Part assembling apparatus and part assembling method

Abstract

This invention provides a parts assembly device and a parts assembly method, relating to the fields of intelligent manufacturing, automated assembly, and robotics. The parts assembly device includes: a base; a robotic arm mounted on the base; a connecting mechanism mounted on the end of the robotic arm away from the base; a camera mounted on the connecting mechanism, configured to follow the robotic arm to a work platform and acquire an initial image of a target workpiece on the work platform; an assembly mechanism mounted on the connecting mechanism; and a processor installed inside the base and electrically connected to the robotic arm, camera, and assembly mechanism. The processor is configured to: determine the orientation information of the surface to be assembled in the target workpiece based on the initial image; control the robotic arm to move the camera to the target position based on the orientation information; and control the robotic arm to move the assembly mechanism to assemble the parts to be assembled into the assembly area based on the target image acquired by the camera at the target position.

Description

Technical Field

[0001] This invention relates to the fields of intelligent manufacturing, automated assembly and robotics, and in particular to a parts assembly device and a parts assembly method. Background Technology

[0002] In automated assembly lines, using robotic arms to fasten screws is a common assembly task. For example, in scenarios where screws are fastened at threaded holes on a workpiece surface, the control system needs to accurately obtain information such as the center position, diameter, depth of the threaded hole, orientation of the workpiece surface, and feed direction of the screw. Therefore, the assembly process requires high precision, but high-precision equipment is expensive, and the control logic is also complex. Summary of the Invention

[0003] This invention provides a parts assembly device and a parts assembly method.

[0004] According to a first aspect, the present invention provides a parts assembly apparatus comprising: a base; a robotic arm mounted on the base; a connecting mechanism mounted on the end of the robotic arm away from the base; a camera mounted on the connecting mechanism, configured to follow the robotic arm to a work platform and acquire an initial image of a target workpiece on the work platform; an assembly mechanism mounted on the connecting mechanism; and a processor installed inside the base and electrically connected to the robotic arm, the camera, and the assembly mechanism, the processor being configured to: determine the orientation information of the surface to be assembled in the target workpiece based on the initial image; control the robotic arm to move the camera to a target position based on the orientation information; and control the robotic arm to move the assembly mechanism to assemble the parts to be assembled into the assembly area based on a target image of the assembly area acquired by the camera at the target position.

[0005] According to a second aspect, the present invention provides a part assembly method, applied to the part assembly apparatus provided by the present invention, the method comprising: determining the orientation information of the surface to be assembled in a target workpiece based on an initial image acquired by a camera; controlling a robotic arm to move the camera to a target position based on the orientation information, wherein the orthographic projection of the target position on the surface to be assembled overlaps with the area to be assembled in the target workpiece; and controlling the robotic arm to move the assembly mechanism to assemble the part to be assembled into the area to be assembled based on a target image of the area to be assembled acquired by the camera at the target position.

[0006] In an embodiment of the present invention, the processor first uses a camera to coarsely locate the surface of the target workpiece, then moves the camera to a position where it is directly facing the workpiece surface. From this direct-view perspective, the camera is then used to precisely locate the assembly area on the surface of the target workpiece. This reduces interference caused by the camera's angle of view and the structure of the target workpiece, lowers the difficulty of locating the assembly area, simplifies assembly, and improves assembly accuracy. Attached Figure Description

[0007] Figure 1 A schematic diagram of a parts assembly apparatus according to an embodiment of the present invention is shown.

[0008] Figure 2 A schematic diagram illustrating the working principle of a parts assembly device according to an embodiment of the present invention is shown.

[0009] Figure 3 A partial structural schematic diagram of a parts assembly apparatus according to an embodiment of the present invention is shown.

[0010] Figure 4 A partial structural schematic diagram of a parts assembly apparatus according to another embodiment of the present invention is shown.

[0011] Figure 5 A schematic diagram of the structure of a floating component according to an embodiment of the present invention is shown.

[0012] Figure 6 A schematic diagram of the structure of the air suction component and the quick-change component according to an embodiment of the present invention is shown.

[0013] Figure 7 A schematic diagram of a parts assembly apparatus according to another embodiment of the present invention is shown.

[0014] Figure 8 A partial structural schematic diagram of a parts assembly apparatus according to another embodiment of the present invention is shown.

[0015] Figure 9 A schematic diagram of the mounting base according to an embodiment of the present invention is shown.

[0016] Figure 10 A schematic diagram of the structure of the tray holder according to an embodiment of the present invention is shown.

[0017] Figure 11 A schematic flowchart of a part assembly method according to an embodiment of the present invention is shown. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the described embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. In the following description, some specific embodiments are for descriptive purposes only and should not be construed as limiting the present invention in any way, but are merely examples of embodiments of the present invention. Conventional structures or constructions will be omitted where they may cause confusion in understanding the present invention. It should be noted that the shapes and dimensions of the components in the figures do not reflect actual size and proportion, but are only schematic representations of the embodiments of the present invention.

[0019] Unless otherwise defined, the technical or scientific terms used in the embodiments of this invention should have the ordinary meaning understood by those skilled in the art. The terms "first," "second," and similar terms used in the embodiments of this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components.

[0020] Figure 1 A schematic diagram of a parts assembly apparatus according to an embodiment of the present invention is shown.

[0021] like Figure 1 As shown, the parts assembly device 100 includes a base 1, a robotic arm 2, a connecting mechanism 3, a camera 4, an assembly mechanism 5, and a processor 6.

[0022] In an embodiment of the present invention, the robotic arm 2 is mounted on the base 1. The robotic arm 2 can be a 6-degree-of-freedom robotic arm. Within its reach, the robotic arm 2 can be adjusted to various postures to meet different assembly requirements.

[0023] The connecting mechanism 3 is mounted on the end of the robotic arm 2 away from the base 1. The connecting mechanism 3 is used to connect the tools for assembling parts. The camera 4 and the assembly mechanism 5 are mounted on the connecting mechanism 3. The processor 6 drives the camera 4 and the assembly mechanism 5 to work together through the robotic arm 2 to achieve precise assembly of the parts to be assembled.

