Warehouse track car taking and placing mechanism with machine learning goods identification function

Abstract

The utility model belongs to the technical field of intelligent warehousing and logistics, concretely relates to a warehousing rail car taking and placing mechanism with machine learning goods identification function, including goods shelf and the rail car body of setting in goods shelf one side, the rail car body top is provided with height adjusting assembly, is provided with first threaded bush on height adjusting assembly, first threaded bush one side is provided with electric telescopic handle and industry camera body, and electric telescopic handle telescopic end is installed with clamping assembly, is provided with first hollow casing on clamping assembly, is provided with replacement assembly in first hollow casing inside, is provided with clamping plate on replacement assembly, and replacement assembly includes the pivot of setting in first hollow casing inside and four rotating holes of setting on first hollow casing. The device can be based on real-time visual information, accurate calculation goods actual state and surrounding space, and automatically, quickly select optimal solution from a variety of execution scheme, realize safe, reliable, efficient gripping operation finally.

Description

Technical Field

[0001] This utility model belongs to the field of intelligent warehousing and logistics technology, specifically relating to a warehouse railcar picking and placing mechanism with machine learning for goods recognition. Background Technology

[0002] In modern intelligent warehousing and logistics systems, automated storage and retrieval equipment is a core component, among which warehouse railcars (also known as stacker cranes or shuttle cars) are widely used due to their flexibility and efficiency. Traditional warehouse railcar retrieval mechanisms are usually based on preset programmed control, and their actuators are mostly fixed structures or clamping devices with a single drive method. This design can work effectively when dealing with standardized goods that are uniform in size and neatly arranged.

[0003] However, in actual warehousing operations, due to differences in model, batch, or human placement errors, the actual position and posture of goods within the storage location often deviate from the ideal state, resulting in uncertain and non-standardized gaps between the goods and the rack uprights. Traditional fixed grippers, lacking the ability to sense and adapt to actual working conditions, are prone to several challenges in such scenarios: First, if the gripper thickness is greater than the actual gap, it will be unable to insert smoothly, leading to retrieval failure or even damage to equipment or goods; second, if the gripper thickness is much smaller than the gap, the gripping contact area may be too small, resulting in weak gripping and the risk of goods slipping during transfer. These problems severely restrict the improvement of the operational efficiency and reliability of warehousing systems.

[0004] To address these issues, machine vision technology has been gradually introduced into warehousing systems for product identification and coarse positioning. However, most current applications remain focused on determining the presence and type of goods, failing to deeply integrate visual information into the dynamic closed-loop control of subsequent grasping decisions and execution. While some research has attempted to enhance adaptability by employing multi-degree-of-freedom robotic arms or complex libraries of replaceable end effectors, these solutions typically suffer from drawbacks such as complex system structures, high costs, and lengthy tool changeover processes, making it difficult to meet the demands of high-cycle, high-frequency warehousing operations.

[0005] To address this, we propose a warehouse railcar retrieval and placement mechanism with machine learning-based cargo recognition capabilities. This device can accurately calculate the actual state of the cargo and its surrounding space based on real-time visual information, and automatically and quickly select the optimal solution from multiple execution schemes, ultimately achieving safe, reliable, and efficient retrieval operations. Utility Model Content

[0006] The purpose of this invention is to provide a warehouse railcar picking and placing mechanism with machine learning for cargo recognition. This device can accurately calculate the actual state of the cargo and the surrounding space based on real-time visual information, and automatically and quickly select the optimal solution from multiple execution schemes to ultimately achieve safe, reliable and efficient picking operations.

[0007] The specific technical solution adopted by this utility model is as follows:

[0008] A warehouse railcar retrieval and placement mechanism with machine learning-based goods recognition function includes a shelf and a railcar body disposed on one side of the shelf. A height adjustment component is disposed on the top of the railcar body, and a first threaded sleeve is disposed on the height adjustment component. An electric telescopic rod and an industrial camera body are disposed on one side of the first threaded sleeve. A clamping component is installed on the telescopic end of the electric telescopic rod, and a first hollow shell is disposed on the clamping component. A replacement component is disposed inside the first hollow shell, and a clamping plate is disposed on the replacement component.

