Stacking robot high-precision positioning method based on multi-perception information

Abstract

The invention discloses a stacking robot high-precision positioning method based on multi-perception information, which comprises the following steps: driving a mechanical arm combination through a servo closed-loop control system, and driving a grabbing disc to pick up a target component with a length extension direction from an initial position. And according to the initial position and the set target position, initial pose control information of the mechanical arm is generated. Distance information of the target component is collected through the laser distance measuring sensor. A target component image is collected through a visual sensor. And generating pose positioning information of the mechanical arm of the servo closed-loop control system at the extending position according to the target component image and the distance information. The problems that an existing stacking robot is poor in control precision and cannot adapt to the complex industrial environment are solved. Cooperative processing is carried out on vision and laser ranging information through a multi-sensor data fusion algorithm, and space coordinates of an end effector of the mechanical arm are corrected in real time. And in combination with a dynamic feedback compensation mechanism, positioning deviation caused by mechanical vibration or environment illumination change is eliminated.

Description

Technical Field

[0001] This invention relates to the field of robotics. Specifically, it relates to a high-precision positioning method and system for a stacking robot based on multi-sensor information. Background Technology

[0002] Stacking robots, as key equipment, are widely used in material handling and palletizing operations in logistics, warehousing, and manufacturing. Traditional stacking robot positioning technology mainly relies on pre-set location teaching or simple visual recognition methods, but these methods have significant limitations. For example, by using location teaching, the robot can only perform grasping operations along fixed paths and in fixed postures, making it unable to adapt to scenarios with irregular object shapes, chaotic arrangements, or dynamically changing environments. This easily leads to grasping failures, low efficiency, and uneven palletizing, increasing operation time and costs.

[0003] Therefore, it can be seen that existing stacking robots have poor accuracy in recognizing multi-shaped (shaft) products and cannot meet the handling needs in complex industrial environments. Summary of the Invention

[0004] The purpose of this invention is to provide a high-precision positioning method and system for stacking robots based on multi-sensor information, which solves the problems of poor control accuracy and inability to adapt to complex industrial environments in existing stacking robots.

[0005] A first aspect of the invention provides a high-precision positioning method for a stacking robot based on multi-sensor information. The stacking robot includes a robotic arm assembly capable of moving in three dimensions and a gripping disk. The robotic arm has a gripping end. The gripping disk is disposed at the gripping end. A vision sensor is disposed on the gripping disk. The robotic arm is controlled by a servo closed-loop control system. The stacking robot is used in conjunction with a conveyor belt. The conveyor belt has a conveying surface. A laser rangefinder sensor is disposed on the conveying surface.

[0006] High-precision positioning methods for stacking robots based on multi-sensor information include: The robotic arm assembly is driven by a servo closed-loop control system, which moves the gripping disk to pick up a target component with a length extension direction from the starting position.

[0007] Based on the starting position and the set target position, initial pose control information for the robotic arm is generated. This initial pose control information includes rotation and extension position information. The robotic arm assembly is then driven by this initial pose information to move the gripper disk above the set target position.

[0008] Distance information of the target component is acquired using a laser rangefinder. Images of the target component are acquired using a vision sensor.

[0009] Using the target component image and distance information, the pose and positioning information of the robotic arm in the extended position of the servo closed-loop control system is generated. This pose and positioning information includes the gripping angle information, which enables the gripping disk to lower the target component based on the gripping angle.

[0010] In one embodiment of the high-precision positioning method for a stacking robot based on multi-sensor information of the present invention, a vision sensor is set at the end of the gripping disk that faces the conveying surface. The vision sensors are respectively set at the edge and the center of the gripping disk. The vision sensor at the center position acquires an image of the target component. The vision sensor at the edge position acquires an auxiliary positioning image.

[0011] In one embodiment of the high-precision positioning method for a stacking robot based on multi-sensor information of the present invention, the step of generating the pose positioning information of the robotic arm of the servo closed-loop control system at the extended position by using the target component image and distance information includes the following steps.

[0012] The target component image is segmented to obtain the segmented regions of the target component. The center position of the target component is obtained based on the segmented regions.

