Automatic loading and unloading system for mine car transportation box warehouse and control method of automatic loading and unloading system

Abstract

The invention provides an automatic loading and unloading system for a mine car transportation box warehouse and a control method of the automatic loading and unloading system, belongs to the technical field of logistics automation, and aims to solve the problems of low efficiency and poor safety of traditional manual or semi-manual loading and unloading. The system comprises six modules including a lifting appliance, a crown block, a mechanical arm, a box body, a flat car and a sensor. The box body adopts a standardized design, and is quickly locked and separated from a lifting appliance and a flat car through upper and lower corner fittings; the flat car is provided with a lifting twist lock mechanism and is compatible with five types of box structures; the four groups of mechanical arms fuse UWB, IMU and multi-depth camera sensing data to cooperatively complete twist lock picking, dynamic posture adjustment and automatic locking operation; the crown block module is integrated with a visual monitoring and early warning system, and intelligent closed-loop control is constructed. Through multi-sensor fusion positioning and mechanical arm cooperative control, full-automatic precise loading and unloading of the transportation box body are achieved, efficiency and safety are remarkably improved, and the intelligent logistics management system is suitable for intelligent logistics management in complex scenes such as mines and warehouses.

Description

Technical Field

[0001] The present invention relates to the field of automated transportation and loading and unloading of mine materials, and in particular to an automated loading and unloading system for a mine car transport box warehouse and a control method thereof. Background Art

[0002] At present, material transportation in mines and warehouses generally adopts manual or semi-manual loading and unloading methods, which have prominent problems such as low operating efficiency, high human resource consumption, and insufficient inherent safety. In particular, for materials of different forms such as minerals, various materials, and equipment accessories, traditional mine cars are difficult to adapt to the differentiated loading requirements of multiple categories of materials due to structural limitations, resulting in low space utilization. Its scheduling system lacks an intelligent coordination mechanism, and it is easy to waste transportation capacity when multiple vehicles work together. At the same time, complex geological conditions and harsh environment also bring great challenges to workers' operations. Material transportation has never been able to effectively form an overall fully automated operation process, and it is difficult to achieve automated loading and unloading relying on the existing loading system. Summary of the invention

[0003] The present invention proposes an automatic loading and unloading system for a mine car transport box warehouse and a control method thereof. Through highly modular hardware design, multi-sensor perception and recognition, and intelligent architecture of mechanical arm collaborative control, the full-process automatic loading and unloading of multi-configuration transport boxes in mining scenarios is realized. The system takes multi-sensor deep fusion as the core, combines the dynamic posture adjustment of the mechanical arm and the operation technology of the locking device, and effectively solves the problems of high manual dependence, insufficient precision, and poor adaptability in traditional loading and unloading operations.

[0004] The system consists of six major parts: a hoist module, a crane module, a robotic arm module, a box module, a flatbed car module, and a sensor module. Each module is integrated into the entire system to work together. The box module is designed to be a coverless type, a double-door type, a side-unloading type, a column type, a canned type, and other configurations according to the physical properties of the material and the requirements of loading and unloading. All configurations follow the unified external dimension specifications. High-strength steel corner pieces are welded at the top and bottom corners, which are fixedly matched with the locking mechanisms of the hoist module and the flatbed car module. The flatbed car module is provided with a wheel pair that is compatible with the track, and the four corners of the frame are equipped with a lifting and twisting lock device. After the box is placed on the flatbed car, the lock head can rise vertically into the corner piece and then rotate the lock head to complete the locking. The lock head can be unlocked by rotating the lock head in the opposite direction, and the lock head can be lowered below the surface of the flatbed car. The spreader module is connected to the crane hook through four sets of high-strength ropes. The four corners of the spreader frame are equipped with locking devices, and a rotating shaft is provided in the middle, which is connected by a connecting rod. The locking device is a rotatable lock head, and the whole adopts a two-crank slider mechanism in series. When the crane hook drives the middle rotating shaft to rotate, the rotating shaft transmits power to the four sets of lock heads, and the lock heads rotate synchronously to clamp the angle pieces to complete the locking and fixing, and rotate counterclockwise to achieve unlocking and separation. Positioning baffles are also provided on the outside of the four corners to improve the matching accuracy of the spreader module and the box module.