[0024] In an embodiment of the present invention, processor 6 is installed inside base 1 and is electrically connected to robotic arm 2, camera 4, and assembly mechanism 5. Camera 4 follows robotic arm 2 to the work platform and acquires an initial image of the target workpiece on the work platform. Processor 6 controls robotic arm 2, camera 4, and assembly mechanism 5 to perform part assembly process, which may include: determining the orientation information of the surface to be assembled in the target workpiece based on the initial image; controlling robotic arm 2 to move camera 4 to the target position based on the orientation information, wherein the orthographic projection of the target position on the surface to be assembled overlaps with the assembly area of ​​the target workpiece; and controlling robotic arm 2 to move assembly mechanism 5 to assemble the part to be assembled into the assembly area based on the target image of the assembly area acquired by camera 4 at the target position.

[0025] In an embodiment of the present invention, the target workpiece is placed on a work platform, and the target workpiece needs to be assembled. For example, the part to be assembled can be a screw, the surface of the target workpiece has a screw hole, and the part assembly device 100 performs the assembly process of driving the screw into the screw hole.

[0026] The initial image is captured by camera 4 at its initial position. For example, after the parts assembly device 100 is activated, it moves to the vicinity of the work platform. The robotic arm 2 can maintain its initial posture. For example, the initial posture is fixed or preset, and the processor 6 does not need to adjust the initial posture of the robotic arm 2 according to the target workpiece. Alternatively, the user can store a preset initial posture in the processor 6 according to the work platform, so that for different work platforms, the robotic arm 2 can move camera 4 to a designated location on the work platform, ensuring the target workpiece is within the field of view of camera 4, without needing to adjust the initial posture according to the target workpiece on the work platform.

[0027] In the initial posture, the viewing angle of camera 4 for different target workpieces may be different, resulting in different angles for the initial images acquired. In the initial posture, the end effector of robotic arm 2 remains stationary, which can reduce motion blur and depth jitter issues of camera 4.

[0028] In embodiments of the present invention, the orientation information may include the orientation and distance information of the target workpiece relative to the assembly mechanism 5, and may also include the placement posture of the target workpiece on the work platform. Based on this, the processor 6 can determine the distance and orientation that the robotic arm 2 needs to move, and determine the orientation of the surface to be assembled, thereby determining the target position.

[0029] Because the orthographic projection of the target position onto the surface to be assembled overlaps with the assembly area of ​​the target workpiece, camera 4 is in a state of looking directly at the assembly area from the target position. For example, the orientation information may include the normal information of the assembly area. Based on the normal information of the assembly area and the normal information of camera 4 itself, processor 6 determines the target position. At the target position, the normal of camera 4 and the normal of the assembly area can overlap in opposite directions, so that camera 4 can look directly at the assembly area.

[0030] For example, if the area to be assembled is a screw hole, with camera 4 viewing the screw hole directly, camera 4 re-captures the details of the screw hole to obtain a target image. Based on the target image, processor 6 can perform detection, segmentation, and geometric measurement on the screw hole to determine information such as the center, radius, and depth of the screw hole.

[0031] The processor 6 calculates the assembly trajectory that the assembly mechanism 5 needs to execute based on information such as the center, radius and depth of the screw hole, and controls the robotic arm 2 to adjust its posture so that the assembly mechanism 5 can move according to the assembly trajectory to perform part assembly.

[0032] For example, the part to be assembled is a screw, and the screw is connected to the assembly mechanism 5. With the movement of the robotic arm 2, the assembly mechanism 5 moves the screw along the assembly trajectory to the screw hole. The assembly mechanism 5 performs the screw-fastening action, combining force control, torque control, and displacement control to complete the screw fastening.

[0033] In an embodiment of the present invention, the processor 6 controls the camera 4 to observe the target workpiece in the initial state of the robotic arm 2, and performs coarse positioning of the target workpiece. Based on the coarse positioning information, the processor 6 then controls the robotic arm 2 to move the camera 4 to a position that directly views the target workpiece, improving the accuracy of the camera 4's observation of the target workpiece. This avoids the need for the processor 6 to adjust the posture of the robotic arm 2 multiple times, simplifies the control logic, reduces interference caused by the camera 4's perspective and the target workpiece's structure, lowers the positioning difficulty of the assembly area, simplifies the assembly difficulty, and improves the assembly accuracy.

[0034] Figure 2 A schematic diagram illustrating the working principle of a parts assembly device according to an embodiment of the present invention is shown.

[0035] like Figure 2 As shown, under the observation of camera 4, assembly mechanism 5 assembles parts in the assembly area A of the surface S of the target workpiece W located on the work platform P.

[0036] In this embodiment of the invention, under the initial posture of the robotic arm 2, the camera 4 acquires an initial image including a color image and a depth image. The processor 6, based on the initial image, determines the orientation information of the surface to be assembled in the target workpiece, which may include: detecting the color image and the depth image to obtain an image mask of the surface to be assembled; using the image mask to remove background from the color image and the depth image to obtain a feature image of the surface to be assembled; and determining the orientation information of the surface to be assembled based on the feature image.

[0037] In this embodiment of the invention, the processor 6 controls the robotic arm 2 to have an initial posture, so that the camera 4 is in an initial observation position. The initial posture can be preset through offline teaching, computer-aided design (CAD) models, or production line layout. At the initial observation position, the angle at which the camera 4 observes the target workpiece is not fixed. For example, the camera 4 acquires an initial image of the target workpiece from an oblique perspective.

[0038] For example, camera 4 could be a color-depth (RGB-D) camera. An RGB-D camera acquires one or more frames of RGB images and depth maps for a target workpiece. In the case where the RGB-D camera acquires multiple initial images, processor 6 can perform temporal filtering, median filtering, or confidence fusion on the multiple depth maps.

[0039] In this embodiment of the invention, processor 6 can use a convolutional neural network to segment the color image and depth image to obtain an image mask. The image mask is used to segment the target region and background region in the color image and depth image. Applying the image mask to the color image and depth image yields a feature image of the surface to be assembled.

[0040] For example, the feature image may include a filtered color image and a filtered depth image. The filtered color image is a color image that has been masked to retain only the target area, and the filtered depth image is a depth image that has been masked to retain only the target area.