[0009] The replacement component includes a rotating shaft disposed inside the first hollow shell and four rotating holes disposed on the first hollow shell. Four first sector blocks distributed at different angles are arranged sequentially from left to right on the rotating shaft. An elastic component is disposed inside each rotating hole. An arc plate is disposed at one end of the elastic component and a clamping plate is installed at the other end of the elastic component. The four clamping plates have different thicknesses.

[0010] Furthermore, the height adjustment assembly includes a U-shaped frame disposed on the top of the railcar body, a first stepper motor disposed on the U-shaped frame, a first threaded rod installed at the output end of the first stepper motor, and a first threaded sleeve sleeved on the first threaded rod.

[0011] Furthermore, a first sliding groove is provided on both sides of the U-shaped frame, and a first slider is provided inside the first sliding groove. One side of the first slider is connected to the first threaded sleeve.

[0012] Furthermore, the clamping assembly includes a connecting plate disposed at the telescopic end of the electric telescopic rod. The connecting plate is provided with a second slide groove and a second stepper motor. A rotating rod connected to the output end of the second stepper motor is disposed inside the second slide groove. The rotating rod is provided with opposite threads, and a matching second threaded sleeve is fitted on the opposite threads. One side of the second threaded sleeve is connected to the first hollow shell.

[0013] Furthermore, the elastic component includes a movable rod disposed inside the rotating hole, one end of the movable rod being connected to the arc-shaped plate, the other end of the movable rod being connected to the clamping plate, and a spring being sleeved on the movable rod between the arc-shaped plate and the inner wall of the first hollow shell.

[0014] Furthermore, the four clamping plates are provided with anti-slip textures.

[0015] Furthermore, a third stepper motor is provided on the first hollow housing, and the output end of the third stepper motor is connected to the rotating shaft.

[0016] Furthermore, the arc-shaped plate and the first sector block are provided with smooth surfaces.

[0017] Furthermore, a support plate is provided on one side of the bottom of the first threaded sleeve, and two sets of symmetrical fixing blocks are provided on one side of the support plate. A rotating shaft is provided between the two fixing blocks, and a fixing sleeve and a gear are sleeved on the rotating shaft. A protective plate is provided on one side of the fixing sleeve. A connecting rod is provided on the telescopic end of the electric telescopic rod, and a moving plate is provided on one side of the connecting rod. The moving plate is provided with meshing teeth that mesh with the gear.

[0018] Furthermore, it also includes a PLC controller.

[0019] The technical effects achieved by this utility model are as follows:

[0020] The system WMS (Warehouse Management System) issues an instruction to retrieve or place a specific item at a certain storage location. The railcar moves horizontally to the target shelf row, the height adjustment component moves vertically to adjust the entire retrieval and placement mechanism to the height of the target shelf, and the industrial camera is activated to take a picture of the target storage location. The built-in machine learning vision algorithm starts working, confirming whether the goods in the storage location are the target goods (to avoid mis-picking). Through image analysis, it estimates the length, width, and height of the goods (especially the width). Based on the known standard storage location width and the estimated goods width, the algorithm calculates the remaining space (gap) on both sides of the goods in real time. The control system receives the gap value calculated by the vision system. The system compares this gap value with the thickness of the four clamping plates and selects the clamping plate with a thickness smaller than the gap value and closest to the gap value. According to the decision result, the control system drives the rotating shaft to rotate. The first sector block at a specific angle on the rotating shaft rotates accordingly, pressing the elastic component and arc plate at the corresponding position. The pressed elastic component pushes the clamping plate at that position to extend outward. The electric telescopic rod moves, pushing the entire clamping assembly forward, so that the selected clamping plate is inserted into the gap on both sides of the goods. Then, the clamping assembly itself moves the clamping plates on both sides towards each other, clamping the goods. Finally, the electric telescopic rod retracts, removing the goods from the shelf. Based on real-time visual information, the device can accurately calculate the actual status of the goods and the surrounding space, and automatically and quickly select the optimal solution from multiple execution plans to achieve safe, reliable and efficient grasping operations. Attached Figure Description

[0021] Figure 1This is a schematic diagram of the overall structure of this utility model;

[0022] Figure 2 This is a schematic diagram of the structure of the railcar body of this utility model;

[0023] Figure 3 This is a schematic diagram of the replacement component of this utility model;

[0024] Figure 4 This is a schematic diagram of the structure of the clamping assembly of this utility model;

[0025] Figure 5 This is a structural schematic diagram of the protective plate of this utility model.

