Automatic loading method for van and related products

Abstract

The invention belongs to the technical field of automatic transportation, and particularly relates to an automatic loading method for a van, which comprises the following steps: S1, an AGV (Automatic Guided Vehicle) receives a system instruction, and analyzes and confirms an operation space coordinate of a target loading operation point; s2, the AGV carries out path planning based on the obtained operation space coordinates, and then autonomously moves to a preset position of the tail of the van; s3, after the AGV arrives at the preset position, the pose of the AGV is adjusted according to camera image recognition analysis, a camera is used for collecting a compartment image, and the compartment image is converted into compartment space coordinates; s4, carrying out combined measurement and calculation on the compartment space coordinates and cargo stacking requirements, generating an optimal stack type layout according to a combined measurement and calculation result, and carrying out operation by the AGV according to the stack type layout; and S5, after the operation is finished, the AGV autonomously navigates and returns to the preset base. According to the method, the compartment is scanned in advance, the optimal stack type layout is generated after combined measurement and calculation, the space of the compartment can be utilized to the maximum degree, and the method is suitable for fragile goods, electronic products and other special goods.

Description

Technical Field

[0001] This invention belongs to the technical field of automated transportation, specifically relating to an automated loading method for vans and related products. Background Technology

[0002] An Automated Guided Vehicle (AGV) is a transport vehicle equipped with electromagnetic or optical automatic guidance devices. It can be controlled by a controller or its movement and route can be set by using electromagnetic tracks, enabling it to move along a prescribed guidance path. It has safety protection and various mobile transport functions.

[0003] Chinese invention patent CN120562480A discloses an automatic loading method for trucks based on deep learning and stereo vision, including the following steps: S1, acquiring the pixel coordinates of each cargo location and transforming them to the cargo location coordinates in the AGV forklift coordinate system through spatial transformation; S2, based on the transformed cargo location coordinates, calculating the driving trajectory of the AGV forklift in the truck area and generating a forward station, so that the AGV forklift moves to the forward station along the driving path; S3, collecting point cloud data when moving in the truck area, denoising the point cloud data, using an improved Transformer-based point cloud registration network for feature extraction and matching, calculating the three-dimensional feature points of the cargo location and pallet, and transforming them to the AGV forklift coordinate system; S4, combining the transformed three-dimensional feature point information, calculating the AGV forklift's driving trajectory in the truck area. The process involves adjusting rotation and displacement compensation parameters to guide the AGV forklift's trajectory and dock it with the target storage location. A global attention mechanism is used to optimize point cloud feature calculation results and correct docking deviations. S5: An improved adaptive local attention weighting strategy is employed in the point cloud registration network to perform fine-grained matching between the target storage location boundary and the pallet slot, correcting the pose in real time to ensure the AGV forklift forks align with the pallet slot and complete cargo grabbing. S6: Combining cargo grabbing information from the point cloud registration network, the AGV forklift's placement strategy is adjusted to optimize cargo stacking posture. The loading angle is calculated based on the center of gravity distribution to control the AGV forklift to complete loading. S7: After loading, the AGV forklift is controlled to exit the loading area along the adjusted trajectory, and loading process data is recorded to continuously optimize the loading strategy and complete automated loading. This solution has the following drawbacks: 1. Lack of pre-treatment and testing of the cargo compartment environment can easily lead to loading risks due to abnormalities in the cargo compartment. Disadvantages: The basic environmental conditions such as the flatness of the truck bed, obstacles, and step height are not tested before directly entering the docking and loading stage. If there are protruding obstacles in the truck bed, uneven ground, or step height that exceeds the adaptability range of the AGV, it may lead to collision damage to the goods and failure of fork / robotic arm docking.

[0004] Causes: The core modules of existing technologies (point cloud processing, pose optimization, etc.) only focus on local coordinate matching and grasping control of "cargo location-pallet-AGV", relying on stereo vision cameras to collect the pixel coordinates of cargo locations and ToF cameras to collect point clouds of truck areas (only used for cargo location / pallet feature extraction). There is no design for scanning and analysis of the overall environment of the truck, and there is a lack of a prediction mechanism for "loading preconditions".