[0041] For example, the feature image may include a filtered color-depth tensor. The filtered color-depth tensor is a fused image of the color image and the depth image after being simultaneously cropped and filtered using the same mask. The fused image only includes the three-dimensional data of the target workpiece.

[0042] For example, a convolutional neural network can include a segmentation network model that binarizes color and depth images and outputs a mask image. When multiple objects or obstructions exist in the initial image, the segmentation network model can determine the image mask of the target object based on its category, location, area, depth range, and task ID or instructions from the host computer.

[0043] Processor 6 applies the image mask to both the color image and the depth image. It sets grayscale values ​​outside the masked area in the color image to zero or the mean, and sets depth values ​​outside the masked area in the depth image to invalid depth or identifies them using an additional channel. It filters outlined depth values ​​within the masked area of ​​the image mask and removes pixels that blend with the target workpiece edge and background, forming a 4-channel RGB-D tensor, or a multi-channel tensor including color, depth, masked area, and coordinate encoding.

[0044] For example, processor 6 can first fuse the depth image and the color image to form a color-depth image with 4 channels, and then input the color-depth image into a convolutional neural network to extract and segment the color RGD features and depth D features in the color-depth image, and output a 4-channel RGD-D tensor.

[0045] For example, processor 6 can also input the color image and the depth image into two separate convolutional neural networks. One convolutional neural network extracts and segments the color RGD features to obtain a 3-channel RGB tensor, while the other convolutional neural network extracts and segments the depth D features to obtain a 1-channel D tensor. Feature fusion of the 3-channel RGB tensor and the 1-channel D tensor can also yield a 4-channel RGD-D tensor.

[0046] In this embodiment of the invention, the convolutional neural network may further include a normal prediction network model. The processor 6 can input the feature image into the normal prediction network model and output the normal vector z of the surface S to be assembled in the target workpiece W. The processor 6 performs normalization, orientation consistency judgment, and confidence evaluation on the normal vector z to obtain the normal vector z3 of the surface S to be assembled.

[0047] For example, the z-vector can preferably be defined in the camera 4-coordinate system, and the output of the normal prediction network model can include a 3D unit normal vector, normal confidence, optional surface tilt angle, plane offset, or quality score. The training loss of the normal prediction network model can include cosine similarity loss, L2 loss, angle loss, sign consistency constraint, and confidence loss.

[0048] During the training of the normal prediction network model, training data can be obtained through at least the following methods: real poses provided by calibration fixtures or high-precision jigs; planar poses obtained through robotic arm teaching or external measurement systems; robust planar fitting of high-quality point clouds as weak labels; synthetic RGB-D data generated in a simulation environment under different angles, lighting, textures, depth noise, and occlusion conditions; simulation pre-training plus fine-tuning with real data; and label correction using geometric consistency after successful orthogonal viewing through self-supervised or semi-supervised methods. Training data can be enhanced through methods such as brightness, contrast, noise, depth holes, mask perturbation, random cropping, random rotation, background replacement, and camera noise simulation.

[0049] In this embodiment of the invention, the processor 6 is configured to control the robotic arm 2 to move the camera 4 to a target position based on orientation information. This may include: calculating the target pose of the end effector of the robotic arm 2 according to the orientation information; and adjusting the end effector of the robotic arm 2 to the target pose. Under the target pose, the robotic arm 2 moves the camera 4 to the target position, such that the normal of the surface to be assembled coincides with the normal of the camera 4 in the opposite direction.

[0050] In this embodiment of the disclosure, the processor 6 can calculate the target posture of the end effector of the robotic arm 2 based on the normal vector z3, thereby obtaining the target position, and making the z-axis of the coordinate system of the camera 4 coincide or approximately coincide with the normal vector z3 of the surface to be assembled. For example, Figure 2 The normal vector z1 shown is the normal vector of camera 4. Normal vector z1 coincides with normal vector z3 in opposite directions.

[0051] Processor 6 can define the optical axis direction of camera 4 as the normal vector z1 of camera 4. Based on the current pose of camera 4 and normal vector z3, it calculates the rotation angle of camera 4 from the original direction to the target direction. The original direction is the normal vector z1 of camera 4 before the movement, and the target direction is the normal vector z1 of camera 4 after the movement. The normal vector z1 of camera 4 after the movement coincides with the normal vector z3 in the opposite direction, that is, z1 = -z3.

[0052] For example, the rotation angle can be calculated and determined using axis angle, quaternion, or rotation matrix methods. If a is the original direction and b is the target direction, then the rotation axis v = a × b, and the rotation angle θ = arccos(a·b). The processor 6 determines the target pose of the camera 4 based on the rotation axis and rotation angle, converts the target pose of the camera 4 into the target pose of the end effector of the robotic arm 2, and sends the target pose of the end effector of the robotic arm 2 to the motion processor of the robotic arm 2 for execution.

[0053] In this embodiment of the invention, the processor 6 is configured to control the robotic arm 2 to drive the assembly mechanism 5 to assemble the part to be assembled into the assembly area based on the target image of the assembly area acquired by the camera 4 at the target position. This may include: detecting the assembly area based on the target image to obtain size information, position information and depth information; generating an assembly trajectory for the part to be assembled based on the size information, position information and depth information; and controlling the robotic arm 2 to drive the assembly mechanism 5 to assemble the part to be assembled into the assembly area based on the assembly trajectory.

[0054] For example, the assembly area A can be a screw hole. The size information can be the diameter of the screw hole, the position information can be the coordinates of the screw hole center in the target image, and the depth information can be the depth of the screw hole.

[0055] Since camera 4 is essentially viewing the surface S to be assembled directly, the deformation of the screw hole in the target image is small, making the screw hole closer to a circle, and the depth measurement is more stable. Therefore, accurate information about the assembly area A can be extracted from the target image obtained from the direct viewing angle. For example, screw hole positioning can be achieved using a boundary detection algorithm. The coordinates and depth of the screw hole are obtained by converting the imaging parameters of camera 4, and then the coordinates in the camera 4 coordinate system are converted to the coordinates in the robotic arm 2 coordinate system.