[0013] The position information of the laser sensor is identified and acquired using assisted positioning images and a pre-trained assisted positioning model. Based on the position information of the laser sensor relative to the center position, the offset adjustment information of the target component is determined. The orientation information of the target component is determined based on the center position and the segmented region of the target component. Based on the offset adjustment information and the orientation information, the pose positioning information of the robotic arm in the extended position of the servo closed-loop control system is generated.

[0014] In one embodiment of the high-precision positioning method for a stacking robot based on multi-sensor information of the present invention, the target component image is segmented to obtain the segmented region of the target component, which includes: The target object image is segmented by a pre-trained target object segmentation model, which is trained by collecting target object sample images arranged in different ways under different environments.

[0015] Edge image processing is performed on the segmented target object region to obtain the segmented region of the target component with edge information.

[0016] In one embodiment of the high-precision positioning method for a stacking robot based on multi-sensor information of the present invention, the input to the training of the auxiliary positioning model is an image of the transport surface with markers. The markers and the laser rangefinder have a specific distance correspondence. The output of the auxiliary positioning model is the position information from the laser rangefinder.

[0017] In one embodiment of the high-precision positioning method for a stacking robot based on multi-sensor information of the present invention, the robotic arm includes a base, a bottom rotating platform, a lower moving arm, an upper moving arm, and an end moving arm. The base has a support surface. The upper part of the lower moving arm and the rotating platform are rotatably connected to the lower moving arm about a lower axis. The lower axis is parallel to the support surface. The lower moving axis is driven by a lower motor based on lower drive information. The lower drive information includes the rotation angle information of the lower moving arm, the starting position information, and the target position information.

[0018] In one embodiment of the high-precision positioning method for a stacking robot based on multi-sensor information of the present invention, the bottom rotating platform and the upper part of the base are rotatably connected about a rotation axis. The rotation axis is perpendicular to the support surface. The bottom rotating platform is driven by a bottom motor based on bottom drive information. The bottom drive information includes base rotation angle information, starting angle information, and target angle information.

[0019] In one embodiment of the high-precision positioning method for a stacking robot based on multi-sensor information of the present invention, the upper part of the upper motion arm and the lower motion arm are rotatably connected to the upper motion arm about an upper axis. The upper axis is parallel to the support surface. The upper motion axis is driven by an upper motor based on upper drive information. The upper drive information includes the rotation angle information of the upper motion arm, the starting position information, and the target position information.

[0020] The ends of the end arms and the upper arm are rotatably connected to the end arms about an end axis. The end axis is parallel to the support surface. The end motion axis is driven by an end motor based on end drive information. The end drive information includes the end arm rotation angle, starting position, and target position.

[0021] In one embodiment of the high-precision positioning method for a stacking robot based on multi-sensor information of the present invention, the servo closed-loop control system includes: The main controller's outputs are connected to the lower motor, upper motor, end motor, and bottom motor, respectively. It generates pose and positioning information by combining the lower drive information, upper drive information, end drive information, and bottom drive information.

[0022] In a second aspect, the present invention provides a stacking robot based on multi-sensor information, which includes a stacking robot controlled by the high-precision positioning method for stacking robots based on multi-sensor information of the present invention.

[0023] The following section will further explain the characteristics, technical features, advantages, and implementation methods of the high-precision positioning method and system for stacking robots based on multi-sensor information in a clear and easy-to-understand manner, with the aid of accompanying figures. Attached Figure Description

[0024] Figure 1 This is a schematic diagram illustrating the structure of a stacking robot based on multi-sensor information in one embodiment of the present invention. Detailed Implementation

[0025] To provide a clearer understanding of the technical features, objectives, and effects of the invention, specific embodiments of the invention are now described with reference to the accompanying drawings. In the drawings, the same reference numerals indicate components with the same or similar structures but the same function.

[0026] In this document, "illustrative" means "serving as an example, illustration, or description," and any illustrations or embodiments described herein as "illustrative" should not be construed as a more preferred or advantageous technical solution. For the sake of brevity, each figure only schematically shows the parts relevant to this exemplary embodiment, and they do not represent the actual structure or true proportions of the product.

