tracking capabilities during pipeline positioning.

[0037] 6. The system is not adaptable enough to changes in pipe size and posture In real-world industrial scenarios, pipelines often experience changes in posture due to process errors, thermal deformation, installation deviations, and other factors. Traditional systems rely on fixed fixtures or manual calibration, lacking dynamic perception and adaptive adjustment capabilities. This invention, based on multi-point laser ranging data and a posture calculation algorithm, constructs a spatial reference plane in real time and adjusts the robot's end-effector posture through inverse kinematics. This allows for rapid alignment and adaptive assembly of non-standard pipeline conditions, demonstrating excellent adaptability to industrial sites.

[0038] The technical solutions provided by various embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0039] Figure 1 This is a flowchart for implementing a robot control system for industrial pipelines in the present invention. The system includes: a robot and a host computer; a flange is installed on the robot's end effector, a binocular vision camera is installed at the front center of the flange, and four point laser displacement sensors are arranged symmetrically on the flange with different heights, and the four point laser displacement sensors are located on both sides of the robot's end effector.

[0040] Binocular vision camera is used to collect pipeline images of industrial pipelines and send the pipeline images to the host computer.

[0041] The host computer is used to determine the pipeline position and the angle between the robot and the industrial pipeline based on the pipeline image, and to determine the initial posture control result of the robot based on the pipeline position and the angle between the robot and the industrial pipeline to control the robot to perform preliminary movement.

[0042] Four point laser displacement sensors are used to collect spatial displacement data of industrial pipelines at different heights after the robot performs initial movement, and send the spatial displacement data to the host computer.

[0043] The host computer is used to determine the angular deviation and position offset between the robot end effector and the axis of the industrial pipeline through spatial displacement data, and determine the target posture control result of the robot end effector based on the angular deviation and position offset between the robot end effector and the axis of the industrial pipeline, so as to control the robot end effector to adjust its posture.

[0044] The flange is mounted directly on the robot's end effector (the tool interface at the end of the sixth axis of the robotic arm) and serves as a mounting platform for sensor integration. "End" here specifically refers to the physical end of the robot's end effector, the mechanical interface where the robot directly interacts with the target object (pipeline) when completing a task. The flange is rigidly connected to the end effector, ensuring that the sensor module is fully synchronized with the robot's end motion.

[0045] Binocular Vision Camera Installation Position: The binocular vision camera is fixed to the front center of the flange (facing the pipeline) to ensure that its field of view covers the target area of ​​the pipeline. The camera's optical axis is parallel to the default operating direction of the robot end effector.

[0046] The system also includes a programmable logic controller (PLC), which communicates with the host computer and enables high-speed, two-way communication between the robot and the PLC. The PLC is used to receive pipeline images sent by the binocular vision camera and upload them to the host computer, receive initial posture control results sent by the host computer, and send them to the robot; receive spatial displacement data sent by four point laser displacement sensors, upload them to the host computer, receive target posture control results sent by the host computer, and send them to the robot.

[0047] Among them, the robot is a six-axis industrial robot.

[0048] The overall architecture of the system is as follows Figure 2 As shown in the figure, from left to right, it consists of a host computer, PLC, robot, binocular vision system, laser sensor module, and positioning algorithm. The system builds an intelligent positioning system that deeply integrates perception and control through the five logical pathways of perception-transmission-computation-control-execution.

[0049] The system is divided into two main collaborative paths: the upper path (coarse positioning path): completed by the binocular vision system and the robot in collaboration to identify the approximate spatial position of the pipeline and output the initial pose estimation result; the lower path (fine positioning path): the point laser displacement sensor carried by the end flange of the robot collaborates with the robot control system to collect multi-point spatial distances and achieve high-precision axis recognition and posture fine-tuning.

[0050] exist Figure 2 In the system, the modules are connected through seven key interaction points, which are as follows: Label 1: The host computer communicates with the PLC through the network port to achieve system parameter setting, command issuance and status monitoring.

[0051] Label 2: The PLC establishes real-time two-way communication with the robot via the Profinet high-speed industrial protocol, transmitting positioning data and control instructions.

[0052] Label 3: Establish a coordinate conversion relationship between the robot and the binocular vision system (Robot-Vision Calibration) to achieve accurate mapping of the visual coordinates to the robot coordinate system.

[0053] Label 4: The flange at the end of the robot serves as the execution platform, and four point laser displacement sensors are installed to form the basis for precise positioning execution.

[0054] Label 5: The binocular vision system is responsible for initial recognition and coarse positioning, and uses AI algorithms to extract the approximate position, direction, and shape characteristics of the pipeline.