[0005] 2. Insufficient palletizing flexibility, unable to adapt to scenarios involving mixed palletizing of multiple box sizes. Disadvantages: It can only adjust the stacking posture of single or similar goods through point cloud feedback (such as to avoid center of gravity shift), does not support mixed stacking of boxes of different sizes and shapes, and has no clear stacking mode, making it difficult to maximize the use of the carriage space.

[0006] Cause: Existing technologies do not incorporate the logic of "box size input - stack type intelligent calculation", and only rely on torque optimization when the forks are gripping and passive adjustment of stacking posture. They do not consider the needs of mixed loading of multi-specification goods in actual scenarios and lack stacking path planning algorithms for different boxes.

[0007] 3. Weak anomaly handling capabilities, lacking real-time intervention and automatic recovery mechanisms. Disadvantages: It only prevents fork tilting and cargo slippage through torque optimization. If abnormalities such as cargo tilting or material falling have already occurred, it lacks real-time detection and automatic processing capabilities, requiring manual intervention to interrupt operations, resulting in reduced efficiency.

[0008] Cause: The existing "data feedback" module does not have sensors (such as visual cameras + laser rangefinders) to monitor the status of goods in real time, nor does it have a closed-loop control logic of "anomaly detection - stop operation - re-grab". Anomaly handling relies on post-event data rather than real-time intervention.

[0009] 4. Navigation accuracy and adaptability to complex environments need improvement. Disadvantages: Navigation relies on "stereo vision coordinate transformation + dynamic trajectory planning" and does not introduce a global environment scanning and attitude calibration mechanism. In scenes with dense dynamic obstacles and large changes in lighting, coordinate deviations may occur, and there are no clear navigation accuracy indicators.

[0010] Causes: Existing technologies do not employ laser SLAM systems, making it impossible to globally scan the environment and match pre-built maps; they also lack inertial navigation systems, making it impossible to continuously calibrate the AGV's attitude and direction of movement; relying solely on local visual data and dynamic path algorithms for navigation, they lack global positioning benchmarks and attitude stability guarantees, making it difficult to cope with complex environmental interference.

[0011] Therefore, there is an urgent need to propose a new technical solution to address the above problems. Summary of the Invention

[0012] This application discloses an automatic loading method and related products for vans, which can maximize the use of the van's space.

[0013] The first aspect of this application discloses an automatic loading method for a van, the method comprising the following steps: S1. The AGV receives system instructions, parses and confirms the workspace coordinates of the target loading point; S2 and AGV plan their paths based on the obtained work space coordinates and then move autonomously to the predetermined position at the rear of the van. S3. After arriving at the predetermined position, adjust the AGV's pose based on camera image recognition and analysis, use the camera to collect images of the carriage, and convert the carriage images into carriage space coordinates. S4. Combine the spatial coordinates of the carriage with the cargo stacking requirements for calculation, generate the optimal stacking layout based on the combined calculation results, and the AGV performs operations according to the stacking layout. S5. After the operation is completed, the AGV autonomously navigates back to the preset base.

[0014] As an optional implementation, in the first aspect of the embodiments of this application, in step S1, after the operator issues a cargo transfer task on the remote control tablet, the system issues a work instruction to the AGV. After receiving the instruction, the AGV starts the laser SLAM navigation system to scan the environment. Combined with the pre-built environment map, it identifies its own coordinates, the spatial coordinates of the box truck, and environmental obstacles. The scheduling system path planning module plans the optimal movement path for the AGV based on the feedback information. In addition, the visual auxiliary positioning module collects images of the target area and matches them with map features to further calibrate the coordinate accuracy.

[0015] As an optional implementation, in the first aspect of the embodiments of this application, in step S2, the AGV travels along the planned path, the lidar scans ahead in real time, and when an obstacle is detected, the motion control unit adjusts the speed or direction through the dynamic window method, and continues to travel after bypassing the obstacle. The inertial navigation system continuously provides AGV attitude and motion state information so that the travel direction is consistent with the path.

[0016] As an optional implementation, in the first aspect of the embodiments of this application, in step S3, after the AGV identifies and successfully moves to the rear of the van, it identifies the first stopping position and adjusts the vehicle's posture based on the camera image recognition analysis to achieve precise docking; during the navigation process at the van entrance, the height of the van steps is detected, and the climbing length required for the AGV to move to the first stacking position is calculated.