[0056] In this embodiment of the invention, the normal vector z2 of the assembly mechanism 5 is parallel to the normal vector z1 of the camera 4. The plane formed by vectors x1 and y1 in the coordinate system of the camera 4, the plane formed by vectors x2 and y2 in the coordinate system of the assembly mechanism 5, and the plane formed by vectors x3 and y3 in the coordinate system of the area to be assembled A are parallel to each other.

[0057] The normal vector z3 of the assembly region A is also used to determine the assembly direction of the parts to be assembled. For example, processor 6 determines the screw installation direction based on the normal vector z3. The driving direction is opposite to the normal vector z3, i.e., the installation direction is -z3. Alternatively, the direction of locking from the outside of the surface to be assembled inward is defined as the installation direction.

[0058] Based on information such as the installation direction, the offset between camera 4 and assembly mechanism 5, the length of the part to be assembled, and the safety distance, processor 6 generates the contact point, feed segment, locking segment, and exit segment of the part to be assembled during the assembly process, thus obtaining the assembly trajectory. During the assembly process, force control and torque control are combined to compensate for attitude errors and assembly deviations, completing the locking process.

[0059] For example, during assembly, the normal vector z2 and normal vector z3 coincide in opposite directions, which aligns the axis of the part to be assembled with the assembly area A. If the assembly operation is a screw-driving task, the assembly mechanism 5 can be set to the forward tightening mode and perform forward tightening to screw the screw into the threaded hole. If the assembly operation is a screw-loosening task, the assembly mechanism 5 can be set to the reverse disassembly mode and perform reverse tightening to screw the screw out of the threaded hole.

[0060] In embodiments of the present invention, a mask image is used to segment the initial image and extract the target region including the target workpiece. This reduces background interference in the normal vector determination process and reduces noise in image recognition. In the target image acquired from a frontal view, the deformation of the region to be assembled is relatively small. Based on the target image, relevant information about the region to be assembled can be accurately obtained. By using the normal vector of the region to be assembled, the assembly trajectory can be planned, achieving unification between the perception result and the execution trajectory.

[0061] Figure 3 A partial structural schematic diagram of a parts assembly apparatus according to an embodiment of the present invention is shown.

[0062] like Figure 3 As shown, the assembly mechanism 5 includes a floating component 51 and an assembly component 52. The floating component 51 is mounted on the connecting mechanism 3, and the assembly component 52 is mounted on the floating component 51. The floating component 51 is configured to control the assembly component 52 to float along the axial direction AA' of the end 21 of the robotic arm 2.

[0063] In an embodiment of the present invention, the assembly direction and floating direction of the assembly component 52 are parallel to the axial direction AA' of the end 21 of the robotic arm 2.

[0064] During the assembly process, the assembly component 52 needs to gradually move closer to or further away from the surface to be assembled. Therefore, the floating component 51 can move along the axial direction AA' of the end 21 of the robotic arm 2, thereby driving the assembly component 52 to move along the axial direction AA' of the end 21 of the robotic arm 2.

[0065] In an embodiment of the invention, a sensor 31 is mounted on the connecting mechanism 3, and the sensor 31 is configured to collect torque data of the assembly mechanism 5. Based on the torque data, the processor 6 controls the amount of movement of the robotic arm 2 during the assembly process.

[0066] For example, sensor 31 can be a six-dimensional force sensor for force assessment and collision detection. The six-dimensional force sensor can collect contact force and torque data of the parts to be assembled in real time. Processor 6 determines whether the parts to be assembled have reached a preset locking threshold or are in a disassembly complete state.

[0067] If the assembly process is determined to be complete, the processor 6 can control the assembly component 52 to stop performing the assembly action. If the assembly process is determined to be incomplete, the processor 6 can continue to rotate the part to be assembled and update the determination until the assembly is completed. In addition, abnormal jamming, misalignment, and other phenomena may occur during the assembly process. The processor 6 can calculate the micro-momentum of the robotic arm 2 by outputting the torque vector value to the sensor 31, thereby realizing force-position hybrid control.

[0068] In an embodiment of the invention, the assembly component 52 includes a wrench 521 and a bit 522. The wrench 521 is mounted on the floating component 51, and the axial direction of the wrench 521 is parallel to the axial direction AA' of the end effector 21 of the robotic arm 2. The bit 522 is mounted on the actuating end of the wrench 521 and is configured to mount the part to be assembled.

[0069] The bit 522 can be detachably connected to the part to be assembled. For example, a screw is connected to the bit 522, and the wrench 521 drives the bit 522 to rotate so as to screw the screw into or out of the screw hole.

[0070] In an embodiment of the invention, the robotic arm 2 moves the wrench 521 so that the socket axis of the wrench 521 is aligned with the screw axis, and the socket is inserted into the screw head. The processor 6 starts the wrench 521 and rotates it according to the set torque and speed, monitoring the rotation angle and torque in real time to tighten or loosen the screw.

[0071] The floating component 51 can compensate for depth measurement errors during the assembly process. When the depth deviation is large, the actual depth of the screw hole can be measured by the sensor 31 and the floating component 51, achieving flexible contact between the screw and the screw hole and ensuring that the target workpiece is not damaged.

[0072] Figure 4 A partial structural schematic diagram of a parts assembly apparatus according to another embodiment of the present invention is shown.

[0073] like Figure 4 As shown, the assembly mechanism 5 includes a floating component 51, an assembly component 52, an air suction component 53, and a quick-change component 54.

[0074] In this embodiment of the invention, a suction component 53 is mounted to the actuating end of a wrench 521, and the suction component 53 is connected to a screwdriver bit 522. For example, the wrench 521 and the screwdriver bit 522 are mechanically connected via the suction component 53. The suction component 53 is configured to provide an adsorption force through the screwdriver bit 522, causing the part to be assembled to be adsorbed onto the screwdriver bit 522.

[0075] For example, the suction component 53 provides suction force along the axial direction of the bit 522. The suction direction is from the end of the suction component 53 towards the interior of the suction component 53, for example... Figure 2 The opposite direction of the normal vector z2 shown.