[0005] As the perception and recognition module of the system, the sensor module integrates UWB (ultra-wideband) positioning unit, IMU (inertial measurement unit) and multiple groups of depth cameras to form a multi-modal perception network covering the entire loading and unloading scene. The UWB base station uses the time difference ranging method to solve the three-dimensional spatial coordinates of the spreader module and the box module in real time, with high positioning accuracy; the IMU has a built-in three-axis gyroscope and accelerometer to monitor the changes in the box's attitude angle; the multi-depth camera uses binocular stereo vision and structured light fusion technology, which can still accurately identify the millimeter-level contour features of the box corners, flatbed car twist locks and surrounding obstacles under complex lighting conditions, and provide path points for the robot arm module to dynamically adjust the box posture and operate the twist lock. The overhead crane module is equipped with an industrial-grade human-computer interaction terminal, which displays the box position, attitude Euler angle, locking status and robot arm movement trajectory transmitted by the sensor module in real time. The operator can switch to full-automatic mode or manually intervene in fine-tuning through the touch interface.

[0006] The robot module consists of four six-degree-of-freedom collaborative robot arms, which are symmetrically arranged on both sides of the flatbed module, and are equipped with multi-functional grippers at the end. During the loading and unloading process, the robot arm performs multi-task collaboration based on multi-sensor fusion data: when the box is lifted by the overhead crane to the top of the flatbed, the robot arm dynamically adjusts the end clamping force through the impedance control algorithm to offset the posture deviation caused by the shaking of the hoist; during the lowering of the box, the multi-depth camera captures the spatial coordinates of the corner pieces and the lock holes in real time, and guides the robot arm grippers to operate the twist locks to complete the lifting and screwing operations. The system has fault self-diagnosis and redundant fault tolerance capabilities. When a single robot arm crashes due to an unexpected power outage or communication interruption, the remaining robot arms automatically reallocate task weights and maintain the continuity of the loading and unloading process by increasing the clamping force and adjusting the motion trajectory.

[0007] The core of the control method is to build a closed-loop control architecture of "perception-decision-execution-feedback". After the loading and unloading operation starts, the UWB and IMU fuse data in real time to generate an initial pose estimate of the box, and the overhead crane module lifts the box to the target area above the flatbed truck according to the path planning algorithm. At this stage, multiple depth cameras synchronously scan the surface of the flatbed truck, identify the spatial coordinates of the locking device through a feature matching algorithm, and input the data into the robotic arm motion planner. When the box enters the hovering state, the gripper at the end of the robotic arm accurately grasps the bottom of the box and automatically twists the lock based on vision guidance, and applies a fine-tuning torque in the horizontal plane to the box through a force-position hybrid control strategy to eliminate the pose deviation caused by hoisting inertia. After the box is lowered onto the flatbed module, the multi-depth camera quickly identifies the spatial coordinates of the lock holes at the four corners of the flatbed. The multi-robotic arms plan the clamping point path based on visual guidance, and the clamping mechanism locks the box through a compound motion of pushing and rotating. When unloading the box, the reverse operation is performed. The robotic arm module locates the characteristic points of the locking device at the four corners of the carrying platform under the guidance of the multi-depth camera, and the clamping arm unlocks the box through reverse rotation of the lock handle.

