A palletizing robot, system and method based on a posture adjustment function

Abstract

The application discloses a kind of based on posture adjustment function's stacking robot, system and stacking method, including slide rail, two symmetrical installation plate being arranged on slide rail, and respectively installing driving assembly, left clamping plate and right clamping plate on each installation plate;The driving assembly includes linear cylinder, and the output end of the linear cylinder is connected with horizontal shaft;Two "eight" shaped guide rails are arranged on each installation plate, and sliding plate is slidably connected on each guide rail;Sliding bearing is fixed on the sliding plate, and the horizontal shaft slides through two sliding bearings;The left clamping plate and right clamping plate are respectively installed on two sliding plates, and synchronous inward or outward movement is realized by the pushing of linear cylinder, and four-corner type clamping structure is formed.The application designs four-corner type clamping structure, increases constraint, from a passive "clamping" device, upgrades to an active "positioning" device.It is convenient to adjust the posture of carton.

Description

Technical Field

[0001] This invention relates to the field of palletizing technology, specifically to a palletizing robot, system, and method based on posture adjustment functionality. Background Technology

[0002] In existing technologies, the following are generally used: Figure 5 The flatbed palletizing robot shown uses two symmetrical flat clamps 100 mounted on the bottom of the guide rail 200 to grip the goods. Some also use top suction cups to extract the goods.

[0003] Using flat-plate clamps can easily interfere with the stacking of goods during palletizing and makes it difficult to adjust the posture of the goods. Suction cup clamps have a narrow range of applications, and some goods are not easy to pick up. Summary of the Invention

[0004] The purpose of this invention is to provide a palletizing robot, system, and palletizing method based on posture adjustment function to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a palletizing robot based on posture adjustment function, comprising a robot arm and a gripping assembly installed at the end of the robot arm:

[0006] The clamping assembly includes a slide rail, two mounting plates symmetrically arranged on the slide rail, and a drive assembly, a left clamping plate, and a right clamping plate respectively mounted on each mounting plate;

[0007] The drive assembly includes a linear cylinder, the output end of which is connected to a horizontal shaft;

[0008] Each mounting plate is equipped with two guide rails arranged in a figure-eight shape, and a sliding plate is slidably connected to each guide rail;

[0009] The sliding plate is fixed with sliding bearings, and the horizontal shaft slides through two sliding bearings.

[0010] The left and right clamping plates are respectively mounted on two sliding plates and move synchronously inward or outward by the push of a linear cylinder, forming a four-corner clamping structure.

[0011] Preferably, the clamping surface of the left clamping plate is an upwardly flared inclined structure, and the clamping surface of the right clamping plate is a flat structure.

[0012] Preferably, the right clamp is rotatably connected to the sliding plate via a horizontal rotating shaft, and is driven by a motor to achieve fine-tuning of rotation around the rotating shaft.

[0013] Preferably, one side of the right clamp is provided with a downwardly extending elastic pressure roller, which is used to press down adjacent cartons when placing cartons to prevent them from shifting.

[0014] Preferably, the left clamp is longer than the right clamp and has multiple contact sensors on its bottom inner side for detecting whether the carton is placed in place.

[0015] Preferably, the left clamp and / or right clamp are configured to rotate about a vertical axis to adjust the angle of the carton in the horizontal plane.

[0016] Preferably, the clamping assembly is rotatably mounted on the robot arm via a slide rail.

[0017] The present invention also discloses a sensing and control system for the above-mentioned palletizing robot, comprising:

[0018] The perception module, including an RGB-D camera and a LiDAR, is used to acquire 3D images and point cloud data of the cargo stack;

[0019] The data processing module is used for multi-source data fusion, carton posture parameter extraction, and offset calculation.

[0020] The stability assessment module is used to calculate intra-layer alignment, inter-layer support ratio, and center of gravity stability margin.

[0021] The decision control module is used to generate placement strategies based on stability scores and control the robot's execution.

[0022] The learning optimization module is used to optimize the stability prediction model based on historical data.