[0017] As an optional implementation, in the first aspect of the embodiments of this application, in step S3, the binocular camera is used to collect images of the carriage and convert them into spatial coordinates of the carriage, including carriage flatness detection and obstacle recognition. If an abnormal situation is detected, the AGV will automatically send an alarm message to the remote control tablet. After the alarm message is manually processed or the abnormality is skipped, the robot arm enters the predetermined position for stacking operations.

[0018] As an optional implementation, in the first aspect of the embodiments of this application, in step S4, the AGV plans the stacking type based on the box size given manually or measured by 3D scanning, combined with the calculated box volume and special cargo stacking requirements, and adopts the same box type Z-shaped stacking or different box types mixed stacking.

[0019] As an optional implementation, in the first aspect of the embodiments of this application, in step S4, the palletizing operation begins, the roller conveyor starts, and the neatly stacked goods are transferred to the loading position. The robot arm collects the shape and position information of the goods through a vision camera and a laser rangefinder, adjusts its posture to grab the goods and stack them to the predetermined palletizing position in the cargo compartment. After detecting that the current row of goods has been stacked, the AGV will retreat one working distance and start the next row of goods stacking operation according to the stacking plan. The entire operation includes real-time obstacle recognition, goods tilt detection, and material drop detection. If the grabbing of goods is detected to have fallen, the original operation will stop, and the image will be collected by a binocular camera to identify the position of the fallen goods. After the fallen goods are grabbed again and accurately stacked, the original operation process will resume. If it cannot be identified, an alarm will be automatically sent to the remote control tablet to remind manual intervention.

[0020] As an optional implementation, in the first aspect of the embodiments of this application, in step S5, after the goods operation is completed, the AGV sends a work completion signal to the scheduling system, and the scheduling system triggers the return to the storage location process: Storage location coordinate identification: The scheduling system sends a return command to the AGV, and the AGV identifies the spatial coordinates of the storage location through a laser SLAM navigation system and vision-assisted positioning; Return path planning: The path planning module plans the optimal return path based on the AGV's current position, warehouse coordinates, and real-time environment; Navigation Return: The AGV travels along the return path, with lidar and inertial navigation systems ensuring driving accuracy and obstacle avoidance. Finally, it stops at the warehouse location and reports back to the scheduling system that it has returned to the warehouse location, completing the transfer process.

[0021] As an optional implementation, in the first aspect of the embodiments of this application, in step S3, after the AGV arrives at the rear of the van and precisely docks, the binocular camera and lidar ranging and acquisition module are activated, and the two-dimensional van image is converted into three-dimensional spatial coordinates through an image coordinate transformation algorithm, and the following core operations are performed: Carriage parameter detection: Accurately measure the length, width, and height of the carriage interior, determine the flatness of the ground through point cloud density analysis, and identify fixed obstacles inside the carriage; Loading condition calculation: Detect the height of the steps at the entrance of the truck bed, and combine the height of the AGV's own chassis and climbing ability to deduce the climbing length required for the AGV to move to the first stacking position, so as to ensure that the path is compatible with the entrance of the truck bed; Abnormal alarm control: If an abnormality in flatness is detected, an obstacle is not removed, or the height of a step exceeds the AGV's climbing ability, the AGV will send an alarm message containing the abnormality type and location coordinates to the remote control tablet via wireless communication, suspend the stacking operation, and wait for manual handling or skipping the abnormality before the robotic arm can enter the predetermined stacking position.

[0022] The second aspect of this application discloses an automated guided vehicle, comprising: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the automatic loading method for vans disclosed in the first aspect of the embodiments of this application.

[0023] A third aspect of this application discloses a computer-readable storage medium storing a computer program, wherein the computer program causes a computer to execute the automatic loading method for a van disclosed in the first aspect of this application.