[0055] Label 6: The point laser displacement sensor is arranged symmetrically at different heights to perform real-time distance measurement on multiple key positions at the upper and lower ends of the pipeline.

[0056] Label 7: Based on the laser sensor data, the axis positioning and posture fitting algorithm is run to output a high-precision axial posture solution result to guide the robot to perform end-point posture adjustment.

[0057] The system modules specifically include: Host computer (human-machine interface and system dispatching center): The host computer communicates with the PLC through industrial Ethernet to realize process logic management, command issuance, status monitoring and data recording. It is the human-machine interaction and dispatching control center of the system.

[0058] PLC (Logic Control and Communication Hub): Using the Profinet high-speed communication protocol, PLC serves as the convergence center for information and instructions, responsible for connecting the perception end and the execution end, forwarding binocular vision and laser sensor data in real time, and controlling the robot's action logic.

[0059] Robot (actuator): As the core actuator, the robot completes the tasks of aligning, approaching, and docking targets within the pipeline. Its end is integrated with a custom-designed flange for mounting a laser sensor module and responding to precise posture adjustment commands output by the perception system.

[0060] Binocular vision system (coarse positioning perception mode): As a global perception mode, the binocular vision system captures the large-scale position information of the pipeline, performs deep image processing through AI recognition algorithms, achieves preliminary perception of the pipeline contour, central axis and spatial posture, and provides coarse positioning guidance for the robot.

[0061] AI pipeline recognition algorithm (visual semantic modeling): Based on visual acquisition images, combined with deep neural network algorithms, it can realize rapid recognition and spatial reconstruction of pipeline structures and output three-dimensional coordinate information ( X , Y , Z ) provides the robot with posture guidance and preliminary plug-in operations.

[0062] Self-designed terminal flange (multi-sensor integration platform): The flange serves as the terminal device, used to securely deploy four point laser displacement sensors arranged at different heights, ensuring stable sampling and redundant acquisition of multi-point data. It is the hardware foundation for achieving multimodal fine perception.

[0063] Point laser displacement sensor array (precise positioning perception mode): Four laser sensors are arranged in an array of different heights around the pipeline to obtain spatial displacement data at different heights of the pipeline, build a multi-point three-dimensional geometric perception model, and achieve high-precision fitting and posture refinement of the pipeline axis.

[0064] Axis positioning algorithm (spatial fitting and deviation solution engine): Based on multi-point data from laser sensors, a spatial fitting model (least squares plane, normal vector calculation, etc.) is constructed to calculate the angular deviation and position offset between the robot end effector and the pipeline axis in real time, driving the robot end to make dynamic adjustments.

[0065] The present invention also provides a robot control method for industrial pipelines, which is applied to the above system, such as Figure 3 As shown, the method includes the following steps: S301, collecting a pipeline image of the industrial pipeline through a binocular vision camera; determining the pipeline position and the angle between the robot and the industrial pipeline based on the pipeline image; and determining the initial position control result of the robot based on the pipeline position and the angle between the robot and the industrial pipeline to control the robot to perform preliminary movement.

[0066] Optionally, the pipeline image includes a left image frame and a right image frame, and the pipeline position and the angle between the robot and the industrial pipeline are determined based on the pipeline image, including: inputting the left image frame and the right image frame into the end-to-end deep network PSMNet to obtain a pixel-level disparity map; obtaining a depth map based on the pixel-level disparity map, the focal length and baseline length of the binocular vision camera; inputting the depth map into the image semantic segmentation model to obtain a pipeline target bounding box; the image semantic segmentation model is constructed based on the lightweight network YOLOv8-Seg; determining the pipeline position based on the pipeline target bounding box; calculating the three-dimensional coordinates of the pipeline center point based on the intrinsic parameter matrix of the binocular vision camera, the pipeline position and its corresponding disparity value; and determining the angle between the robot and the industrial pipeline based on the three-dimensional coordinates of the pipeline center point.

[0067] Specifically, the method includes the following steps: (1) Image input and disparity calculation: Left image frame: I_L( x , y ), right image frame: I_R( x , y ), Camera parameters: focal length f , baseline length B and the internal parameter matrix K .

[0068] The end-to-end deep network PSMNet calculates the pixel-level disparity map D( x , y ), using binocular triangulation principle to calculate the depth map : .