[0026] The attached diagram lists the components represented by each number as follows:

[0027] 1. Shelf; 2. Railcar body; 3. First threaded sleeve; 4. Electric telescopic rod; 5. Industrial camera body; 6. First hollow shell; 7. Clamping plate; 8. Rotating shaft; 9. First sector block; 10. Arc plate; 11. U-shaped frame; 12. First stepper motor; 13. First threaded rod; 14. First slide groove; 15. First slider; 16. Connecting plate; 17. Second slide groove; 18. Second stepper motor; 19. Rotating rod; 20. Second threaded sleeve; 21. Moving rod; 22. Spring; 23. Third stepper motor; 24. Support plate; 25. Rotating shaft; 26. Gear; 27. Protective plate; 29. ​​Moving plate; 28. Meshing teeth. Detailed Implementation

[0028] To make the objectives and advantages of this utility model clearer, the following detailed description is provided in conjunction with embodiments. It should be understood that the following text is merely used to describe one or more specific embodiments of this utility model and does not strictly limit the scope of protection specifically claimed by this utility model.

[0029] like Figures 1-5 As shown, a warehouse railcar retrieval and placement mechanism with machine learning goods recognition function includes a shelf 1 and a railcar body 2 set on one side of the shelf 1. A height adjustment component is set on the top of the railcar body 2, and a first threaded sleeve 3 is set on the height adjustment component. An electric telescopic rod 4 and an industrial camera body 5 are set on one side of the first threaded sleeve 3. A clamping component is installed on the telescopic end of the electric telescopic rod 4. A first hollow shell 6 is set on the clamping component. A replacement component is set inside the first hollow shell 6. A clamping plate 7 is set on the replacement component.

[0030] The replacement component includes a rotating shaft 8 disposed inside the first hollow housing 6 and four rotating holes disposed on the first hollow housing 6. The rotating shaft 8 is provided with four first sector blocks 9 distributed at different angles from left to right. Each rotating hole is provided with an elastic component. One end of the elastic component is provided with an arc plate 10, and the other end of the elastic component is provided with a clamping plate 7. The four clamping plates 7 have different thicknesses.

[0031] The height adjustment component includes a U-shaped frame 11 set on the top of the railcar body 2, a first stepper motor 12 set on the U-shaped frame 11, a first threaded rod 13 installed at the output end of the first stepper motor 12, and a first threaded sleeve 3 fitted on the first threaded rod 13.

[0032] The first stepper motor 12 drives the first threaded rod 13 to rotate, and the first threaded rod 13 drives the first threaded sleeve 3 to rise or fall, thereby achieving height adjustment.

[0033] Meanwhile, the U-shaped frame 11 has a first slide groove 14 on both sides, and a first slider 15 is provided inside the first slide groove 14. One side of the first slider 15 is connected to the first threaded sleeve 3. This arrangement makes the first threaded sleeve 3 move more smoothly.

[0034] The clamping assembly includes a connecting plate 16 disposed at the telescopic end of the electric telescopic rod 4. The connecting plate 16 is provided with a second slide groove 17 and a second stepper motor 18. The second slide groove 17 is provided with a rotating rod 19 connected to the output end of the second stepper motor 18. The rotating rod 19 is provided with opposite threads. A matching second threaded sleeve 20 is fitted on the opposite threads. One side of the second threaded sleeve 20 is connected to the first hollow housing 6.

[0035] The second stepper motor 18 drives the rotating rod 19 to rotate, and the rotating rod 19 drives the two second threaded sleeves 20 to move relative to each other or in opposite directions, thereby causing the first hollow shell 6 to drive the clamping plate 7 to move relative to each other or in opposite directions.

[0036] The elastic component includes a movable rod 21 disposed inside the rotating hole. One end of the movable rod 21 is connected to the arc plate 10, and the other end of the movable rod 21 is connected to the clamping plate 7. A spring 22 is sleeved on the movable rod 21 between the arc plate 10 and the inner wall of the first hollow housing 6.

[0037] When the arc plate 10 moves, it drives the moving rod 21 to move inside the rotating hole and squeezes the spring 22, thereby causing the clamping plate 7 to extend and work.

[0038] The thickness of the four clamping plates 7 has been carefully calculated and designed, ranging from 10mm to 40mm. Furthermore, the four clamping plates 7 are equipped with anti-slip textures to increase friction and improve the clamping effect.