[0023] The stability assessment module calculates the overall stability score using the following formula: ,

[0024] Where S is the overall stability score. For effective positional deviation, For stability margin, These are the weighting coefficients.

[0025] This invention also discloses an adaptive palletizing method for palletizing robots, comprising the following steps:

[0026] The three-dimensional position and orientation of the cardboard box are obtained through multi-sensor fusion.

[0027] Calculate the deviation of the carton from the ideal position, the alignment within the layers, the interlayer support ratio, and the center of gravity stability margin;

[0028] Assess placement risk level based on stability scoring model;

[0029] Generate a set of candidate placement locations based on the scoring results, and select the optimal location;

[0030] Control the robot to adjust the posture of the end effector to achieve precise placement of the cardboard box;

[0031] Record the actual stacked data and optimize the scoring model parameters through machine learning.

[0032] Compared with the prior art, the beneficial effects of the present invention are: the present invention designs a four-corner clamping structure, which increases constraints and completely limits the in-plane movement (X, Y, θz) of the carton, upgrading it from a passive "clamping" device to an active "positioning" device. This facilitates the adjustment of the carton's posture. Attached Figure Description

[0033] Figure 1 This is a schematic diagram showing the working state of the present invention;

[0034] Figure 2 This is a schematic diagram of the clamping component structure of the present invention;

[0035] Figure 3 This is a schematic diagram of the clamp mounting structure of the present invention;

[0036] Figure 4 This is a schematic diagram of the right clamping plate mounting structure of the present invention;

[0037] Figure 5 This is a schematic diagram of an existing flat plate clamping structure. Detailed Implementation

[0038] 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 embodiments of the present invention, and not all embodiments. Based on the 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.

[0039] like Figures 1 to 4 As shown, the present invention provides a palletizing robot based on posture adjustment function. Its main inventive concept is to grip goods by clamping them at the four corners, mainly for cardboard box packaged goods, or other similar goods with straight edges.

[0040] Reference Figure 2 As shown, it includes a slide rail 1, two mounting plates 9 respectively disposed on the slide rail 1, each mounting plate 9 having a drive assembly, a left clamping plate 3 and a right clamping plate 2.

[0041] The drive assembly includes a linear cylinder 8, the output end of which is connected to a horizontal shaft 4. It also includes a sliding plate 6, the bottom of which is slidably connected to a guide rail 7. Two sliding plates 6 and two guide rails 7 are each mounted on a mounting plate 9, arranged separately on the left and right sides, and the two guide rails 7 are arranged in a figure-eight shape. Figure 3 As shown.

[0042] A sliding bearing 5 is fixedly connected to the upper end face of the sliding plate 6, and the horizontal shaft 4 is slidably inserted into the two sliding bearings 5. (Refer to...) Figure 3 Two sliding bearings 5 ​​are located on the left and right sides of the horizontal shaft 4, respectively, and are slidably connected to the horizontal shaft 4.

[0043] The left clamp 3 is mounted on one of the sliding plates, and the right clamp 2 is mounted on the other sliding plate.

[0044] Continue to refer to Figure 3 When the linear cylinder 8 pushes forward, the two mounting plates 9, the writer's plate 3 and the right clamping plate 2, move inward and outward, respectively.

[0045] The core advantage of the four-corner gripper in this embodiment lies in the added constraint, completely limiting the in-plane movement (X, Y, θz) of the carton, upgrading it from a passive "clamping" device to an active "positioning" device. The required clamping force is less than that of traditional planar clamping-plate robots. Even with a slightly lower clamping force, the carton is confined within the space formed by the four right angles, exhibiting strong resistance to interference (such as robot acceleration and vibration). This significantly reduces the risk of the carton shifting or rotating during handling, improving process reliability. Forced "centering" is achieved, ensuring that the carton's posture relative to the gripper is unique and precisely known after each gripping.