[0069] (2) Pipeline target identification: Figure 4 As shown, Figure 4The network structure diagram of YOLOv8-Seg, where YOLOv8-Seg includes a backbone network (Backbone) and a head network (Head). The CBS module in the figure includes a convolution layer (Convolution), a batch normalization layer (Batch Normalization, BN) and a SiLU activation function connected in series: C2f module: a cross-stage local fusion module. It divides the input features into two parts, one part is directly passed, and the other part is spliced ​​and fused with the directly passed part after multiple convolution operations. In this way, the feature extraction capability is enhanced while reducing the amount of calculation; SPPF module: the Spatial Pyramid Pooling Fast module; Upsample represents the upsampling operation, Concat represents the splicing operation, and Segment represents instance segmentation. Based on the previously extracted and fused feature maps, the final instance segmentation result is output, for example, the predicted target category and the corresponding mask, to achieve accurate segmentation of different instances in the image. The depth map is input into the image semantic segmentation model to obtain the pipeline target bounding box bbox = [ x min , y min , x max , y max ] and semantic mask M( x , y ).

[0070] Specifically, the semantic mask M( x , y ) is a binary matrix of the same size as the image, marking whether each pixel belongs to the pipeline target (the pipeline area is 1 and the background is 0), realizing pixel-level pipeline contour segmentation.

[0071] The physical meaning of semantic mask: M( x , y ) represents the precise pixel-level area of ​​the pipeline target in the depth map, which is used to extract the pipeline edge contour point set for subsequent three-dimensional direction calculation (such as principal component analysis fitting the pipeline main axis direction), and combined with the depth map Z( x , y ), the pixels of the pipeline area are mapped into a three-dimensional point cloud, which is used to calculate the coordinates of the pipeline center point ( X , Y , Z ).

[0072] YOLOv8-Seg generates M( x , y): Extract features from the input depth map and output the bounding box and segmentation mask of the pipeline target. The final semantic mask M is obtained by mask thresholding (such as Sigmoid activation + 0.5 binarization). x , y ).

[0073] (3) Three-dimensional coordinate recovery (image coordinates → space coordinates) The pipeline position can include the center point of the image, and the image center point is calculated ( x c , y c ): .

[0074] Three-dimensional coordinate recovery (image coordinates → space coordinates): According to the intrinsic parameter matrix K of the binocular vision camera, the image center point ( x c , y c ) and its corresponding disparity value D( , )Calculate the three-dimensional coordinates of the center point of the pipeline ( X , Y , Z ). Deep recovery: ; Pixel coordinate projection: The internal parameter matrix K includes the principal point ( c x , c y ) and pixel size ratios f x and f y : .

[0075] (4) Rough direction estimation Extract the edge contour point set from the mask M(x, y) and use linear fitting or principal component analysis (PCA) to obtain the two-dimensional principal axis vector: .

[0076] Output the angle between the robot and the industrial pipeline: .

[0077] in, The unit vector representing the two-dimensional principal axis is used to describe the main extension direction of the pipeline on the image plane. Z ( x , y ), which can be converted into a pipeline direction vector in three-dimensional space.

[0078] Specifically, The methods for obtaining include: In one embodiment, principal component analysis is used to obtain: Input: mask M( x , y )=1.

[0079] Steps: Calculate the covariance matrix of pixel coordinates, perform eigenvalue decomposition on the covariance matrix, and the eigenvector corresponding to the maximum eigenvalue is the principal axis direction ( , ).

[0080] In another embodiment, the following is obtained by linear fitting: Input: Mask M ( x , y )=1, steps: Assuming that the pipeline contour is approximately a straight line, use the least squares method to fit the straight line equation y = kx + b .

[0081] The main axis direction vector is v=(1, k ), after normalization, we get .

[0082] In one embodiment, the pipeline position includes the three-dimensional coordinates of the pipeline center point; the initial posture control result of the robot is determined based on the pipeline position and the angle between the robot and the industrial pipeline, including: based on the three-dimensional coordinates of the pipeline center point and the angle between the robot and the industrial pipeline, the robot base coordinate system is aligned with the pipeline coordinate system, and the linear motion path parameters from the current position of the robot to the industrial pipeline are generated; the path parameters include the translation and rotation of the target posture of the end effector; the translation and rotation of the target posture are converted into angle instructions for each joint of the end effector to form the initial posture control result.

[0083] Specifically, the three-dimensional coordinates of the pipeline center point calculated by the binocular vision system ( X , Y , Z ) and the angle θ between the robot and the pipe, align the robot base coordinate system with the pipe coordinate system, and generate a linear motion path from the robot's current position to the rough positioning point at the near end of the pipe. The path parameters include the translation of the end effector's target pose ΔP = [ X , Y , Z ]^T and the rotation amount ΔR = R_z(θ), and the motion mode adopts joint space interpolation or Cartesian space linear motion.