[0027] One aspect of the present invention provides a high-precision positioning method for a stacking robot based on multi-sensor information, referring to... Figure 1 The stacking robot 100 includes a robotic arm assembly capable of movement in three dimensions and a gripping disk 10. The robotic arm has a gripping end. The gripping disk 10 is disposed at the gripping end. A vision sensor is disposed on the gripping disk 10. The robotic arm is controlled by a servo closed-loop control system. The stacking robot 100 is used in conjunction with a conveyor belt. The conveyor belt has a conveying surface 201. A laser rangefinder sensor is disposed on the conveying surface.

[0028] High-precision positioning methods for stacking robots based on multi-sensor information include: Reference Figure 1 The robotic arm assembly is driven by a servo closed-loop control system, which drives the gripping disk 10 to pick up a target component 200 with a length extension direction from the starting position.

[0029] Based on the starting position and the set target position, initial pose control information for the robotic arm is generated. This initial pose control information includes rotation and extension position information. The robotic arm assembly, driven by this initial pose information, moves the gripper 10 above the set target position. At this time, a vision sensor acquires edge image information of the target component 200 in real time, and a laser rangefinder dynamically compensates for changes in the height of the conveyor surface. A multi-source data fusion algorithm corrects rotational deviations and height errors in the initial pose control information. When the gripper 10 descends to the set target position, its posture is precisely adjusted according to the updated pose control commands to ensure that the length extension direction of the target component 200 is completely consistent with the stacking requirements, achieving high-precision positioning and placement.

[0030] Distance information of the target component is acquired using a laser rangefinder. Images of the target component are acquired using a vision sensor. The image information and distance information are spatiotemporally registered to construct a 3D point cloud model of the target component, and its attitude feature parameters are extracted.

[0031] Using the target component image and distance information, the pose and positioning information of the robotic arm in the extended position is generated by the servo closed-loop control system. This pose and positioning information includes the gripping angle information, which enables the gripping disk 10 to lower the target component 200 according to the gripping angle. By combining real-time feedback from gyroscopes and accelerometers on the attitude angle changes of the gripping disk, minor offsets during the dynamic process are further corrected to ensure the continuity and stability of the pose and positioning information. Based on the constructed 3D point cloud model, the central axis of the target component is fitted using the least squares method, and its spatial deviation from the ideal stacking pose is calculated. A compensation vector is generated and input into the servo control system. Finally, multi-axis linkage fine-tuning is completed under PLC coordination, achieving sub-millimeter-level positioning accuracy to meet the requirements of high-density stacking processes.

[0032] As can be seen from the above technical solution, the implementation process of the high-precision positioning method for stacking robots based on multi-sensor information of the present invention is as follows: First, the robotic arm is driven by a servo closed-loop control system to move the gripper disk above the target component. Image and distance information are simultaneously acquired using a visual sensor and a laser rangefinder to construct a three-dimensional point cloud model and extract posture features. Combined with the initial pose control information, rotational deviation and height error are corrected in real time, and the posture of the gripper disk is precisely adjusted under dynamic compensation.

[0033] By fitting the central axis using the least squares method, calculating the spatial deviation, and generating a compensation vector, this vector is input into the servo system for multi-axis linkage fine-tuning, ultimately achieving high-precision stacking positioning of the target component along its length. This method, through multi-source sensor fusion and real-time feedback control, significantly improves the accuracy and stability of stacking operations, effectively reduces the risk of stacking collapse due to positioning errors, and enhances warehouse space utilization and operational efficiency.

[0034] Therefore, the advantages of the high-precision positioning method for stacking robots based on multi-sensor information in this invention are that it achieves accurate perception and dynamic compensation of the spatial pose of the target component through multi-sensor fusion technology, effectively overcomes the problem of limited accuracy of a single sensor, significantly improves the positioning accuracy and operational stability of the stacking robot under complex working conditions, and has good real-time performance and scalability, making it suitable for automated high-density stacking scenarios of components of various specifications.