[0008] The advantages of the present invention are as follows: Through modular design, the efficiency and safety of material transportation and handling are improved simultaneously. The overall structure adopts a unified specification box body and a universal locking structure for flatbed trucks, which can adapt to the transportation requirements of various configurations of materials, reduce waste of transportation capacity, and improve the handling efficiency. The sensor fusion technology (UWB / IMU / depth vision) constructs a centimeter-level positioning network. Combined with the coordinated posture adjustment and fault-tolerant control of four robotic arms, the lifting sway error is dynamically eliminated, ensuring the precise docking of the box body and the flatbed truck module. The locking device of the flatbed truck cooperates with the vision guidance system to realize the automatic locking and unlocking operations of the robotic arm, avoiding manual intervention in high-risk operations. The overhead crane monitoring and pedestrian warning system form a three-dimensional protection. Through the visual interface, seamless switching between human-machine collaboration and full-automatic mode is achieved. While ensuring operation safety, the system can also continuously optimize the operation process through reinforcement learning, forming an efficient and closed-loop intelligent handling system. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 FIG. 1 is a schematic structural diagram of an automated loading and unloading system for a mine car transportation box body warehouse provided by an embodiment of the present invention;

[0010] Figure 2 is Figure 1 an enlarged view of the partial structure at A in FIG. 1;

[0011] Figure 3 FIG. 2 is a schematic diagram of the locking of the flatbed truck box body of an automated loading and unloading system for a mine car transportation box body warehouse provided by an embodiment of the present invention;

[0012] Figure 4 FIG. 3 is a schematic diagram of the key component structure of the flatbed truck module of an automated loading and unloading system for a mine car transportation box body warehouse provided by an embodiment of the present invention;

[0013] Figure 5 FIG. 4 is a schematic diagram of the locking of the spreader box body of an automated loading and unloading system for a mine car transportation box body warehouse provided by an embodiment of the present invention;

[0014] Figure 6 FIG. 5 is a schematic diagram of the structure of the robotic arm module of an automated loading and unloading system for a mine car transportation box body warehouse provided by an embodiment of the present invention;

[0015] Figure 7 FIG. 6 is a schematic diagram of the control framework of an automated loading and unloading system for a mine car transportation box body warehouse provided by an embodiment of the present invention.

[0016] DESCRIPTION OF THE REFERENCE NUMERALS:

[0017] Spreader module (1), overhead crane module (2), robotic arm module (3), box body module (4), flatbed truck module (5), sensor module (6), corner fitting (7), twist lock head (8), wheel set (9), twist lock (10), rotary lock head (11), positioning baffle (12). Detailed Implementation Modes

[0018] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following will describe in detail the specific embodiments of the present invention with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0019] An automated loading and unloading system for a mine car transportation box warehouse, characterized in that it includes a spreader module (1), a crane module (2), a robotic arm module (3), a box module (4), a flatbed truck module (5), and a sensor module (6), as Figure 1 shown, wherein, the flatbed truck module (5) is provided with a locking device, which can fix and carry the box module (4), the materials are loaded in the box module (4), and are transported through the flatbed truck module (5); the spreader module (1) is hung under the crane module (2), and the spreader module (1) can lift the box module (4), and the loading and unloading of the box module (4) is realized by the movement of the crane module (2); the box module (4) has a variety of configurations and can load different types of materials; the robotic arm module (3) is arranged around the flatbed truck module (5) and realizes a series of automated operations during the loading and unloading of the box module (4), including picking up automatic twist locks, righting the box module (4), locking and unlocking; the sensor module (6) has sensing and recognition functions, can measure the pose of the box module (4), identify the characteristics of the box module (4), and provide the motion path points of the robotic arm module (3).

[0020] The box module (4) is designed with a variety of configurations according to the types of materials, specifically divided into lidless type, double-door type, side-dumping type, column type, and canned type. Each configuration has the same external dimensions and can be locked and separated from the flatbed truck module (5) and the spreader module (1) respectively. Specifically, corner fittings (7) are provided at the four corners of the top and bottom of the box. The flatbed truck module (5) and the spreader module (1) are respectively provided with twist lock heads (8) and rotary lock heads (11) at corresponding positions, as Figure 3 shown, the two types of lock heads can rotate within the corner fittings (6) to realize the locking or detachment of the two modules; the flatbed truck module (5) can run on the track and can fix and transport the box module (4). Specifically, the bottom of the flatbed truck is provided with a wheel set (9) structure, and twist locks (10) are provided on the vehicle frame according to the positions of the corner fittings (7) at the bottom of the box module (4), as Figure 4 shown, which can realize the locking of each configuration of the box module (4). The twist lock head (8) can rise and rotate within the corner fitting (7) to clamp, and the twist lock head (8) can rotate in the reverse direction and fall below the upper surface of the flatbed truck module (5), so that the flatbed truck module (5) can be used as a general flatbed truck.