[0046] Furthermore, in practical applications, this structure proves suitable for lightweight cardboard boxes with smooth surfaces or those prone to deformation. It can even correct slight deformations of the box itself, ensuring reliable robot vision recognition. Moreover, this clamping method only clamps a portion of the box (in actual clamping, only the upper part needs to be clamped), leaving most of the box exposed to the vision camera. This significantly aids in coordinate recognition by the algorithm and avoids obstruction of the box by existing clamping components.

[0047] It's worth noting that this structure provides a reliable "initial state value" for the robot's recognition algorithm. With a two-plate gripper, even if the algorithm accurately detects the position and angle of the cartons on the conveyor belt through the vision system, the cartons may slide or rotate within the gripper during gripping and movement. Therefore, the actual posture of the robot when placing the cartons is an uncertain value. With a four-corner gripper, the angles of the gripper itself are precisely calibrated, allowing the algorithm to be 100% confident that the carton posture used in its calculations is the posture at the time of placement. This eliminates a significant uncertainty, greatly improving the accuracy of subsequent placement position predictions and stability calculations. When the algorithm detects a rotational offset in the lower layer of cartons on the stack, it attempts to compensate by adjusting the angle of the upper layer of cartons. With a two-plate gripper, this strategy is difficult to execute due to the inability to precisely control the carton angles. The four-corner gripper makes precise posture control possible. The algorithm can work in this way.

[0048] In summary, this clamping method not only provides a stable clamping effect but also a more stable algorithm implementation environment. Since no new variables (position and angular offset) are introduced during the clamping process itself, the control model of the entire robot system is simplified. The algorithm can focus more on handling visual detection errors and the robot's absolute positioning accuracy, without needing to build a complex compensation model for an uncertain "clamping slippage," thus reducing the difficulty of system debugging and maintenance.

[0049] Based on this, the following design was made for the structure of the clamping plate in this embodiment:

[0050] First refer to Figure 1-2 As shown, the clamping surface of the left clamping plate 3 is inclined, specifically an upward-flaring inclined structure, while the clamping surface of the right clamping plate 2 is flat. The purpose of this design is to provide freedom of adjustment for the clamping plates and the carton's posture while ensuring stable clamping.

[0051] Reference Figure 4 As shown, the right clamp 2 is designed as a rotatable and fine-tunable structure. As shown in the figure, the right clamp 2 and the sliding plate 6 are rotatably connected by a horizontal rotating shaft 21 and driven by a motor 22. The length direction of the rotating shaft 21 is the extension direction of the sliding plate 6.

[0052] Reference Figure 1 As shown. When the robot needs to lower the cardboard box, motor 22 drives the right clamping plate 2 to rotate at a certain angle, while linear cylinder 8 makes a slight advance to ensure stable clamping, causing the cardboard box to deflect from a vertical position. Figure 1 The angle shown allows the right side of the carton to rest against the outer side wall of another carton adjacent to it on the pallet, and then move vertically downwards until the bottom left side is about to contact the lower support surface. At this point, the left and right clamps move away from each other, releasing the carton.

[0053] This design avoids the clamping plate being positioned between two cartons when placing them down, preventing large gaps between them. Furthermore, the horizontal displacement of the clamping plate when releasing a carton can push adjacent cartons, causing them to shift. When placing the cartons, they are positioned at a certain angle to each other, providing space for the right-side clamping plate to move aside. This arrangement also ensures the cartons are close together, resulting in better overall stability.

[0054] The specific design of the carton's flipping angle can be determined based on the length of the clamping plate. The purpose is to make the two cartons close together and leave space for the clamping plate to avoid each other.

[0055] In another preferred embodiment, the clamp can be further designed to be retractable and driven by a gear rack or cylinder (this is also quite common in existing clamping flat-plate robots, so the specific structure will not be described in detail in this embodiment). When placed, the clamp retracts a certain space, which can reduce the flipping angle of the carton design to a certain extent.

[0056] Based on this, a downward-extending pressure roller with an elastic cantilever is added to one side of the right clamping plate. After descending to a certain height, the pressure roller presses on the upper part of the adjacent cardboard box, which can prevent the adjacent cardboard box from being impacted and displaced when it is fully lowered.