[0035] In one embodiment of the high-precision positioning method for a stacking robot based on multi-sensor information of the present invention, a vision sensor is provided at the end of the gripping disk 10 that faces the conveying surface. The vision sensors are respectively located at the edge and center of the gripping disk 10. The vision sensor at the center position acquires images of the target component. The vision sensor at the edge position acquires auxiliary positioning images. Therefore, simultaneous capture of the central region and boundary features of the target component can be achieved, improving the integrity and reliability of image acquisition. The vision sensor at the center position is used to accurately identify the surface texture and contour information of the component, while the sensor at the edge position assists in monitoring the relative posture of the gripping disk and the conveying surface, preventing positioning inaccuracies due to tilting or offset.

[0036] In one embodiment of the high-precision positioning method for a stacking robot based on multi-sensor information of the present invention, the step of generating the pose positioning information of the robotic arm of the servo closed-loop control system at the extended position through the target component image and distance information includes: The target component image is segmented to obtain segmented regions. The center position of the target component is determined based on these segments. The lateral and longitudinal deviations of the target component relative to the gripping disk are calculated based on the offset of the center position from the origin of the image coordinate system. Combined with depth information obtained from a distance sensor, initial pose values ​​of the target component in 3D space are constructed. Using a calibrated vision-manipulator hand-eye relationship matrix, pixel coordinates are converted to spatial coordinates in the manipulator's base coordinate system. This allows for the calculation of the required attitude angles and translational amounts for the manipulator's end effector, which are then output to the servo closed-loop control system for real-time correction.

[0037] The position information of the laser sensor is identified and acquired using assisted positioning images and a pre-trained assisted positioning model. Based on the position information of the laser sensor relative to the center position, the offset adjustment information of the target component is determined. The orientation information of the target component is determined based on the center position and the segmented region of the target component. Based on the offset adjustment information and the orientation information, the pose positioning information of the robotic arm in the extended position of the servo closed-loop control system is generated.

[0038] Therefore, this method achieves high-precision joint calculation of the target component's pose by fusing visual segmentation results with laser ranging data, further suppressing recognition interference caused by changes in ambient lighting and surface reflection, and can still stably output sub-millimeter-level positioning parameters on a dynamic conveyor line.

[0039] The generated pose and positioning information is fed back to the servo closed-loop control system in real time, driving the robotic arm to continuously correct its trajectory during extension, effectively avoiding the risk of grasping failure or collision due to initial positioning errors. Simultaneously, the auxiliary positioning model employs a lightweight neural network structure, enabling efficient operation on edge computing units and meeting the stringent response latency requirements of industrial scenarios. The entire process requires no manual calibration, possesses self-learning and adaptive capabilities, and can quickly adapt to component types of different sizes, shapes, and stacking patterns, significantly improving the system's intelligence and engineering practicality.

[0040] In one embodiment of the high-precision positioning method for a stacking robot based on multi-sensor information of the present invention, the target component image is segmented to obtain the segmented region of the target component, which includes: The target object image is segmented by a pre-trained target object segmentation model, which is trained by collecting target object sample images arranged in different ways under different environments.

[0041] Edge image processing is performed on the segmented target object regions to obtain segmented regions of target components with edge information. Contour features of the target components are extracted using the edge information, and morphological operations are combined to eliminate noise interference and accurately locate their geometric center. The contour features are mapped to the grasping coordinate system through coordinate transformation, and the rotational deviation angle between the target component and the standard pose is calculated. Based on the rotational deviation angle and the geometric center offset, an initial adjustment vector is generated, and depth data is fused to correct for height dimension errors.

[0042] In one embodiment of the high-precision positioning method for a stacking robot based on multi-sensor information of the present invention, the input to the training of the auxiliary positioning model is an image of the transport surface with markers. The markers and the laser rangefinder have a specific distance correspondence. The output of the auxiliary positioning model is the position information from the laser rangefinder.