[0084] Initial posture control result generation: The target posture (ΔP, ΔR) is converted into joint angle commands through the robot kinematics forward solution to form the initial posture control result. The calculation formula is:

[0085] Among them, IK represents the inverse kinematics solution function, is the joint angle vector.

[0086] Safety constraints and collision detection: Robot workspace limitations and joint velocity / acceleration constraints are introduced into path planning to ensure collision-free and smooth coarse positioning.

[0087] S302, after the robot performs preliminary movement, spatial displacement data of the industrial pipeline at different height points are collected through four point laser displacement sensors.

[0088] S303, determine the angular deviation and position offset between the robot end effector and the axis of the industrial pipeline through spatial displacement data, and determine the target posture control result of the robot end effector based on the angular deviation and position offset between the robot end effector and the axis of the industrial pipeline to control the robot end effector to adjust its posture.

[0089] In one embodiment, the spatial displacement data includes the position coordinates of four spatial points; the angular deviation and position offset between the robot end effector and the axis of the industrial pipeline are determined through the spatial displacement data, including: using the least squares method to fit the position coordinates of the four spatial points to obtain the normal vector of the industrial pipeline axis; based on the direction vector of the robot end effector and the normal vector of the industrial pipeline axis, the angular deviation between the robot end effector and the pipeline axis is determined; based on the coordinates of the center point of the pipeline and the coordinate point of the robot end effector, the position offset between the robot end effector and the industrial pipeline axis is determined.

[0090] Optionally, the position coordinates of the four spatial points are fitted using the least squares method to obtain the normal vector of the industrial pipeline axis, including: constructing an objective function based on the three-dimensional space plane equation; solving the objective function based on the position coordinates of the four spatial points using the least squares method to determine the values ​​of the plane parameters in the objective function; and determining the values ​​of the plane parameters in the objective function as the normal vector of the pipeline axis.

[0091] The equation of a plane in three-dimensional space is: , A 、 B 、 C and D is the plane parameter, ( x , y , z ) are three-dimensional plane coordinates.

[0092] The objective function is: ; in,( , , ) indicates the i The position coordinates of a spatial point, 、 、 and All are variables.

[0093] The output of the point laser displacement sensor is the position coordinates of four spatial points: P1( x 1, y 1, z 1), P2( x 2, y 2, z 2) P3( x 3, y 3, z 3) and P4( x 4, y 4, z 4).

[0094] Based on the position coordinates of the four spatial points, the least squares method is used to fit the objective function to obtain the plane parameters. 、 、 and The value of is the normal vector of the industrial pipeline axis. = ( A , B , C ).

[0095] The direction vector of the robot's end effector is , the angle between the industrial pipeline axis and the robot end is calculated by the following formula: .

[0096] Therefore, the angular deviation between the robot end effector and the axis of the industrial pipeline The calculation formula is: ; in, represents the normal vector of the industrial pipeline axis, Represents the orientation vector of the robot's end effector.

[0097] Calculate the position offset between the robot end effector and the industrial pipeline axis: P_center is the center point coordinate of the pipeline ( , , ), P_effector is the coordinate of the robot end effector ( , , ), the position offset ΔT is: .

[0098] The pipeline center coordinates (P_center) are obtained by using the YOLOv8-Seg model to obtain the semantic mask M(x, y), extracting pixels from the pipeline area, and averaging or weighting the depth values ​​of these pixels (from the depth map Z(x, y)) to obtain the three-dimensional center point, i.e., the pipeline center coordinates.

[0099] Apply the camera intrinsic parameter matrix and coordinate transformation to transform the image coordinate system into the robot base coordinate system. The P_effector (robot end effector coordinate) is obtained by the end effector pose information fed back in real time by the robot controller.

[0100] In one embodiment, the target pose control result of the robot end-effector is determined based on the angular deviation and position offset between the robot end-effector and the axis of the industrial pipeline. Based on the angular deviation and position offset between the robot end-effector and the axis of the industrial pipeline, an inverse kinematics algorithm is used to determine the required adjustment of the robot end-effector joint angles. By adjusting the joint angles of the robot end-effector, the end-effector and the axis of the pipeline are aligned.