[0021] The sling module (1) can be mounted on the overhead crane module (2) and can realize the locking and separation functions with the box module (4). Specifically, it is provided with ear plate structures at the four corners and is towed to the hook of the overhead crane module (2) by four ropes. Rotating lock heads (11) are provided at the four corners of the frame according to the positions of the corner fittings at the top of the box module (4). As Figure 5 shown, the rotating lock head (11) is connected to the intermediate rotating shaft through a connecting rod. The hook drives the intermediate rotating shaft to rotate, driving the rotating lock head (11) to rotate synchronously to achieve locking or detachment. During the movement of the sling, there may be shaking, and positioning baffles (12) are also designed at the four corner positions. The overhead crane module (2) includes an overhead crane, a host computer, and a human-computer interaction monitoring software. The overhead crane can mount and move the sling module (1). A monitoring camera is provided at the top of the overhead crane. The monitoring interface can display the monitoring screen, the pose data of the box, and the current loading and unloading process status in real time, and issue an alarm to the personnel entering the working area of the overhead crane. The overhead crane module (2) can be automatically controlled or operated by personnel.

[0022] The sensor module (6) includes UWB, IMU, and multi-depth cameras. UWB and IMU are respectively used to obtain the three-dimensional poses of the sling module (1) and the box module (4). The multi-depth cameras are used to identify and sense the features of the box module (4) and the flatbed truck module (5) to obtain the pose of the feature points. Specifically, based on the pose of the box module (4) obtained by UWB and IMU, the robotic arm module (3) makes dynamic adjustments. The multi-depth cameras identify the features of the box and the flatbed truck to obtain their depth information, providing a reference basis for the motion control of the robotic arm module (3), and ensuring the real-time, rapid, and secure transmission of the measured data. The robotic arm module (3) includes four groups of robotic arms. As Figure 6 shown, it can achieve precise loading and unloading of the box module (4) through coordinated control. The robotic arms are arranged outside the locking device of the flatbed truck module (5). Based on the perception feedback data of the sensor module (6) as described in claim 1, it can simulate the operation process of the operator. Specifically, the robotic arms can automatically pick up and release the automatic twist locks, correct the pose of the box module (4), lock and unlock the twist locks (10) of the flatbed truck module (5), plan the path in real time, have an obstacle avoidance function, and have a certain fault tolerance ability. When one or several robotic arms break down, the remaining robotic arms that can still work normally can still complete the loading operation without affecting the normal work process. By storing and querying historical data, a reinforcement learning model can also be established, and reinforcement training can be carried out according to the historical data to optimize the motion trajectory of the robotic arms.