[0076] When the bit 522 needs to hold the part to be assembled, the suction component 53 provides suction force along the suction direction, so that the part to be assembled is stably connected to the bit 522. When the bit 522 needs to replace the part to be assembled, the suction component 53 stops providing suction force, so that the part to be assembled is detached from the bit 522.

[0077] In this embodiment of the invention, the quick-change component 54 is sleeved on the bit 522, and the quick-change component 54 is detachably connected to the air suction component 53. The quick-change component 54 is configured to fix the bit 522. For example, the air suction component 53 and the bit 522 are mechanically connected through the quick-change component 54.

[0078] The quick-change part 54 needs to be compatible with the bit 522. The quick-change part 54 is detachable and can be removed and replaced to accommodate different sizes and structures of bit 522.

[0079] For example, quick-change component 54 can be pre-assembled with bit 522. When bit 522 needs to be replaced, it can be done directly by replacing quick-change component 54.

[0080] In this embodiment of the invention, the air suction component 53 provides adsorption assistance for the screwdriver bit 522 to connect with the part to be assembled, ensuring a stable connection between the screwdriver bit 522 and the part, preventing displacement of the part during assembly and improving the stability of subsequent assembly processes. The quick-change component 54 enables rapid replacement of the screwdriver bit 522, simplifying the replacement process.

[0081] Figure 5 A schematic diagram of the structure of a floating component according to an embodiment of the present invention is shown.

[0082] like Figure 5 As shown, the floating component 51 includes a bracket 511, a guide rod 512, a spring 513, a slide rail 514, a slider 515, and a fixing block 516.

[0083] In this embodiment of the invention, the bracket 511 includes a first sub-part 5111 and two second sub-parts 5112 disposed opposite to each other along the axial direction AA' of the end 21 of the robotic arm 2. The first sub-part 5111 has a limiting hole H1, and the two second sub-parts 5112 have through holes H2 disposed opposite to each other.

[0084] The bracket 511 may include four second sub-parts 5112, which are fixedly disposed at the four corners of the first sub-part 5111, with two of the second sub-parts 5112 being disposed opposite to each other. The two opposing second sub-parts 5112 are arranged along the axial direction AA' of the end 21 of the robotic arm 2, and the through-holes H2 disposed opposite to each other on the two second sub-parts 5112 are also disposed along the axial direction AA' of the end 21 of the robotic arm 2.

[0085] The guide rod 512 is installed between the two second sub-parts 5112 through the through hole H2, and the guide rod 512 extends along the axial direction AA' of the end 21 of the robotic arm 2. The guide rod 512 passes through the two through holes H2 that are arranged opposite each other.

[0086] For example, the floating component 51 includes two guide rods 512. The two guide rods 512 pass through two sets of oppositely arranged through holes H2. The guide rods 512 limit the floating direction of the floating component 51, ensuring that the floating direction does not deviate.

[0087] Spring 513 is sleeved on guide rod 512, with one end of spring 513 abutting against second sub-part 5112. Slide rail 514 is installed in first sub-part 5111 near guide rod 512 and is arranged parallel to guide rod 512. Slider 515 is slidably mounted on slide rail 514, configured to slide on slide rail 514, with slider 515 abutting against the other end of spring 513.

[0088] The two ends of the spring 513 abut against the second sub-part 5112 and the slider 515 respectively, and the position of the second sub-part 5112 remains fixed. The slider 515 is also sleeved on the guide rod 512. Under the limitation of the guide rod 512, the slider 515 can slide along the slide rail 514, and drive the spring 513 to deform along the direction of the guide rod 512. The spring 513 provides a buffer for the slider 515, and the slide rail 514 can reduce the sliding friction of the slider 515.

[0089] The fixing block 516 is fixed on the slider 515. The fixing block 516 has a mounting hole H3. The assembly component 52 is installed on the fixing block 516 through the mounting hole H3. The fixing block 516 drives the assembly component 52 to slide on the slide rail 514 following the slider 515, and fixes its position through the limiting hole H1.

[0090] The wrench 521 of the assembly component 52 passes through the mounting hole H3 and is fixedly mounted on the fixing block 516 with screws. During the sliding of the slider 515, the slider 515 drives the wrench 521 to move together along the axial direction AA' of the end 21 of the robotic arm 2 through the fixing block 516, so that the wrench 521 can float up and down, thus solving the error of the z-axis during screw driving.

[0091] After the movement is completed, the screws on the fixing block 516 can be screwed into the limiting hole H1 on the first sub-part 5111, so that the limiting hole H1 limits the position of the fixing block 516.

[0092] A wear-resistant sleeve 517 can also be provided between the spring 513 and the slider 515, and a wear-resistant sleeve 517 can also be provided between the slider 515 and the second sub-part 5112. The wear-resistant sleeve 517 can reduce the friction between the spring 513 and the slider 515 and the slider 515 and the second sub-part 5112.

[0093] In this embodiment of the invention, the floating component 51 includes two guide rods 512 and two sliders 515. The structure of the double guide rods and double sliders allows the floating component 51 to move smoothly and withstand large torques. Furthermore, even if one guide rod or slider fails, the other guide rod and slider can still function effectively.

[0094] Figure 6 A schematic diagram of the structure of the air suction component and the quick-change component according to an embodiment of the present invention is shown.

[0095] like Figure 6 As shown, the air suction component 53 includes an air duct 531, an output shaft 532, a sealing ring 533, a conversion shaft 534, a coupling 535, a connector 536, and an air duct 537.

[0096] In this embodiment of the invention, an air suction channel 5311 is provided inside the air duct 531. An output shaft 532 is located in the air suction channel 5311 and is connected to the actuating end of a wrench 521. A sealing ring 533 is fitted onto the output shaft 532 and abuts against the air duct 531. A conversion shaft 534 is connected to the output shaft 532 in the air suction channel 5311, and is configured to connect a screwdriver bit 522. A coupling 535 is fitted at the connection between the output shaft 532 and the conversion shaft 534. A connector 536 is provided on the outer wall of the air duct 531 and communicates with the air suction channel 5311. An air duct 537 connects to and communicates with the connector 536, and is configured to provide suction force to the air suction channel 5311.