[0024] Compared with related technologies, the embodiments of this application have the following beneficial effects: To address the shortcomings of existing technologies in terms of vehicle environment adaptability, palletizing flexibility, anomaly handling, and navigation stability, this AGV loading solution achieves full-process automation, comprehensive scenario coverage, and predictable risks. Specifically, it includes: 1) Implement pre-processing and detection of the cargo compartment environment: Scan the dimensions, flatness, and obstacles of the cargo compartment with a binocular camera, calculate the step height and AGV climbing length, and automatically alarm when there is an abnormality to ensure that the cargo compartment environment meets the requirements before loading; 2) Improve palletizing scenario adaptability: Based on the box size scanned by 3D or manually entered, intelligently plan the pallet type, support Z-shaped palletizing of the same specification and mixed palletizing of different specifications, and maximize the utilization of the car compartment space; 3) Construct a closed-loop system for handling anomalies: Real-time detection of cargo tilting and falling materials triggers the robotic arm to automatically re-grab the cargo or a manual alarm is triggered, reducing operation interruptions; 4) Improve navigation accuracy and stability: The three-mode fusion navigation is achieved by using laser SLAM + vision assistance + inertial navigation, which controls the coordinate error to ≤10mm. Dynamic obstacle avoidance uses the dynamic window method to adapt to complex environments. 5) Achieve full-process automation: Covering the entire process from instruction reception to navigation to the truck, pre-processing and inspection, palletizing, anomaly handling, and return to the warehouse, without the need for manual intervention, thus improving loading efficiency and reliability. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 This is a schematic flowchart of an automatic loading method for a van disclosed in an embodiment of this application; Figure 2 This is a schematic diagram of the structure of an automated guided vehicle disclosed in an embodiment of this application.

[0027] Among them: 1. Memory; 2. Processor. Detailed Implementation

[0028] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0029] The terms “comprising” and “having”, and any variations thereof, in this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such process, method, product, or device.

[0030] This application discloses an automatic loading method and related products for vans. By scanning the van in advance and performing calculations, an optimal stacking layout can be generated to maximize the use of the van's space, thus adapting to special goods such as fragile items and electronic products.

[0031] The technical solution of this application will be described in detail below with reference to specific embodiments.

[0032] Please see Figure 1 The automated loading method for this van follows a main process: AGV receives instructions → navigates to the van → robotic arm performs palletizing operations → returns to the storage location. This process may include the following steps: S1. The AGV receives system instructions, parses and confirms the workspace coordinates of the target loading point; S2 and AGV plan their paths based on the obtained work space coordinates and then move autonomously to the predetermined position at the rear of the van. S3. After arriving at the predetermined position, adjust the AGV's pose based on camera image recognition and analysis, use the camera to collect images of the carriage, and convert the carriage images into carriage space coordinates. S4. Combine the spatial coordinates of the carriage with the cargo stacking requirements for calculation, generate the optimal stacking layout based on the combined calculation results, and the AGV performs operations according to the stacking layout. S5. After the operation is completed, the AGV autonomously navigates back to the preset base.

[0033] In some embodiments, during step S1, the AGV receives system instructions and identifies spatial coordinates: After the operator issues a cargo transfer task on the remote control tablet, the system issues an operation instruction to the AGV. After receiving the instruction, the AGV starts the laser SLAM navigation system to scan the environment. Combined with the pre-built environment map, it identifies its own coordinates, the spatial coordinates of the box truck, and environmental obstacles. The scheduling system's path planning module plans the optimal movement path for the AGV based on the feedback information. At the same time, the vision-assisted positioning module collects images of the target area and matches them with map features to further calibrate the coordinate accuracy (error ≤ 10mm), providing a foundation for accurate navigation.

[0034] In some embodiments, during step S2, the AGV navigates to the van: the AGV travels along the planned path: the lidar scans ahead in real time, and when an obstacle is detected, the motion control unit adjusts the speed or direction using the dynamic window method (obstacle avoidance algorithm), and continues to travel after bypassing the obstacle. The inertial navigation system continuously provides AGV attitude and motion status information to ensure that the travel direction is consistent with the path.

[0035] In some embodiments, in step S3, after arriving at the van, the docking is accurately identified and the volume of the box is calculated: after the AGV identifies and successfully moves to the rear of the van, the first stopping position is identified and the vehicle posture is adjusted according to the camera image recognition analysis to achieve accurate docking; during the navigation process at the entrance of the van, the height of the van steps is detected and the climbing length required for the AGV to move to the first stacking position is calculated. Simultaneously, the system uses binocular cameras to capture images of the carriage and convert them into spatial coordinates, supporting carriage flatness detection and obstacle recognition. If an anomaly is detected, the AGV will automatically send an alarm message to the remote control tablet. After manual handling of the alarm message or skipping the anomaly, the robotic arm will enter the predetermined stacking position. By automatically calculating the stack type, the optimal stacking scheme can be obtained, and it can achieve Z-shaped stacking of boxes of the same specification or mixed stacking of boxes of different specifications.