[0039] A third stepper motor 23 is provided on the first hollow housing 6. The output end of the third stepper motor 23 is connected to the rotating shaft 8. The rotating shaft 8 is driven to rotate by the third stepper motor 23, thereby rotating the extension of different clamping plates 7.

[0040] The four first sector blocks 9 can form a circle, that is, they are distributed at 90°. When different thicknesses of clamping plates 7 are needed, the rotation of the first sector blocks 9 by the pivot 8 can be selected.

[0041] The arc plate 10 and the first sector block 9 are provided with smooth surfaces, which can reduce friction and make them easy to squeeze and push.

[0042] A support plate 24 is provided on one side of the bottom of the first threaded sleeve 3. Two sets of symmetrical fixing blocks are provided on one side of the support plate 24. A rotating shaft 25 is provided between the two fixing blocks. A fixing sleeve and a gear 26 are fitted on the rotating shaft 25. A protective plate 27 is provided on one side of the fixing sleeve. A connecting rod is provided on the telescopic end of the electric telescopic rod 4. A moving plate 29 is provided on one side of the connecting rod. A meshing tooth 28 that meshes with the gear 26 is provided on the moving plate 29.

[0043] When the electric telescopic rod 4 is retracted, the electric telescopic rod 4 drives the connecting rod to move the moving plate 29. When the moving plate 29 moves, it causes the gear 26 to rotate through the meshing teeth 28. The gear 26 drives the rotating shaft 25 to rotate. The rotating shaft 25 drives the fixed sleeve to protect the bottom of the goods with the protective plate 27 to prevent the goods from falling off.

[0044] Side panels (not shown in the picture) are provided on both sides of the placement board to prevent goods from falling off and further improve safety.

[0045] It also includes a PLC controller, which can control the operation of electrical components such as the first stepper motor 12, the second stepper motor 18, the third stepper motor 23, and the electric telescopic rod 4. Its operating principle and circuit connection method are existing technologies, and will not be elaborated on here.

[0046] It should be noted that the identification of goods by the industrial camera body 5 (MV-CE050-10GM) and the subsequent analysis and learning by the controller is existing technology and will not be elaborated on here.

[0047] Furthermore, the machine learning vision system described in this invention employs a deep learning-based object detection algorithm (such as YOLO or SSD) to accurately identify cargo boundaries and storage space gaps through real-time processing of cargo images. It also achieves closed-loop control with the actuators via a PLC controller. The system boasts an accuracy rate exceeding 99% and a response time of less than 200ms, fully meeting the demands of high-frequency warehousing operations.

[0048] The working principle of this utility model is as follows: the system WMS (Warehouse Management System) issues an instruction to pick up and place specific goods in a certain storage location. The railcar moves horizontally to the first column of the target shelf, the height adjustment component moves vertically, and the entire picking and placing mechanism is adjusted to the height of the target shelf. The industrial camera is activated to take pictures of the target storage location. The built-in machine learning vision algorithm starts working, confirming whether the goods in the storage location are the target goods (to avoid mis-picking). Through image analysis, it estimates the length, width, and height of the goods (especially the width). Based on the known standard storage location width and the estimated goods width, the algorithm calculates the remaining space (gap) on both sides of the goods in real time. The control system receives the gap value calculated by the vision system. The system compares this gap value with the thickness of the four clamping plates 7, and selects the clamping plate 7 whose thickness is less than the gap value and is closest to the gap value. According to the decision result, the control system drives the rotating shaft 8 to rotate. The first sector block 9 at a specific angle on the rotating shaft 8 rotates accordingly, pressing the elastic component and arc plate 10 at the corresponding position. The pressed elastic component pushes the clamping plate 7 at that position to extend outward. The electric telescopic rod 4 moves, pushing the entire clamping assembly forward, so that the selected clamping plate 7 is inserted into the gap on both sides of the goods. Then, the clamping assembly itself moves the clamping plates 7 on both sides towards each other, clamping the goods. Finally, the electric telescopic rod 4 retracts, removing the goods from the shelf 1. Based on real-time visual information, the device can accurately calculate the actual status of the goods and the surrounding space, and automatically and quickly select the optimal solution from multiple execution plans to achieve safe, reliable and efficient grasping operations.

[0049] The above description is merely a preferred embodiment of this utility model. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of this utility model, and these improvements and modifications should also be considered within the scope of protection of this utility model. Structures, devices, and operating methods not specifically described or explained in this utility model, unless otherwise specified or limited, shall be implemented using conventional methods in the field.