[0097] The conversion shaft 534 uses its own set screw to fix the output shaft 532 and the coupling 535 together, ensuring that the conversion shaft 534, the output shaft 532 and the coupling 535 are coaxial.

[0098] The trachea 537 is connected to the inhalation channel 5311 via the connector 536, and the trachea 537 provides air to the inhalation channel 5311. Figure 6 The thick black line shown represents the airflow path R. It should be noted that the thick black line is only schematically shown to indicate the airflow path R; there is no actual thick black line in the air intake component 53.

[0099] The sealing ring 533 is located away from the direction of the air passage R. The sealing ring 533 abuts against the outer wall of the output shaft 532 and the inner wall of the air duct 531 to seal the gap between the output shaft 532 and the air duct 531 and prevent gas from leaking from the port of the air duct 531 near the output shaft 532.

[0100] The output shaft 532, conversion shaft 534, and screwdriver bit 522 are all located within the air suction channel 5311. The wrench 521 drives the output shaft 532 to rotate, and the output shaft 532 then drives the screwdriver bit 522 to rotate via the conversion shaft 534. The rotation of the output shaft 532, conversion shaft 534, and screwdriver bit 522 all occurs within the air suction channel 5311. During the rotation of the output shaft 532, conversion shaft 534, and screwdriver bit 522, the air duct 531 remains stationary.

[0101] During assembly, the air duct 531 will not rotate with the bit 522, which prevents the air duct 537 from getting tangled on the air duct 531. In addition, since the air duct 531 does not rotate, the output shaft 532, the conversion shaft 534 and the bit 522 only rotate within the air suction channel 5311, which ensures the airtightness of the air duct 531.

[0102] In this embodiment of the invention, the quick-change component 54 includes a housing 541, a flange 542, a suction cup 543, and a suction nozzle 544.

[0103] The outer casing 541 is fitted onto the bit 522, and the inner wall of the outer casing 541 abuts against the suction unit 53. A flange 542 secures the bit 522 to the interior of the outer casing 541. A suction cup 543 is installed in the outer casing 541 at the end furthest from the suction unit 53 and is fitted onto the bit 522. A suction nozzle 544 is mounted on the suction cup 543 and is configured to suction the part to be assembled.

[0104] The inner wall of the outer casing 541 abuts against the outer wall of the duct 531, allowing air from the duct 531 to be drawn into the outer casing 541. A rotary sealing ring 545 may also be fitted onto the duct 531. The rotary sealing ring 545 is located in a groove on the inner side of the outer casing 541 to improve the sealing between the duct 531 and the outer casing 541.

[0105] Flange 542 is fitted onto bit 522, and bit 522 is fixed to housing 541 via flange 542. Suction cup 543 is fitted onto bit 522, and the outer wall of suction cup 543 abuts against the inner wall of housing 541 to ensure the airtightness of housing 541 and guide air into suction cup 543.

[0106] The screwdriver bit 522 extends from the air suction channel 5311 of the air duct 531 into the housing 541 and is fixed to the suction cup 543. The air path R extends along the air duct 537 to the extension direction of the screwdriver bit 522 to provide air suction assistance for the screwdriver bit 522. The output port of the suction cup 543 can be filled in the middle of the screwdriver bit 522 to ensure the air path is sealed.

[0107] The suction cup 543 can be made of silicone and can assist the screwdriver bit 522 in holding screws or washers while assembling parts. A sealing lip is provided at the bottom of the suction cup 543 to improve airtightness and its own stability.

[0108] The suction nozzle 544 guides the airflow path R. The nozzle 544 matches the suction cup 543 to adsorb different parts to be assembled. The nozzle 544 is detachable; by replacing the nozzle with one of different structures, different parts can be adsorbed, thus reducing costs. If the suction cup 543 is damaged, it can also be quickly replaced, further reducing costs.

[0109] Figure 7 A schematic diagram of a parts assembly apparatus according to another embodiment of the present invention is shown.

[0110] like Figure 7 As shown, the parts assembly device 100 may further include a moving mechanism 8. The moving mechanism 8 is mounted below the base 1 and is configured to move the base 1.

[0111] In this embodiment of the disclosure, the processor 6 can control the moving mechanism 8 to move autonomously, so that the moving mechanism 8 drives the assembly component 52 to move to the vicinity of the corresponding work platform.

[0112] A first lidar 81 is mounted on the mobile mechanism 8. The first lidar 81 is configured to scan the features of obstacles on a specific plane. For example, the first lidar 81 is a lidar with a 2D field of view, and the specific plane can be the plane on which the mobile mechanism 8 travels.

[0113] During the movement of the mobile mechanism 8, the first lidar 81 can detect the environment along the movement path for real-time positioning and map building. The first lidar 81 sends the collected planar environmental information, such as obstacle information, to the processor 6, which, in conjunction with the processor 6, provides navigation and obstacle avoidance services to the mobile mechanism 8.

[0114] In this embodiment of the disclosure, a second lidar 10 is mounted on the base 1. The second lidar 10 is configured to perform feature scanning of obstacles in the workspace of the parts assembly device 100.

[0115] For example, the second lidar 10 is a lidar with a 3D view, and its working space is the three-dimensional space where the parts assembly device 100 is located. The second lidar 10 can detect obstacle information and environmental information in the front, back, left, right, up, and down directions of the parts assembly device 100, and send the detected information to the processor 6. Based on the obstacle information and environmental information from the second lidar 10, the processor 6 analyzes the current position and current environment of the parts assembly device 100, and the mobile mechanism 8 provides navigation and obstacle avoidance services. For example, based on the map obtained by real-time localization and map construction, the processor 6 controls the mobile mechanism 8 to autonomously travel to the coarse positioning point corresponding to the working platform, so that the working space of the robotic arm covers the target area.

[0116] Combination Figure 8 and Figure 9 The mounting base 7 is illustrated schematically. Figure 8 A partial structural schematic diagram of a parts assembly apparatus according to another embodiment of the present invention is shown. Figure 9 A schematic diagram of the mounting base according to an embodiment of the present invention is shown.