[0036] In some embodiments, in step S4, the stacking pattern is planned and the stacking operation is completed according to the box size: The AGV intelligently plans the stacking pattern based on the box size given manually or measured by 3D scanning, combined with the calculated box volume and special cargo stacking requirements (such as fragile items like ceramics or goods that are not suitable for compression, such as fruits and electronic products, which need to be stacked on the top layer of the cargo compartment), and adopts the Z-shaped stacking of the same box type or the mixed stacking of different box types to maximize the rational use of the cargo compartment space.

[0037] Palletizing begins with the roller conveyor starting up, transporting the neatly stacked goods to the loading position. The robotic arm, using a vision camera and laser rangefinder, collects the shape and position information of the goods, adjusts its posture to grasp the goods, and precisely places them into the designated palletizing position within the cargo compartment. Once the current row of goods is completed, the AGV will retreat one working distance and begin stacking the next row according to the planned pallet pattern. The entire operation features real-time obstacle recognition, goods tilt detection, and drop detection. If a dropped item is detected, the operation stops, and images are captured by a binocular camera to identify the position of the dropped item. After re-grabbing and accurately stacking the dropped item, the original operation resumes. If no detection is found, an alarm is automatically sent to the remote control tablet to prompt manual intervention.

[0038] In some embodiments, during step S5, after the task is completed, the AGV autonomously identifies and navigates back to the warehouse: after the goods operation is completed, the AGV sends a task completion signal to the scheduling system, and the scheduling system triggers the return-to-warehouse process. Warehouse location coordinate identification: The scheduling system sends a return command to the AGV. The AGV uses a laser SLAM navigation system and vision-assisted positioning to identify the spatial coordinates of the warehouse location (a designated stopping position in the warehouse, pre-stored in the environmental map). Return path planning: The path planning module plans the optimal return path (obstacle avoidance, shortest distance) based on the AGV's current position, warehouse coordinates, and real-time environment. Navigation Return: The AGV travels along the return path, with LiDAR and inertial navigation systems ensuring driving accuracy and obstacle avoidance. It then accurately stops at the warehouse location and reports back to the scheduling system that it has returned to the warehouse location, completing the transfer process.

[0039] The loading method of the present invention has the following technical advantages: A. Carriage Environment Pretreatment Technology In some embodiments, in step S3, the preprocessing of the cargo compartment environment is performed as follows: After the AGV arrives at the rear of the van and precisely docks, the binocular camera and lidar ranging acquisition module are activated. The two-dimensional cargo compartment image is converted into three-dimensional spatial coordinates through an image coordinate transformation algorithm, and the following core operations are performed: Carriage parameter detection: Accurately measure the length, width, and height of the carriage interior, determine the flatness of the ground through point cloud density analysis (flatness deviation > 5mm is considered abnormal), and identify fixed obstacles inside the carriage (such as abandoned tools or cargo fragments). Loading condition calculation: Detect the height of the steps at the entrance of the truck bed, and combine the height of the AGV's own chassis and climbing ability to deduce the climbing length required for the AGV to move to the first stacking position, so as to ensure that the path is compatible with the entrance of the truck bed; Anomaly Alarm Control: If an anomaly is detected in the flatness, obstacles are not removed, or the step height exceeds the AGV's climbing capacity, the AGV will send an alarm message containing the anomaly type and location coordinates to the remote control tablet via wireless communication, pausing the stacking operation. The robotic arm can only enter the predetermined stacking position after manual handling or skipping the anomaly. A binocular camera enables omnidirectional scanning of the vehicle's dimensions, flatness, and obstacles. Combined with climbing length calculation and anomaly alarms, this allows for proactive risk prediction and avoidance, ensuring the vehicle environment meets operational requirements before loading.