[0057] In another preferred embodiment, the length of the left clamping plate 3 is greater than that of the right clamping plate 2. Several contact sensors are arranged on the bottom inner side of the left clamping plate 3. When the carton is fully lowered, the left clamping plate 3 is moved to contact the corners of the carton. If each contact sensor receives feedback, it indicates that the carton is placed correctly. Otherwise, the carton's posture is inaccurate.

[0058] It is worth mentioning that the above only discloses part of the robot gripping component, which is installed on the robot. Specifically, it can use slide rail 1 as the mounting point and is configured to be rotatably mounted on the robot arm.

[0059] In another preferred embodiment, each clamp (including all the left clamps 3 and the right clamps 2) is designed as a rotatable structure with its rotation axis vertical, so that it can be used to adjust the angle of the carton.

[0060] Therefore, based on the above design, a four-corner clamping method is adopted, which makes it more flexible to use.

[0061] Based on the design of the clamping component described above, this embodiment also proposes a sensing algorithm for palletizing, which specifically includes the following steps:

[0062] Multi-sensor data acquisition and fusion

[0063] Color images and depth point cloud data of the cargo are acquired using an RGB-D camera, and combined with LiDAR scanning data to construct a 3D model of the cargo stack.

[0064] The Kalman filter algorithm is used to perform time-series alignment and fusion of multi-source sensor data, and the three-dimensional center coordinates of each carton are output. and the rotation angle around the Z-axis .

[0065] Stacking posture and cost-effectiveness analysis

[0066] Single-box attitude parameter extraction:

[0067] For the i-th carton, its ideal position is defined as ( The actual location is ( ).

[0068] The ideal position is the ideal center coordinates and rotation angle of the i-th carton; the latter is its actual center coordinates and rotation angle.

[0069] Calculate the position and angle deviations of the X and Y coordinates:

[0070]

[0071] Calculation of effective position deviation:

[0072] Considering the actual space occupied after rotation, the maximum positional deviation of the four corner points is calculated as the effective deviation:

[0073]

[0074] For effective positional deviation, Let K be the actual coordinates of the k-th corner point. Let be the ideal coordinates of the k-th corner point.

[0075] The derivative of corner coordinates through rotation matrix transformation:

[0076]

[0077] This represents the rotation angle of the cardboard box.

[0078] Interlayer alignment and interlayer support analysis

[0079] Calculate the standard deviation of the center coordinates of all cartons in the same layer:

[0080]

[0081] in, ;

[0082] n is the number of cartons in this layer;

[0083] This is the average of the center coordinates of all the cartons in this layer;

[0084] Let the coordinates be the center coordinates of the i-th carton;

[0085] Inter-layer bracing ratio assessment

[0086] Project the upper cardboard box onto the lower plane and calculate the overlap ratio between the projected area and the lower support area:

[0087]

[0088] For support rate;

[0089] Proj(U) is the projection of the upper carton onto the lower plane;

[0090] L represents the lower support area;

[0091] Area(...) is the area calculation function;

[0092] Central stability margin

[0093] Calculate the shortest distance from the overall centroid to each side of the supporting polygon:

[0094] ;

[0095] For stability margin;

[0096] C represents the coordinate of the overall center of gravity of the cargo stack;

[0097] To support the i-th edge of the polygon;

[0098] d(C, ) is the vertical distance from the centroid to the i-th edge.

[0099] Stability overall score and risk classification:

[0100] Construct a weighted scoring model

[0101]

[0102] S represents the overall stability score (0-1).

[0103] These are weighting coefficients, and their sum is 1. They can be set according to the specific product and on-site conditions.

[0104] Risk levels are determined based on the scoring results:

[0105] Low risk (S > 0.8S > 0.8): Place according to the original plan;

[0106] Medium risk (0.5 < S ≤ 0.80.5 < S ≤ 0.8): Fine-tune the placement position;

[0107] High risk (S ≤ 0.5S ≤ 0.5): Re-plan the stacking strategy or give an alarm

[0108] Finally, based on the stability score, a set of candidate placement positions is generated. The stability of each candidate position is predicted through simulated stacking, and the optimal position is selected. The robot is controlled to adjust the position and posture of the end effector to achieve precise placement of the carton.