[0023] Through the deep integration of the identification and positioning technology of the sensor module (6) as described in claim 4 and the coordinated control technology of the robotic arm module (3), a closed-loop feedback intelligent control architecture is constructed. As Figure 7As shown in the figure, the precise positioning, dynamic pose adjustment of the box module (4) and the automatic locking operation between the flatbed truck module (5) and the box module (4) are realized, specifically as follows: Based on the three-dimensional pose of the box module (4) obtained in real time by the IMU and UWB in the sensor module (6), combined with the visual recognition of the box module (4) by multiple depth cameras, multi-modal perception data is formed and transmitted to the visualization interface in the overhead crane module (2), dynamically displaying the box position, attitude angle and surrounding environment for the operator to control the overhead crane in real time to complete the preliminary lifting and positioning of the box module (4); At the same time, the system calculates the target pose error between the box module (4) and the flatbed truck module (5) in real time through the space coordinate transformation algorithm, and inputs the error vector into the robotic arm module (3) to generate the joint motion trajectory command of the robotic arm, driving the robotic arm module (3) to synchronously execute the box righting, fine-tuning and dynamic balance compensation actions, and continuously correcting the control parameters through the IMU and UWB feedback loops until the error converges to the preset threshold range; During this process, the multiple depth cameras synchronously identify the twist locks (10) of the flatbed truck module (5) and the hole positions of the bottom corner fittings (7) of the box module (4), generate the precise position information of the twist locks (10) through feature matching, and guide the end effector of the robotic arm to plan the grasping path to complete the lifting, unlocking or locking of the twist locks (10); The system supports the switching between the human-machine collaboration and the full-automatic mode. In the manual control mode, the operator manipulates the box to be lifted to the target area and slowly lowers it through the overhead crane console. The multiple robotic arms adjust their postures in real time to offset the lifting disturbance and compensate for the end positioning error through visual servo; In the full-automatic mode, the overhead crane is taken over by the central controller, and based on the path planning algorithm, it autonomously completes the lifting and lowering. At the same time, the robotic arm module (3) can optimize the control strategy through reinforcement learning to achieve the high-precision dynamic alignment of the box and the flatbed truck, and the operator only needs to monitor the anomalies through the safety interface.

[0024] The system constructs a multi-source sensor network through the UWB positioning system, multi-depth camera recognition system, and IMU inertial measurement unit in the sensor module (6), and combines the collaborative control technology of the robotic arm module (3) to achieve fully automatic and precise loading and unloading operations of the transportation box. During loading, first, the UWB and multi-depth cameras are used to monitor the movement trajectory of the transportation box in real time. When it is recognized that the box module (4) hovers above the flatbed truck module (5), the robotic arm module (3) collaborates under the guidance of the sensor module (6) to complete the automatic unlocking operation of the bottom twist locks of the box. Subsequently, the pose deviation between the box and the bearing platform in the horizontal plane is calculated in real time through the attitude solution algorithm that fuses the IMU and multi-depth cameras, generating a collaborative motion trajectory for multiple robotic arms. A multi-point compliant control strategy is adopted to dynamically correct and adjust the pose of the box. After the box is lowered onto the flatbed truck module (5), the multi-depth camera quickly identifies the spatial coordinates of the lock holes at the four corners of the flatbed truck. Multiple robotic arms plan the path of the clamping points based on visual guidance, and the gripper mechanism realizes locking through the combined movement of pushing and rotating. During unloading, the reverse operation is performed. The robotic arm module (3) locates the feature points of the twist locks at the four corners of the flatbed truck module (5) under the guidance of the multi-depth camera, and the gripper realizes reverse rotation unlocking of the lock handle through reverse rotation. After the overhead crane module (2) hoists the box module (4) to the specified height, the robotic arm module (3) precisely installs the automatic twist locks at the bottom of the box under the fusion positioning guidance of the IMU and multi-depth cameras to complete the box unloading operation.

Claims

1. An automated loading and unloading system for a mine car transportation box warehouse, characterized in that, It includes a spreader module (1), a crane module (2), a robotic arm module (3), a box module (4), a flatbed truck module (5) and a sensor module (6). Among them, the flatbed truck module (5) is equipped with a locking device that can fix and carry the box module (4). The material is loaded in the box module (4) and transported by the flatbed truck module (5). The spreader module (1) is suspended below the crane module (2), and the spreader module (1) can lift the box module (4) to realize the loading and unloading of the box module (4) by the movement of the crane module (2). The box module (4) has multiple configurations and can load different types of materials. The robotic arm module (3) is arranged around the flatbed truck module (5) to realize a series of automated operations during the loading and unloading of the box module (4), including picking up automatic twist locks, righting the box module (4), locking and unlocking. The sensor module (6) has sensing and recognition functions, can measure the pose of the box module (4), identify the characteristics of the box module (4), and provide the motion path points of the robotic arm module (3).