[0040] B. Container stacking planning and multi-mode palletizing technology A complete technical solution based on box size, stack type calculation, and stacking execution: Size Acquisition: Supports two input methods: one is to automatically scan the cabinet with a 3D laser scanner (accuracy ≤2mm) to generate length, width and height data; the other is to manually input the dimensions on a remote control tablet, and the system will automatically verify the data. Stacking planning: Maximize the rational use of the target cargo space, and generate the optimal stacking plan by combining the cargo box size, container size and special cargo rules (such as placing fragile items such as ceramics on the top layer of the cargo box, and intelligently planning to place non-compressible goods such as fruits / electronic products on the upper layer of the cargo box); Stacking execution: The robotic arm can execute two modes according to the plan. For boxes of the same specification, a Z-shaped stacking is used (the first row is left → right, and the second row is right → left alternately). For boxes of different specifications, a layered mixed stacking is used (the boxes are layered from low to high according to their height, and the same layer is arranged according to the width). After each row is completed, the AGV automatically retreats one unit and continues to work after the laser ranging data meets the requirements of the next row.

[0041] By constructing an active palletizing planning logic of size input, rule matching, and multi-mode execution, the following are achieved: First, accurate box dimensions are obtained through 3D scanning / manual input, solving the problem of lack of size data support in existing technologies; second, the specific execution methods of Z-shaped palletizing and layered mixed palletizing are clarified, covering multi-specification box mixed palletizing scenarios not covered by existing technologies; and third, special cargo rule planning is combined to balance space utilization and cargo safety, breaking through the limitations of the single palletizing mode in existing technologies.

[0042] C. Real-time cargo anomaly detection and automatic recovery technology Deploy a vision camera and a laser rangefinder in the robotic arm's palletizing area to create a closed loop for anomaly handling: Anomaly detection: The vision camera captures the posture of the goods in real time. If the tilt angle is greater than 5°, it is considered that the goods are tilted. The laser rangefinder detects the relative position of the goods and the robotic arm. If the goods are out of the gripping range and the height drops by more than 10mm, it is considered that the goods are dropped. Tiered processing: If the goods are tilted, the robotic arm pauses its operation, re-collects the goods' pose data, adjusts the gripping angle, and then replenishes the weight; if the goods are dropped, the binocular camera identifies the three-dimensional coordinates of the dropped goods, the robotic arm matches the gripping force according to the weight of the goods, and re-grabs and replenishes the weight; if the binocular camera cannot identify the dropped position or the dropped position is difficult to automatically re-grab, the AGV automatically sends a manual intervention alarm (with an image of the dropped area) to the remote control tablet.

[0043] Existing technologies only prevent fork tilting and cargo slippage by optimizing fork torque, lacking real-time anomaly detection capabilities. Once an anomaly occurs, operations must be manually interrupted and adjustments made, lacking an automatic recovery mechanism.

[0044] The breakthroughs of this invention are: first, the combination of a vision camera and a lidar dual sensor enables precise quantitative judgment of anomalies (tilt angle, drop height), solving the problem of existing technologies lacking real-time monitoring methods; second, the design incorporates a hierarchical logic that recovers if it can self-heal and alarms if it cannot self-heal, allowing the robotic arm to automatically replenish code, reducing human intervention, overcoming the shortcomings of existing technologies that rely entirely on manual intervention for anomalies, and improving the continuity of operations.

[0045] D. Laser SLAM + Visual Assistance + Inertial Navigation Fusion Navigation Technology The AGV adopts a three-mode fusion navigation architecture to ensure precise and stable driving. Positioning calibration: After receiving the instruction, the AGV starts the laser SLAM navigation system to scan the environment and match the pre-built warehouse map to identify its own spatial coordinates; at the same time, the vision-assisted positioning module is started to collect the image of the truck's rear, match it with map features, and calibrate the coordinate error to ≤10mm. The vision-assisted positioning module can be a binocular camera. Dynamic obstacle avoidance: During driving, the lidar detects dynamic obstacles (such as pedestrians or other AGVs) in real time. The motion control unit calls the dynamic window method to calculate the speed-direction feasible window and adjust the AGV's driving speed and direction. Stable attitude: The inertial navigation system continuously collects AGV angular velocity and acceleration data and calibrates the driving direction in real time (if a 2° deviation in direction is detected, the wheel steering is immediately adjusted); when returning to the warehouse, the laser SLAM navigation system locates the warehouse coordinates and visually assists in matching the warehouse ground markings (such as black positioning lines), and the final docking error is ≤5mm.

[0046] Existing technologies rely solely on stereo vision coordinate transformation and dynamic trajectory planning navigation, lacking a global environment scanning and attitude calibration mechanism. This makes them prone to coordinate deviations in scenarios with dense dynamic obstacles and large changes in lighting, and there are no clear accuracy indicators.