[0101] In one embodiment, a point laser displacement sensor array and an angle calculation 1). Horizontal sensor calculation Known: Fixed distance between level sensors: Distance values ​​measured in the horizontal direction: and The spatial coordinates of sensor 1: x 1, y 1, z 1) The spatial coordinates of sensor 2: x 2, y 2, z 2) Calculation steps: Calculate the angle with respect to the horizontal: For each horizontal sensor, calculate the angle between it and the pipeline axis. The direction of the pipeline axis is n pipe = [ a , b , c ], calculate the angle θ h :

[0102] Coordinate transformation: Coordinate transformation is performed using the spatial coordinates of the sensor and the target pipe axis. P is the coordinate of a point on the pipeline axis, T h is the horizontal transformation matrix.

[0103] 2). Vertical sensor calculation Known: Fixed distance between vertical sensors: Distance values ​​measured in the vertical direction: d 3 and d 4 The spatial coordinates of sensor 3: x 3, y 3, z 3) The spatial coordinates of sensor 4: ( x 4, y 4, z 4) Calculation steps: Calculate the vertical angle: For each vertical sensor, calculate the angle between it and the pipe axis θ v : It should be noted that the angle calculation between the horizontal and vertical directions ( θ h 、 θ v ) is a refined decomposition and dynamic compensation of global angle deviation, with the following specific functions: 1. Split-axis control Horizontal angle θ h : Guides the robot to rotate around the Z axis and corrects the deflection of the end effector in the horizontal plane.

[0104] vertical angle θ v : Guides the robot to rotate around the X-axis and corrects the tilt of the end effector in the vertical plane.

[0105] Advantages: Decomposes the comprehensive angle deviation into independent components to avoid adjustment oscillations caused by multi-joint coupled motion.

[0106] 2. High frequency fine-tuning The laser sensor updates at 1kHz θ h and θ v , respond to pipeline vibration or robot motion error in real time to achieve dynamic closed-loop control.

[0107] 3. Related processes Coarse positioning: Visually calculate the global deviation and drive the robot to quickly approach the pipeline.

[0108] Precise positioning: laser direction calculation θ h / θ v , achieving sub-millimeter posture locking.

[0109] In one embodiment, see Figure 2 The system flow follows a closed-loop control of the five logical paths of "perception-transmission-computation-control-execution". The specific process is as follows: Step 1: Initialize communication between the host computer and PLC The host computer sends task parameters, establishes a connection with the PLC via Ethernet, configures the robot's operation process, and initializes the perception module and stands by to respond.

[0110] Step 2: Binocular vision system starts coarse positioning The robot moves to its initial observation position, where a binocular camera captures stereoscopic images of the pipeline area. An AI recognition algorithm extracts the pipeline's contours and axial features, completing initial spatial positioning. A vision-robot calibration matrix is ​​used to transform the coordinate system and initially guide the robot toward the target.

[0111] Step 3: The robot performs coarse pose adjustment Based on the visual perception results, the robot performs rough alignment, moves the end effector (flange) to the near end of the pipe, and completes the near-field docking preparation.

[0112] Step 4: Laser sensor precise positioning start When the robot approaches the target, the four-point laser displacement sensors on the flange are turned on to obtain the precise displacement difference between the upper and lower points at both ends of the pipeline, thereby achieving detailed modeling of the spatial posture of the pipeline.

[0113] Step 5: Axis positioning algorithm fits pipeline posture in real time The system performs spatial calculations on the four-point laser data, fits the pipeline axis, and calculates the relative deviation (angle, distance, direction) from the robot end in real time to form precise positioning feedback.

[0114] Step 6: Fine-tune the robot's end-drive by reverse solution Based on the deviation results, the robot runs the inverse kinematics algorithm to adjust the multi-axis posture and correct the end effector so that it remains parallel to the pipeline axis with high precision and enters the precise insertion area.

[0115] Step 7: Closed-loop control and dynamic feedback correction A high-speed Profinet data channel is maintained between the PLC and the robot, laser sensing data is continuously acquired, high-frequency closed-loop adjustment is achieved, and dynamic micro-deviation and vibration of the pipeline are adapted to achieve truly multi-modal, continuous, and dynamic posture adaptive control.

[0116] This invention proposes a robot control method for industrial pipelines. Through a "phased, two-level" perception strategy, it forms a closed-loop collaborative mechanism of "coarse positioning + fine positioning": (1) Binocular camera and robot collaborate to achieve coarse spatial positioning This invention integrates a binocular vision camera at the front end of the robot's flange to determine the approximate spatial position and orientation of the pipeline. Using a stereo vision algorithm to calculate the pipeline's center point and axial orientation, the robot rapidly approaches the target pipeline, establishing an initial positioning framework and significantly improving the robot's response speed and target acquisition efficiency.