[0043] In one embodiment of the high-precision positioning method for a stacking robot based on multi-sensor information of the present invention, the robotic arm includes a base 21, a bottom rotating platform 22, a lower moving arm 23, an upper moving arm 24, and an end moving arm 25. The base 21 has a support surface. The lower moving arm 23 and the upper part of the rotating platform are rotatably connected to the lower moving arm 23 about a lower axis. The lower axis is parallel to the support surface. The lower moving axis is driven by a lower motor based on lower drive information. The lower drive information includes the rotation angle information, starting position information, and target position information of the lower moving arm 23.

[0044] In one embodiment of the high-precision positioning method for a stacking robot based on multi-sensor information of the present invention, the upper parts of the bottom rotating platform 22 and the base 21 are rotatably connected about a rotation axis. The rotation axis is perpendicular to the support surface. The bottom rotating platform 22 is driven by a bottom motor based on bottom drive information. The bottom drive information includes the rotation angle information of the base 21, the starting angle information, and the target angle information.

[0045] In one embodiment of the high-precision positioning method for a stacking robot based on multi-sensor information of the present invention, the upper part of the upper moving arm 24 and the lower moving arm 23 are rotatably connected to the upper moving arm 24 about an upper axis. The upper axis is parallel to the support surface. The upper moving axis is driven by an upper motor based on upper driving information. The upper driving information includes the rotation angle information of the upper moving arm 24, the starting position information, and the target position information.

[0046] The ends of the end motion arm 25 and the upper motion arm 24 are rotatably connected to the end motion arm 25 about an end axis. The end axis is parallel to the support surface. The end motion axis is driven by an end motor based on end drive information. The end drive information includes the rotation angle information, starting position information, and target position information of the end motion arm 25.

[0047] In one embodiment of the high-precision positioning method for a stacking robot based on multi-sensor information of the present invention, the servo closed-loop control system includes: The main controller's outputs are connected to the lower motor, upper motor, end motor, and bottom motor, respectively. It generates pose and positioning information by combining the lower, upper, end, and bottom drive information. This servo closed-loop control system can dynamically adjust the combination of lower, upper, end, and bottom drive information by comparing the encoder feedback data of each moving arm with the target pose information in real time, thereby generating pose and positioning information and achieving high-precision positioning of the stacking robot's end effector.

[0048] The encoder feedback data includes the actual rotation angle and position offset of each moving arm. The main controller performs PID adjustment based on the deviation value to ensure dynamic stability and positioning accuracy during movement. Through multi-sensor information fusion technology, the system can adapt to changes in the external environment, effectively eliminate accumulated errors, and further improve repeatability.

[0049] In a second aspect, the present invention provides a stacking robot based on multi-sensor information, which includes a stacking robot 100 controlled by the high-precision positioning method for stacking robots based on multi-sensor information of the present invention.

[0050] It should be understood that although this specification describes various embodiments, not every embodiment contains only one independent technical solution. This way of describing the specification is only for clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

[0051] The detailed descriptions listed above are merely specific descriptions of feasible embodiments of the present invention, and are not intended to limit the scope of protection of the present invention. All equivalent embodiments or modifications made without departing from the spirit of the present invention should be included within the scope of protection of the present invention.

Claims

1. A high-precision positioning method for a stacking robot based on multi-sensor information, characterized in that, The stacking robot includes a robotic arm assembly capable of moving in three dimensions and a gripping disk; the robotic arm has a gripping end; the gripping disk is disposed on the gripping end; a vision sensor is disposed on the gripping disk; the robotic arm is controlled by a servo closed-loop control system; the stacking robot is used in conjunction with a conveyor belt; the conveyor belt has a conveying surface; a laser rangefinder is disposed on the conveying surface; The high-precision positioning method for stacking robots based on multi-sensor information includes: The robotic arm assembly is driven by a servo closed-loop control system, which drives the gripping disk to pick up a target component with a length extension direction from the starting position. Based on the starting position and the set target position, initial pose control information for the robotic arm is generated; the initial pose control information includes rotation information and extension position information; the robotic arm assembly is driven by the initial pose information to move the gripping disk above the set target position; Distance information of the target component is acquired using a laser rangefinder; images of the target component are acquired using a vision sensor. Using the target component image and the distance information, the pose positioning information of the robotic arm in the extended position of the servo closed-loop control system is generated; the pose positioning information includes gripping angle information, so that the gripping disk can put down the target component through the gripping angle information.