[0047] This invention achieves three-mode fusion navigation that combines global positioning, accuracy calibration, and attitude stabilization. First, it introduces a laser SLAM system to achieve global environment scanning and map matching, solving the problem of the lack of a global positioning reference in existing technologies. Second, it adds an inertial navigation system to continuously calibrate attitude, avoiding directional deviations caused by uneven ground and improving adaptability to complex environments.

[0048] E. AGV Autonomous Return to Storage and Data Feedback Technology Integrate inventory return and data feedback into automated processes: Return to warehouse trigger: After palletizing is completed, the AGV sends a work completion signal to the remote scheduling system; the scheduling system uses a binocular camera to take a panoramic view of the cargo compartment, compares and verifies it with the preset full-pallet image, and then sends a return to warehouse instruction to the AGV; Warehouse location positioning: After the AGV receives the return command, the laser SLAM system matches the warehouse map to locate the warehouse location coordinates, and the vision assistance module collects the warehouse location identification image to calibrate the warehouse location recognition error to ≤3mm; Data Feedback: After the AGV docks at the storage location, it reports its return to the storage location to the scheduling system and uploads data on the return process (navigation error, docking accuracy). This data is integrated with loading data (loading time, number of anomalies) to form a complete operational data chain. While the AGV is docked at the storage location, it automatically enters wireless charging mode. To ensure uninterrupted operation during high-intensity work, multiple AGVs can be scheduled in an orderly manner.

[0049] Existing technologies lack an autonomous warehousing process and only provide feedback on loading data (such as docking errors and torque distribution), resulting in an incomplete data chain that fails to meet the needs of intelligent AGV charging planning in real-world application scenarios.

[0050] This invention integrates autonomous warehousing and complete data feedback. By detecting full stacks of cargo compartments and accurately locating warehouse positions, it enables AGVs to autonomously return to the warehouse, filling the gaps in existing technologies.

[0051] Obviously, the loading method of the present invention has the following advantages: Enhanced adaptability to different cargo compartment environments: Thanks to the newly added binocular camera compartment scanning module, flatness, obstacles, and step height can be detected in advance, avoiding cargo damage or loading failure caused by the failure of existing technology to anticipate compartment problems, thus ensuring reliable loading conditions; Wider coverage of palletizing scenarios: Through 3D scanning / manual input of box dimensions and intelligent stacking planning, it supports Z-shaped palletizing and mixed stacking of multiple specifications, solving the limitation of existing technologies that can only stack single items. It can be adapted to fragile items, electronic products and other special goods, maximizing the space in the carriage. Higher operational continuity: Real-time anomaly detection using a vision camera and laser rangefinder, along with the robotic arm's automatic re-grabbing function, reduces the frequency of reliance on manual intervention in existing technologies and improves operational efficiency; Superior navigation accuracy and stability: The global positioning of laser SLAM, the millimeter-level calibration with vision assistance (error ≤10mm) and the attitude stabilization of inertial navigation are better able to cope with dynamic obstacles and lighting interference than the existing vision + trajectory planning technology, ensuring accurate AGV docking and safe driving.

[0052] Please see Figure 2 , Figure 2 This is a schematic diagram of the structure of an automated loading vehicle disclosed in an embodiment of this application. Figure 2 As shown, the automated loading truck may include: a memory 1 storing executable program code, and a processor 2 coupled to the memory 1; wherein the processor 2 calls the executable program code stored in the memory 1 to execute the automated loading method of the van disclosed in the above embodiments.

[0053] This application discloses a computer-readable storage medium storing a computer program that causes a computer to execute the automatic loading method for a van disclosed in the above embodiments.

[0054] This application also discloses an application publishing platform, which is used to publish computer program products. When the computer program products are run on a computer, the computer performs some or all of the steps of the methods described in the above method embodiments.

[0055] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Those skilled in the art should also recognize that the embodiments described in the specification are optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0056] In the various embodiments of this application, it should be understood that the sequence number of each process does not necessarily imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0057] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they can be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0058] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0059] If the aforementioned integrated units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-accessible memory. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several requests to cause a computer device (which can be a personal computer, server, or network device, specifically a processor in the computer device) to execute some or all of the steps of the methods described in the various embodiments of this application.