[0117] (2) Laser sensor and robot work together to achieve precise posture positioning Once the robot reaches the pipeline, four laser displacement sensors are symmetrically mounted on either side of the end effector to collect distance data from the upper and lower points at each end of the pipeline. A four-point fitting algorithm calculates the precise axis of the pipeline in real time. Combined with the robot's end-of-pipe position and orientation calculation system, this allows the end effector to achieve precise parallel alignment with the pipeline axis.

[0118] (3) Sensor fusion and closed-loop control mechanism A high-speed Profinet communication channel is established between the PLC controller and the robot's control system, enabling real-time fusion of binocular vision and laser ranging information. This data is processed and converted into position adjustment commands for the robot's end-point. This inverse kinematics-driven joint linkage achieves closed-loop response and dynamic posture compensation.

[0119] The technical advantages and breakthroughs of the present invention include: (1) The multimodal perception system has a clear structure and clear functional division of labor: the binocular camera focuses on global positioning, and the laser sensor focuses on local fine-tuning, which synergistically enhances the system stability and adaptability.

[0120] (2) Adaptability to irregular targets: The system does not rely on regular planes or calibration devices and is suitable for industrial pipelines of different sizes, postures, and materials.

[0121] (3) Easy deployment and rapid reuse: The sensor module is integrated into the end of the robot, ready for installation and flexible deployment, and suitable for a variety of industrial sites.

[0122] (4) High robustness and real-time control: Eliminate response delays through a closed-loop control system, and achieve real-time tracking and adjustment of slight posture differences in dynamic scenarios.

[0123] The key points of the present invention are: (1) Multimodal collaborative perception system architecture for positioning tasks The present invention designs a robot control method for industrial pipelines, focusing on the task of high-precision positioning of axisymmetric targets such as pipelines. This method introduces a heterogeneous fusion mechanism of visual perception and laser displacement perception. Vision is responsible for constructing macroscopic target cognition, and laser is responsible for microscopic posture analysis and local error modeling, realizing a full-chain closed loop of "task orientation-multimodal perception-posture reasoning-control decision-making". By simultaneously introducing visual sensing and laser displacement perception, the target coarse recognition and microscopic posture modeling are completed respectively. Through data fusion, coordinate alignment and error collaborative compensation mechanism, the dynamic posture adjustment of the robot end effector is realized, which is suitable for the docking scenario of cylindrical pipeline targets. Specifically, the binocular vision camera is first used to locate the pipeline entrance area and identify the initial value of the pipeline posture, guiding the laser mode to arrange the docking position; the laser sensor performs precise positioning with high-frequency sampling, and completes the robot posture solution and adjustment based on the multimodal collaborative strategy.

[0124] (2) Coarse-fine collaborative positioning strategy under heterogeneous sensor fusion In a multimodal sensing architecture, the vision system first performs coarse identification and spatial estimation of the pipeline area, providing a preliminary search area for the laser subsystem. Four point laser displacement sensor arrays perform microscopic high-frequency sampling based on coarse positioning, and a model of the pipeline's true axis is established through spatial geometry calculations. This strategy not only improves positioning accuracy but also significantly enhances the robot's adaptability in unstructured environments. Based on the fixed arrangement relationship between sensors and the differences in measurement data, this method directly constructs a model of the target's relative spatial posture, avoiding traditional coordinate conversion and calibration processes, improving deployment flexibility and model generalization.

[0125] (3) Structured spatial reasoning mechanism constructed by four-point laser array with different heights By symmetrically placing four point laser displacement sensors along the flange of the robot's end fixture, a spatial "non-coplanar measurement array at different heights" is constructed. The geometric differences between the measurement points enable structural reconstruction of the three-dimensional spatial posture, supporting active real-time adjustment of axial parallelism. The four point laser displacement sensors are arranged at different heights along the edge of the pipe flange, forming a structurally stable non-coplanar measurement array. A spatial model is constructed by measuring the distance difference between the laser points on each end surface and the pipe surface, enabling fitting of the target axis and real-time posture error feedback.

[0126] (4) “Perception internal reference autonomous mechanism” without external calibration benchmark This invention breaks the traditional robotic system's reliance on external calibration plates and global coordinate systems, proposing a "perception-based autonomous" deployment approach. This approach, based entirely on the relative structural relationships between sensors, estimates the spatial posture of the pipeline through internal geometric modeling and data fitting, enabling rapid deployment and adaptability to diverse specifications. The attitude error parameters calculated through real-time perception serve as the robot's inverse kinematic control input. This, combined with a high-speed communication protocol between the PLC and the robot, enables high-frequency data updates, thereby implementing dynamic attitude compensation control for the robot's end effector.