2. The high-precision positioning method for a stacking robot based on multi-sensor information according to claim 1, characterized in that, The vision sensor is disposed at one end of the gripping disk that faces the conveying surface; the vision sensor is disposed at the edge and the middle of the gripping disk respectively; the vision sensor at the middle position acquires the image of the target component; the vision sensor at the edge position acquires the auxiliary positioning image.

3. The high-precision positioning method for a stacking robot based on multi-sensor information according to claim 2, characterized in that, The step of generating the pose positioning information of the robotic arm in the extended position of the servo closed-loop control system using the target component image and the distance information includes: The target component image is segmented to obtain the target component segmentation region; The center position of the target component is obtained based on the segmented region of the target component. The location information of the laser sensor is identified and obtained through the assisted positioning image and the pre-trained assisted positioning model. Based on the position information of the laser sensor at the center position, the offset adjustment information of the target component is determined; The orientation information of the target component is determined based on the center position and the segmented region of the target component; Based on the offset adjustment information and the orientation information, the pose positioning information of the robotic arm in the extended position of the servo closed-loop control system is generated.

4. The high-precision positioning method for a stacking robot based on multi-sensor information according to claim 3, characterized in that, The segmentation of the target component image to obtain the target component segmentation region includes: The target object image is segmented by a pre-trained target object segmentation model, wherein the target object segmentation model is trained by collecting target object sample images arranged in different ways under different environments; Edge image processing is performed on the segmented target object region to obtain the segmented region of the target component with edge information.

5. The high-precision positioning method for a stacking robot based on multi-sensor information according to claim 3, characterized in that, During the training of the auxiliary positioning model, the input is a transmission surface image with markers; the markers have a specific distance correspondence with the laser ranging sensor; the output of the auxiliary positioning model is the position information of the laser ranging sensor.

6. The high-precision positioning method for a stacking robot based on multi-sensor information according to claim 1, characterized in that, The robotic arm includes a base, a bottom rotating platform, a lower moving arm, an upper moving arm, and an end moving arm; the base has a support surface; the upper part of the lower moving arm and the rotating platform are rotatably connected to the lower moving arm about a lower axis; the lower axis is parallel to the support surface; the lower moving axis is driven by a lower motor based on lower drive information; the lower drive information includes the rotation angle information of the lower moving arm, the starting position information, and the target position information.

7. The high-precision positioning method for a stacking robot based on multi-sensor information according to claim 6, characterized in that, The bottom rotating platform and the upper part of the base are rotatably connected around a rotating axis; the rotating axis is perpendicular to the support surface; the bottom rotating platform is driven by a bottom motor based on bottom drive information; the bottom drive information includes base rotation angle information, starting angle information and target angle information.

8. The high-precision positioning method for a stacking robot based on multi-sensor information according to claim 7, characterized in that, The upper part of the upper moving arm and the upper part of the lower moving arm are rotatably connected to the upper moving arm about an upper axis; the upper axis is parallel to the support surface; the upper moving axis is driven by an upper motor based on upper driving information; the upper driving information includes the rotation angle information of the upper moving arm, the starting position information, and the target position information; The ends of the end moving arm and the upper moving arm are rotatably connected to the end moving arm about one end axis; the end axis is parallel to the support surface; The end motion axis is driven by an end motor based on end drive information; the end drive information includes the end motion arm rotation angle information, starting position information and target position information.

9. The high-precision positioning method for a stacking robot based on multi-sensor information according to claim 8, characterized in that, The servo closed-loop control system includes: The main controller has its output terminals connected to the lower motor, upper motor, end motor, and bottom motor respectively; it generates the pose positioning information by combining the lower drive information, upper drive information, end drive information, and bottom drive information.

10. A stacking robot based on multi-sensor information, characterized in that, It includes the stacking robot controlled by the high-precision positioning method for stacking robots based on multi-sensor information according to any one of claims 1 to 9.

Images