[0109] In addition, feedback learning and optimization can be introduced. Record the actual stability data of each palletizing operation, and update the scoring model parameters through machine learning algorithms to continuously improve the prediction accuracy.

[0110] This embodiment also proposes a system, including a perception module: containing sensors such as RGB-D cameras and lidar, used to collect three-dimensional data of the cargo stack.

[0111] Data processing module: used for multi-source data fusion, attitude parameter extraction, and offset calculation.

[0112] Stability evaluation module: used for in-layer alignment analysis, inter-layer support rate calculation, and center-of-gravity stability evaluation.

[0113] Decision control module: used to generate an adaptive placement strategy and control the robot to execute.

[0114] Learning and optimization module: used to optimize the stability prediction model based on historical data.

[0115] Through multi-sensor fusion and three-dimensional modeling, the position and posture of the carton are accurately perceived in real time, overcoming the limitations of traditional two-dimensional vision. A stability scoring model based on effective position deviation, support rate, and center-of-gravity margin is proposed to quantitatively evaluate the tipping risk. Adaptive palletizing control based on real-time stability prediction is achieved, significantly improving the safety of the cargo stack and the intelligent level of robot operations. It has the ability of online learning and can adapt to different carton specifications and stacking scenarios by continuously optimizing the model parameters.

[0116] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A palletizing robot based on posture adjustment function, comprising a robot arm and a gripping assembly mounted at the end of the robot arm, characterized in that: The clamping assembly includes a slide rail (1), two mounting plates (9) symmetrically arranged on the slide rail (1), and a drive assembly, a left clamping plate (3) and a right clamping plate (2) respectively mounted on each mounting plate (9); The drive assembly includes a linear cylinder (8), the output end of which is connected to a horizontal shaft (4); Each mounting plate (9) is provided with two guide rails (7) arranged in a figure-eight shape, and each guide rail (7) is slidably connected with a sliding plate (6); A sliding bearing (5) is fixed on the sliding plate (6), and the horizontal shaft (4) slides through the two sliding bearings (5). The left clamping plate (3) and the right clamping plate (2) are respectively installed on two sliding plates (6), and move synchronously inward or outward by the push of the linear cylinder (8) to form a four-corner clamping structure; The clamping surface of the left clamping plate (3) is an upwardly flared inclined structure, and the clamping surface of the right clamping plate (2) is a flat structure. The right clamp (2) is rotatably connected to the sliding plate (6) via a horizontal rotating shaft (21), and is driven by a motor (22) to achieve fine-tuning of rotation around the rotating shaft (21); The right clamp (2) has a downwardly extending elastic pressure roller on one side, which is used to press the adjacent cartons together when placing them to prevent them from shifting. The left clamp (3) is longer than the right clamp (2) and has multiple contact sensors on its bottom inner side for detecting whether the carton is placed in place.

2. A palletizing robot based on posture adjustment function according to claim 1, characterized in that: The clamping assembly is rotatably mounted on the robot arm via a slide rail (1).

3. A sensing and control system for the palletizing robot based on posture adjustment function as described in claim 1, characterized in that, include: The perception module, including an RGB-D camera and a LiDAR, is used to acquire 3D images and point cloud data of the cargo stack; The data processing module is used for multi-source data fusion, carton posture parameter extraction, and offset calculation. The stability assessment module is used to calculate intra-layer alignment, inter-layer support ratio, and center of gravity stability margin. The decision control module is used to generate placement strategies based on stability scores and control the robot's execution. The learning optimization module is used to optimize the stability prediction model based on historical data.

4. The sensing and control system according to claim 3, characterized in that: The stability assessment module calculates the overall stability score using the following formula: , Where S is the overall stability score. For effective positional deviation, For stability margin, These are the weighting coefficients.

Images