Claims

1. A warehouse railcar retrieval and placement mechanism with machine learning-based goods recognition function, comprising a shelf (1) and a railcar body (2) disposed on one side of the shelf (1), characterized in that: The top of the railcar body (2) is provided with a height adjustment component, and a first threaded sleeve (3) is provided on the height adjustment component. An electric telescopic rod (4) and an industrial camera body (5) are provided on one side of the first threaded sleeve (3). A clamping component is installed on the telescopic end of the electric telescopic rod (4). A first hollow shell (6) is provided on the clamping component. A replacement component is provided inside the first hollow shell (6). A clamping plate (7) is provided on the replacement component. The replacement component includes a rotating shaft (8) disposed inside the first hollow shell (6) and four rotating holes disposed on the first hollow shell (6). The rotating shaft (8) has four first sector blocks (9) arranged at different angles from left to right. Each rotating hole is provided with an elastic component. One end of the elastic component is provided with an arc plate (10), and the other end of the elastic component is provided with a clamping plate (7). The four clamping plates (7) have different thicknesses.

2. The warehouse railcar retrieval and placement mechanism with machine learning goods recognition function according to claim 1, characterized in that: The height adjustment assembly includes a U-shaped frame (11) disposed on the top of the railcar body (2), a first stepper motor (12) disposed on the U-shaped frame (11), a first threaded rod (13) installed at the output end of the first stepper motor (12), and a first threaded sleeve (3) sleeved on the first threaded rod (13).

3. A warehouse railcar retrieval and placement mechanism with machine learning goods recognition function according to claim 2, characterized in that: The U-shaped frame (11) has a first sliding groove (14) on both sides, and a first slider (15) is provided inside the first sliding groove (14). One side of the first slider (15) is connected to the first threaded sleeve (3).

4. The warehouse railcar retrieval and placement mechanism with machine learning goods recognition function according to claim 1, characterized in that: The clamping assembly includes a connecting plate (16) disposed at the telescopic end of the electric telescopic rod (4). The connecting plate (16) is provided with a second slide groove (17) and a second stepper motor (18). The second slide groove (17) is provided with a rotating rod (19) connected to the output end of the second stepper motor (18). The rotating rod (19) is provided with opposite threads. A matching second threaded sleeve (20) is fitted on the opposite threads. One side of the second threaded sleeve (20) is connected to the first hollow shell (6).

5. A warehouse railcar retrieval and placement mechanism with machine learning goods recognition function according to claim 1, characterized in that: The elastic component includes a movable rod (21) disposed inside the rotating hole. One end of the movable rod (21) is connected to the arc plate (10), and the other end of the movable rod (21) is connected to the clamping plate (7). A spring (22) is sleeved on the movable rod (21) between the arc plate (10) and the inner wall of the first hollow shell (6).

6. A warehouse railcar retrieval and placement mechanism with machine learning goods recognition function according to claim 1, characterized in that: The four clamping plates (7) are provided with anti-slip textures.

7. A warehouse railcar retrieval and placement mechanism with machine learning goods recognition function according to claim 1, characterized in that: A third stepper motor (23) is provided on the first hollow housing (6), and the output end of the third stepper motor (23) is connected to the rotating shaft (8).

8. A warehouse railcar retrieval and placement mechanism with machine learning goods recognition function according to claim 1, characterized in that: The arc-shaped plate (10) and the first sector block (9) are provided with smooth surfaces.

9. A warehouse railcar retrieval and placement mechanism with machine learning goods recognition function according to claim 1, characterized in that: A support plate (24) is provided on one side of the bottom of the first threaded sleeve (3). Two sets of symmetrical fixing blocks are provided on one side of the support plate (24). A rotating shaft (25) is provided between the two fixing blocks. A fixing sleeve and a gear (26) are fitted on the rotating shaft (25). A protective plate (27) is provided on one side of the fixing sleeve. A connecting rod is provided on the telescopic end of the electric telescopic rod (4). A moving plate (29) is provided on one side of the connecting rod. A meshing tooth (28) that meshes with the gear (26) is provided on the moving plate (29).

10. A warehouse railcar retrieval and placement mechanism with machine learning goods recognition function according to claim 1, characterized in that: It also includes a PLC controller.

Images