[0127] The present invention provides a robot control method for industrial pipelines, which has the following effects: 1. Dual-modal collaborative perception improves positioning accuracy and spatial understanding By combining a binocular vision system with a four-point laser displacement sensor, the present invention achieves high-precision three-dimensional positioning of pipeline targets, with a positioning accuracy of ±0.01mm. Compared with traditional single-sensor systems (where the positioning accuracy is typically ±0.5mm), this significantly improved accuracy ensures the robot's reliability and accuracy in scenarios such as high-precision assembly, pipeline docking, and welding, effectively avoiding operation failures and quality issues caused by positioning errors.

[0128] 2. Real-time response and efficient dynamic control This invention uses the Profinet protocol to achieve high-speed, low-latency data transmission between the PLC and the robot control system, reducing communication latency to less than 5ms. This low-latency response significantly improves the system's dynamic response capabilities, enabling the robot to quickly respond to changes in pipeline posture and accurately adjust the end effector's posture on high-speed production lines, ensuring efficient and stable production operations.

[0129] 3. Enhance system stability and robustness Through a closed-loop control system, the present invention monitors laser sensor data in real time and automatically adjusts the robot's end-effector's posture to ensure the robot maintains high-precision parallelism with the pipeline axis. This control system effectively suppresses errors caused by external interference, pipeline deformation, equipment deviation, and other factors, improving system stability and robustness. Even under long-term, high-load operations, positioning accuracy remains at or above 99.5%.

[0130] 4. Significantly improve production efficiency and reduce manual intervention Compared with traditional manual calibration or low-precision automated systems, the present invention uses real-time data feedback and closed-loop control to enable the robot to complete pipeline positioning and position calibration in seconds, reducing the time consumption by >80% compared to traditional systems (which usually take several minutes).

[0131] 5. Improve system adaptability and compatibility The system design is highly adaptable, supporting pipes of varying diameters, materials (e.g., steel, plastic, and composite), and shapes (e.g., straight and curved), ensuring versatility across diverse industrial scenarios. Its modular design allows for rapid adaptation to pipe targets of varying sizes and supports a wider range of workpiece sizes, improving adaptability by over 50%.

[0132] 6. Reduce system failure rate and maintenance costs By utilizing high-precision laser sensors and a stable communication protocol, the system's failure rate is reduced by approximately 70% compared to traditional mechanical touch or optical sensors. The system's real-time error compensation mechanism reduces robot positioning errors and deviations, reducing frequent adjustments and maintenance, lowering long-term maintenance costs and significantly reducing equipment downtime.

[0133] 7. Enhance the system’s intelligence and adaptability This invention integrates an intelligent error compensation algorithm and a dynamic feedback control mechanism, enabling the robot to automatically adjust its operating path and correct target posture deviations in real time, reducing reliance on manual operation. The robot can adaptively optimize its control strategy based on the operating environment and continuously self-optimize based on feedback data, enabling it to autonomously handle complex tasks, further enhancing the system's intelligence and production automation levels.

[0134] The specific definition of the robot control system for industrial pipelines provided by the present invention can be found in the above definition of the robot control method for industrial pipelines, which will not be repeated here.

[0135] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of the present invention.

Claims

1. A robot control method for industrial pipelines, characterized in that: A flange is installed on the end effector of the robot, a binocular vision camera is installed at the front center of the flange, and four point laser displacement sensors are arranged symmetrically on the flange with different heights; the method includes: Collect pipeline images of industrial pipelines through binocular vision cameras; Determine the pipeline position and the angle between the robot and the industrial pipeline based on the pipeline image, and determine the initial position control result of the robot based on the pipeline position and the angle between the robot and the industrial pipeline to control the robot to perform preliminary movement; After the robot performs preliminary movements, the spatial displacement data of the industrial pipeline at different heights are collected by four point laser displacement sensors; The angular deviation and position offset between the robot end effector and the axis of the industrial pipeline are determined through spatial displacement data. Based on the angular deviation and position offset between the robot end effector and the axis of the industrial pipeline, the target posture control result of the robot end effector is determined to control the robot end effector to adjust its posture.