[0117] like Figure 8 and Figure 9 As shown, the parts assembly device 100 also includes a mounting base 7. The mounting base 7 is mounted on one side of the base 1, and the mounting base 7 has a plurality of opening slots 71 for placing a plurality of quick-change parts 54.

[0118] In this embodiment of the present disclosure, the processor 6 can control the end of the robotic arm 2 to move above the mounting base 7, place the quick-change component 54 to be replaced in the empty opening slot 71, and install the quick-change component 54 stored in the opening slot 71 to the end of the air suction component 53.

[0119] The quick-change part 54 that needs to be replaced can be moved in the reverse direction of the opening of the slot 71 and placed in the slot 71. The quick-change part 54 that needs to be newly installed can be removed along the opening direction of the slot 71.

[0120] The opening slot 71 can be a U-shaped structure, and the cross-section of each opening slot 71 can be a gradually changing V-shaped structure. For example, the upper surface of the opening slot 71 is inclined. The gradually changing V-shaped structure facilitates the quick replacement of the quick-change component 54 and can also increase the guidance for taking out and placing the quick-change component 54.

[0121] Combination Figure 6 and Figure 9 A buffer 546 can also be provided on the outer arm of the outer casing 541. The bottom surface of the buffer 546 is inclined and fits with the upper surface of the opening groove 71, so that the quick-change component 54 can be stably placed in the opening groove 71.

[0122] The buffer 546 also includes a spring and a limiting part. When the quick-change component 54 is placed in the opening slot 71, the buffer 546 can move away from the opening slot 71, and the internal spring of the buffer 546 is compressed towards the limiting part. Under the action of the spring, the buffer 546 can press against the upper surface of the opening slot 71, so that the quick-change component 54 can be stably placed in the opening slot 71.

[0123] Combination Figure 8 and Figure 10 A schematic illustration of tray 9 is provided. Figure 10 A schematic diagram of the structure of the tray holder according to an embodiment of the present invention is shown.

[0124] like Figure 8 and Figure 10 As shown, a tray 9 is mounted on the upper surface of the base 1, and the tray 9 is configured to hold the parts to be assembled.

[0125] The tray 9 is provided with limiting holes for placing the parts to be assembled. For example, the limiting holes may be multiple screw holes in which screws are placed.

[0126] For example, during the picking phase, the assembly mechanism 5 can use air suction to remove the screw from the screw hole. During the placement phase, after aligning the assembly mechanism 5 with the through hole, the air suction is disengaged, and the screw is placed in the screw hole.

[0127] For example, during the retrieval phase, the assembly mechanism 5 can rotate the bit 522 and adjust the posture of the robotic arm 2 to unscrew multiple threads from the screw hole and remove the screw. During the placement phase, the assembly mechanism 5 can rotate the bit 522 and adjust the posture of the robotic arm 2 to screw the screw in 2-3 threads and place the screw in, thus achieving vertical placement of the screw and storage of multiple screws.

[0128] Figure 11 A schematic flowchart of a part assembly method according to an embodiment of the present invention is shown.

[0129] like Figure 11 As shown, the part assembly method 200 includes operations S201 to S203.

[0130] In an embodiment of the present invention, the part assembly method 200 can be applied to the part assembly apparatus 100 described above.

[0131] In operation S201, the orientation information of the surface to be assembled in the target workpiece is determined based on the initial image acquired by the camera.

[0132] In operation S202, based on the orientation information, the robotic arm is controlled to move the camera to the target position. The orthographic projection of the target position on the surface to be assembled overlaps with the assembly area of ​​the target workpiece.

[0133] In operation S203, based on the target image of the assembly area captured by the camera at the target position, the robotic arm is controlled to drive the assembly mechanism to assemble the parts to be assembled into the assembly area.

[0134] In the embodiments of the present invention, the assembly process of operations S201 to S203 is similar to the assembly process performed by the part assembly device 100 described above, and will not be repeated here.

[0135] The block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram or flowchart, and combinations of blocks in the block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0136] Those skilled in the art will understand that the features described in the various embodiments of the present invention can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in the present invention. In particular, the features described in the various embodiments of the present invention can be combined and / or combined in various ways without departing from the spirit and teachings of the present invention. All such combinations and / or combinations fall within the scope of the present invention.

[0137] The embodiments of the present invention have been described above. However, these embodiments are merely illustrative and not intended to limit the scope of the invention. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of the invention, and all such substitutions and modifications should fall within the scope of the invention.

Claims

1. A parts assembly apparatus, comprising: Base (1); A robotic arm (2) is mounted on the base (1); A connecting mechanism (3) is installed at the end of the robotic arm (2) away from the base (1); A camera (4) is mounted on the connecting mechanism (3). The camera (4) is configured to follow the robotic arm (2) to the work platform and acquire an initial image of the target workpiece in the work platform. Assembly mechanism (5) is installed on the connecting mechanism (3); as well as A processor (6), installed inside the base (1), is electrically connected to the robotic arm (2), the camera (4), and the assembly mechanism (5). The processor (6) is configured to: Based on the initial image, determine the orientation information of the surface to be assembled in the target workpiece; Based on the orientation information, the robotic arm (2) is controlled to move the camera (4) to the target position, and the orthographic projection of the target position on the surface to be assembled overlaps with the assembly area of ​​the target workpiece. as well as Based on the target image of the assembly area captured by the camera (4) at the target location, the robotic arm (2) is controlled to drive the assembly mechanism (5) to assemble the parts to be assembled into the assembly area.

2. The parts assembly device according to claim 1, wherein, The initial image includes a color image and a depth image; The processor (6) is configured to determine the orientation information of the surface to be assembled in the target workpiece based on the initial image, including: The color image and the depth image are detected to obtain an image mask for the surface to be assembled. The background is removed from the color image and the depth image using the image mask to obtain a feature image of the surface to be assembled; and Based on the feature image, the orientation information of the surface to be assembled is determined.