2. The automated loading and unloading system for a mine car transportation box warehouse according to claim 1, characterized in that, The box module (4) is designed with multiple configurations according to the types of materials, specifically divided into lidless type, double-door type, side-dumping type, column type, and canned type. Each configuration has the same outer dimensions and can be locked and separated from the flatbed truck module (5) and the spreader module (1) respectively. Specifically, corner fittings (7) are provided at the four corners of the top and bottom of the box. Twist lock heads (8) and rotary lock heads (11) are respectively provided at the corresponding positions of the flatbed truck module (5) and the spreader module (1). The two types of lock heads can rotate within the corner fittings (6) to realize the locking or detachment of the two modules. The flatbed truck module (5) can run on the track and can fix and transport the box module (4). Specifically, the bottom of the flatbed truck is provided with a wheel set (9) structure, and twist locks (10) are provided on the vehicle frame according to the positions of the corner fittings (7) at the bottom of the box module (4), which can realize the locking of each configuration of the box module (4). The twist lock heads (8) can rise and rotate into the corner fittings (7) to clamp, and the reverse rotation of the twist lock heads (8) can fall below the upper surface of the flatbed truck module (5), enabling the flatbed truck module (5) to be used as a general flatbed truck.

3. The automatic loading and unloading system for a mine car transportation box warehouse according to claim 1, wherein The sling module (1) can be mounted on the crane module (2) and can realize the locking and separation functions with the box module (4). Specifically, ear plate structures are provided at the four corners and are pulled to the hook of the crane module (2) by four ropes. Rotary locks (11) are provided at the four corners of the frame according to the position of the top corner pieces of the box module (4). The rotary locks (11) are connected to the intermediate shaft through a connecting rod. The hook drives the intermediate shaft to rotate and drives the rotary locks (11) to rotate synchronously to achieve locking or disengagement. The sling may shake during movement, and positioning baffles (12) are also designed at the four corners. The crane module (2) includes a crane, a host computer and human-computer interaction monitoring software. The crane can mount and move the sling module (1). A monitoring camera is provided on the top of the crane. The monitoring interface can display the monitoring screen, the posture data of the box and the current loading and unloading process status in real time, and send an alarm to personnel entering the crane working area. The crane module (2) can be automatically controlled or operated by personnel.

4. The automatic loading and unloading system for a mine car transportation box warehouse according to claim 1, characterized in that, The sensor module (6) includes UWB, IMU and a multi-depth camera. The UWB and IMU are used to obtain the three-dimensional posture of the sling module (1) and the box module (4) respectively. The multi-depth camera is used to identify and perceive the features of the box module (4) and the flatbed vehicle module (5) to obtain the posture of the feature points. Specifically, the posture of the box module (4) obtained based on the UWB and IMU is dynamically adjusted through the mechanical arm module (3). The multi-depth camera identifies the features of the box and the flatbed vehicle to obtain their depth information, providing a reference basis for the motion control of the mechanical arm module (3) and ensuring the real-time, rapid and safe transmission of the measured data. The mechanical arm module (3) includes four groups of mechanical arms, which can realize the control of the box module (4) through collaborative control. For precise loading and unloading, the robot arm is arranged outside the locking device of the flatbed vehicle module (5). Based on the perception feedback data of the sensor module (6) as claimed in claim 1, it can simulate the operator's loading and unloading action process, which is specifically manifested in that the robot arm can automatically remove the automatic twist lock, correct the posture of the box module (4), lock and unlock the twist lock (10) of the flatbed vehicle module (5), plan the path in real time, have obstacle avoidance function, and have a certain fault tolerance capability. When one or several robot arms fail, the remaining robot arms that can work normally can still complete the loading operation without affecting the normal workflow; by storing and querying historical data, a reinforcement learning model can also be established, and reinforcement training can be performed based on historical data to optimize the movement trajectory of the robot arm.