2. The method according to claim 1, characterized in that The pipeline image includes a left image frame and a right image frame. The pipeline position and the angle between the robot and the industrial pipeline are determined based on the pipeline image, including: The left and right image frames are input into the end-to-end deep network PSMNet to obtain a pixel-level disparity map; Obtain a depth map based on the pixel-level disparity map, the focal length of the binocular vision camera, and the baseline length; The depth map is fed into the image semantic segmentation model to obtain the pipeline target bounding box; the image semantic segmentation model is built based on the lightweight network YOLOv8-Seg; Determine the pipeline location based on the pipeline target bounding box; Calculate the three-dimensional coordinates of the center point of the pipeline based on the intrinsic parameter matrix of the binocular vision camera, the pipeline position and its corresponding parallax value; The angle between the robot and the industrial pipeline is determined based on the three-dimensional coordinates of the pipeline center point.

3. The method according to claim 2, characterized in that The pipeline position includes the three-dimensional coordinates of the pipeline center point. Based on the pipeline position and the angle between the robot and the industrial pipeline, the robot's initial posture control results are determined, including: Based on the 3D coordinates of the pipeline center point and the angle between the robot and the industrial pipeline, the robot base coordinate system is aligned with the pipeline coordinate system to generate the linear motion path parameters from the robot's current position to the industrial pipeline. The path parameters include the translation and rotation of the end effector's target position. The translation and rotation of the target posture are converted into angle instructions of each joint of the end effector to form the initial posture control result.

4. The method according to claim 1, wherein The spatial displacement data includes the position coordinates of four spatial points. Through the spatial displacement data, the angular deviation and position offset between the robot end effector and the axis of the industrial pipeline are determined, including: The position coordinates of four spatial points are fitted using the least squares method to obtain the normal vector of the industrial pipeline axis; Determine the angular deviation between the robot end effector and the industrial pipeline axis according to the direction vector of the robot end effector and the normal vector of the pipeline axis; According to the center point coordinates of the pipeline and the coordinate points of the robot's end effector, the position offset between the robot's end effector and the axis of the industrial pipeline is determined.

5. The method according to claim 4, characterized in that The position coordinates of the four spatial points are fitted using the least squares method to obtain the normal vector of the industrial pipeline axis, including: Construct the objective function based on the three-dimensional space plane equation; According to the position coordinates of the four spatial points, the least square method is used to solve the objective function and determine the values ​​of the plane parameters in the objective function; The value of the plane parameter in the objective function is determined as the normal vector of the industrial pipeline axis.

6. The method according to claim 5, characterized in that The objective function is: ； in,( , , ) indicates the i The position coordinates of a spatial point, 、 、 and All are variables.

7. The method according to claim 4, characterized in that Angular deviation between the robot end effector and the axis of the industrial pipeline The calculation formula is: ； in, represents the normal vector of the industrial pipeline axis, Represents the orientation vector of the robot's end effector.

8. The method according to claim 1, characterized in that Based on the angular deviation and position offset between the robot end effector and the industrial pipeline axis, the target posture control result of the robot end effector is determined, including: According to the angular deviation and position offset between the robot's end effector and the axis of the industrial pipeline, the joint angle that needs to be adjusted of the robot's end effector is determined through the inverse kinematics algorithm.

9. A robot control system for industrial pipelines, characterized in that: The system includes: a robot and a host computer; a flange is installed on the end effector of the robot, a binocular vision camera is installed at the front center of the flange, and four point laser displacement sensors are symmetrically arranged on the flange with different heights; Binocular vision camera, used to collect pipeline images of industrial pipelines and send the pipeline images to the host computer; The host computer is used to determine the pipeline position and the angle between the robot and the industrial pipeline based on the pipeline image, and determine the initial posture control result of the robot based on the pipeline position and the angle between the robot and the industrial pipeline to control the robot to perform preliminary movement; Four point laser displacement sensors are used to collect spatial displacement data of industrial pipelines at different heights after the robot performs initial movement, and send the spatial displacement data to the host computer; The host computer is used to determine the angular deviation and position offset between the robot end effector and the axis of the industrial pipeline through spatial displacement data, and determine the target posture control result of the robot end effector based on the angular deviation and position offset between the robot end effector and the axis of the industrial pipeline, so as to control the robot end effector to adjust its posture.

10. The system according to claim 9, characterized in that The system also includes a programmable logic controller, the host computer communicates with the programmable logic controller, and the robot's actuator communicates with the programmable logic controller at high speed and in two directions; The programmable logic controller is used to receive the pipeline image sent by the binocular vision camera and upload it to the host computer; receive the initial posture control result sent by the host computer and send it to the robot's actuator; receive the spatial displacement data sent by the four-point laser displacement sensors and upload it to the host computer; receive the target posture control result sent by the host computer and send it to the robot's actuator.

Images