3. The parts assembly device according to claim 1, wherein, The processor (6) is configured to control the robotic arm (2) to move the camera (4) to the target position based on the orientation information, including: Based on the orientation information, calculate the target posture of the end effector of the robotic arm (2); and Adjust the end effector of the robotic arm (2) to the target posture; In the target posture, the robotic arm (2) drives the camera (4) to the target position, so that the normal of the surface to be assembled coincides with the normal of the camera (4) in the opposite direction.

4. The parts assembly device according to claim 1, wherein, The processor (6) is configured to control the robotic arm (2) to drive the assembly mechanism (5) to assemble the part to be assembled into the assembly area based on the target image of the area to be assembled acquired by the camera (4) at the target location, including: Based on the target image, the area to be assembled is detected to obtain size information, position information and depth information; Based on the size information, position information, and depth information, an assembly trajectory for the part to be assembled is generated; and The robotic arm (2) is controlled to drive the assembly mechanism (5) to assemble the part to be assembled into the assembly area based on the assembly trajectory.

5. The parts assembly device according to claim 1, wherein, The assembly mechanism (5) includes: The floating component (51) is mounted on the connecting mechanism (3); and Assembly component (52) is mounted on the floating component (51); The floating component (51) is configured to control the assembly component (52) to float along the axial direction of the end of the robotic arm (2).

6. The parts assembly apparatus according to claim 5, wherein, The floating component (51) includes: The bracket (511) includes a first sub-part (5111) and two second sub-parts (5112) arranged opposite each other along the axial direction of the end of the robotic arm (2). The first sub-part (5111) has a limit hole (H1) and the two second sub-parts (5112) have through holes (H2) arranged opposite each other. A guide rod (512) is installed between the two second sub-parts (5112) through the through hole (H2), and the guide rod (512) extends along the axial direction of the end of the robotic arm (2); A spring (513) is sleeved on the guide rod (512), and one end of the spring abuts against the second sub-part (5112); A slide rail (514) is installed in the first sub-part (5111) on the side near the guide rod (512) and is arranged parallel to the guide rod (512); A slider (515) is slidably mounted on the slide rail (514), the slider being configured to slide on the slide rail (514), and the slider (515) abuts against the other end of the spring (513); A fixing block (516) is fixed on the slider (515). The fixing block (516) has a mounting hole (H3). The assembly component (52) is installed on the fixing block (516) through the mounting hole (H3). The fixing block (516) drives the assembly component (52) to slide on the slide rail (514) following the slider (515), and fixes its position through the limiting hole (H1).

7. The parts assembly apparatus according to claim 5, wherein, The assembly component (52) includes: A wrench (521) is mounted on the floating component (51), the axial direction of the wrench (521) being parallel to the axial direction of the end of the robotic arm (2); and A bit (522) is mounted on the actuating end of the wrench (521), and the bit (522) is configured to mount the part to be assembled.

8. The parts assembly apparatus according to claim 7, wherein, The assembly mechanism (5) further includes: A suction component (53) is installed on the actuating end of the wrench (521) and connected to the bit (522). The suction component (53) is configured to provide suction force through the bit (522) so that the part to be assembled is adsorbed onto the bit (522).

9. The parts assembly apparatus according to claim 8, wherein, The air intake component (53) includes: The air duct (531) has an internal air intake channel (5311). The output shaft (532) is located in the air intake channel (5311) and is connected to the actuating end of the wrench (521); A sealing ring (533) is fitted onto the output shaft (532) and abuts against the air duct (531); The conversion shaft (534), which connects to the output shaft (532) in the air intake channel (5311), is configured to connect to the bit (522); A coupling (535) is fitted at the connection between the output shaft (532) and the conversion shaft (534); A connector (536) is disposed on the outer wall of the air duct (531) and communicates with the air intake channel (5311); and The air tube (537), which connects to and communicates with the connector (536), is configured to provide adsorption force to the air intake channel (5311).

10. The parts assembly apparatus according to claim 8, wherein the assembly mechanism (5) further comprises: A quick-change component (54) is fitted onto the bit (522) and is detachably connected to the air suction component (53). The quick-change component (54) is configured to fix the bit (522).

11. The parts assembly apparatus according to claim 10, wherein, The quick-change component (54) includes: The outer casing (541) is fitted onto the bit (522), and the inner wall of the outer casing (541) abuts against the air suction component (53); A flange (542) secures the bit (522) to the interior of the housing (541); A suction cup (543) is installed in the housing (541) at one end away from the air suction component (53) and is fitted onto the bit (522); and The suction nozzle (544) is mounted on the suction cup (543) and is configured to adsorb the part to be assembled.

12. The parts assembly apparatus according to claim 10, further comprising: Mounting base (7) is installed on one side of the base (1). The mounting base (7) has multiple opening slots (71) for placing multiple quick-change parts (54).

13. The parts assembly apparatus according to claim 1, wherein, A sensor (31) is mounted on the connecting mechanism (3), and the sensor (31) is configured to collect torque data of the assembly mechanism (5); The processor (6) is also configured to control the amount of movement of the robotic arm (2) during the assembly process based on the torque data.

14. The parts assembly apparatus according to claim 1, further comprising: A moving mechanism (8) is installed below the base (1) and is configured to move the base (1); The moving mechanism (8) is equipped with a first lidar (81), which is configured to perform feature scanning of obstacles on a specific plane.

15. The parts assembly apparatus according to claim 1, wherein, The upper surface of the base (1) is fitted with a tray (9), which is configured to hold the parts to be assembled.

16. The parts assembly apparatus according to claim 1, wherein, A second lidar (10) is mounted on the base (1), and the second lidar (10) is configured to perform feature scanning of obstacles in the workspace of the parts assembly device.

17. A method for assembling parts, applied to the part assembly apparatus according to any one of claims 1 to 16, the method comprising: Based on the initial image acquired by the camera (4), the orientation information of the surface to be assembled in the target workpiece is determined; Based on the orientation information, the robotic arm (2) is controlled to move the camera (4) to the target position, and the orthographic projection of the target position on the surface to be assembled overlaps with the assembly area of ​​the target workpiece. as well as Based on the target image of the assembly area captured by the camera (4) at the target location, the robotic arm (2) is controlled to drive the assembly mechanism (5) to assemble the part to be assembled into the assembly area.

Images