5. A control method for an automated loading and unloading system of a mine car transportation box warehouse, characterized in that, Through the deep integration of the positioning technology identified by the sensor module (6) as described in claim 4 and the collaborative control technology of the robotic arm module (3), a closed-loop feedback intelligent control architecture is constructed to achieve the precise positioning, dynamic pose adjustment of the box module (4) as described in claim 2, and the automatic locking operation between the flatbed truck module (5) and the box module (4). Specifically, it is manifested as follows: Based on the three-dimensional pose of the box module (4) obtained in real time by the IMU and UWB in the sensor module (6), combined with the visual recognition of the box module (4) by multiple depth cameras, multi-modal perception data is formed and transmitted to the visualization interface in the overhead crane module (2), dynamically displaying the box position, attitude angle and surrounding environment for the operator to manipulate the overhead crane in real time to complete the preliminary lifting and positioning of the box module (4); At the same time, the system calculates the target pose error between the box module (4) and the flatbed truck module (5) in real time through the space coordinate transformation algorithm, and inputs the error vector into the robotic arm module (3) to generate the joint motion trajectory command of the robotic arm, driving the robotic arm module (3) to synchronously execute the box righting, fine-tuning and dynamic balance compensation actions, and continuously correct the control parameters through the IMU and UWB feedback loops until the error converges to the preset threshold range; During this process, the multiple depth cameras synchronously identify the twist lock (10) of the flatbed truck module (5) and the hole positions of the bottom corner fittings (7) of the box module (4), generate the precise position information of the twist lock (10) through feature matching, and guide the end effector of the robotic arm to plan the grasping path to complete the lifting and unlocking or locking of the twist lock (10); The system supports the switching between human-machine collaboration and full-automatic mode. In the manual control mode, the operator manipulates the box to be lifted to the target area and slowly lowers it through the overhead crane console. The multiple robotic arms adjust their postures in real time to offset the lifting disturbance and compensate for the end positioning error through visual servo; In the full-automatic mode, the overhead crane is taken over by the central controller, and based on the path planning algorithm, it autonomously completes the lifting and lowering. At the same time, the robotic arm module (3) can optimize the control strategy through reinforcement learning to achieve the high-precision dynamic alignment of the box and the flatbed truck. The operator only needs to monitor the anomalies through the safety interface.

6. The automatic loading and unloading system for the ore car transportation box warehouse according to claim 1, wherein, The system constructs a multi-source sensor network through the UWB positioning system, multi-depth camera recognition system, and IMU inertial measurement unit in the sensor module (6), and combines the collaborative control technology of the robotic arm module (3) to achieve fully automatic and precise loading and unloading operations of the transportation box. During loading, first, the UWB and multi-depth cameras are used to monitor the movement trajectory of the transportation box in real time. When it is recognized that the box module (4) hovers above the flatbed module (5), the robotic arm module (3) collaboratively completes the automatic unlocking operation of the bottom twist lock of the box under the guidance of the sensor module (6). Subsequently, the attitude solution algorithm that fuses the IMU and multi-depth cameras is used to calculate the horizontal pose deviation between the box and the bearing platform in real time, generate the collaborative motion trajectory of multiple robotic arms, and adopt a multi-point compliant control strategy to dynamically correct and adjust the pose of the box. After the box is lowered onto the flatbed module (5), the multi-depth camera quickly recognizes the spatial coordinates of the lock holes at the four corners of the flatbed. Multiple robotic arms plan the clamping point path based on visual guidance, and the gripper mechanism realizes locking through the combined movement of pushing and rotating. During unloading, the reverse operation is performed. The robotic arm module (3) locates the feature points of the twist locks at the four corners of the flatbed module (5) under the guidance of the multi-depth camera, and the gripper realizes the reverse rotation unlocking of the lock handle through the reverse rotation action. After the overhead crane module (2) hoists the box module (4) to the specified height, the robotic arm module (3) precisely installs the automatic twist lock at the bottom of the box under the guidance of the integrated positioning of the IMU and multi-depth cameras to complete the unloading operation of the box.

Images