[0060] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0061] Based on the disclosure and teachings of the foregoing specification, those skilled in the art can make changes and modifications to the above embodiments. Therefore, the invention is not limited to the specific embodiments described above, and any obvious improvements, substitutions, or modifications made by those skilled in the art based on this invention are within the scope of protection of this invention. Furthermore, although some specific terms are used in this specification, these terms are only for convenience of explanation and do not constitute any limitation on the invention.

Claims

1. An automated loading method for vans, characterized in that, Includes the following steps: S1. The AGV receives system instructions, parses and confirms the workspace coordinates of the target loading point; S2 and AGV plan their paths based on the obtained work space coordinates and then move autonomously to the predetermined position at the rear of the van. S3. After arriving at the predetermined position, adjust the AGV's pose based on camera image recognition and analysis, use the camera to collect images of the carriage, and convert the carriage images into carriage spatial coordinates. S4. Combine the spatial coordinates of the carriage with the cargo stacking requirements for calculation, generate the optimal stacking layout based on the combined calculation results, and the AGV performs operations according to the stacking layout. S5. After the operation is completed, the AGV autonomously navigates back to the preset base.

2. The automatic loading method for a van as described in claim 1, characterized in that, In step S1, after the operator issues a cargo transfer task on the remote control tablet, the system issues a work instruction to the AGV. After receiving the instruction, the AGV starts the laser SLAM navigation system to scan the environment. Combined with the pre-built environment map, it identifies its own coordinates, the spatial coordinates of the box truck, and environmental obstacles. The scheduling system's path planning module plans the optimal movement path for the AGV based on the feedback information. In addition, the visual-assisted positioning module collects images of the target area and matches them with map features to further calibrate the coordinate accuracy.

3. The automatic loading method for a van as described in claim 1, characterized in that, In step S2, the AGV travels along the planned path. The lidar scans ahead in real time. When an obstacle is detected, the motion control unit adjusts the speed or direction using the dynamic window method. After bypassing the obstacle, the AGV continues to travel. The inertial navigation system continuously provides the AGV's attitude and motion status information to ensure that the travel direction is consistent with the path.

4. The automatic loading method for a van as described in claim 1, characterized in that, In step S3, after the AGV identifies and successfully moves to the rear of the van, it identifies the first stopping position and adjusts the vehicle's posture based on camera image recognition and analysis to achieve precise docking. During the navigation process at the van entrance, the height of the van steps is detected, and the required ramp length for the AGV to move to the first stacking position is calculated.

5. The automatic loading method for a van as described in claim 4, characterized in that, In step S3, the binocular camera is used to collect images of the carriage and convert them into spatial coordinates of the carriage, including carriage flatness detection and obstacle recognition. If an abnormality is detected, the AGV will automatically send an alarm message to the remote control tablet. After the alarm message is manually processed or the abnormality is skipped, the robotic arm will enter the predetermined position for stacking operations.

6. The automatic loading method for a van as described in claim 1, characterized in that, In step S4, the AGV plans the stacking type based on the box dimensions given manually or measured by 3D scanning, combined with the calculated box volume and special cargo stacking requirements.

7. The automatic loading method for a van as described in claim 6, characterized in that, In step S4, the palletizing operation begins, the roller conveyor starts, and the neatly stacked goods are transferred to the loading position. The robotic arm uses a vision camera and a laser rangefinder to collect the shape and position information of the goods, adjusts its posture to grab the goods and stack them in the predetermined palletizing position in the cargo compartment. After the current row of goods is detected to be stacked, the AGV will retreat one working distance and start the next row of goods stacking operation according to the stacking plan. The entire operation is carried out in real time with obstacle recognition, goods tilt detection, and material drop detection. If the grabbing of goods is detected to have fallen, the original operation will stop, and the image will be collected by a binocular camera to identify the position of the fallen goods. After the fallen goods are grabbed again and accurately stacked, the original operation process will resume. If the detection fails, an alarm will be automatically sent to the remote control tablet to remind manual intervention.

8. The automatic loading method for a van as described in claim 1, characterized in that, In step S5, after the goods operation is completed, the AGV sends a work completion signal to the scheduling system, and the scheduling system triggers the return-to-warehouse process. Storage location coordinate identification: The scheduling system sends a return command to the AGV, and the AGV identifies the spatial coordinates of the storage location through a laser SLAM navigation system and vision-assisted positioning; Return path planning: The path planning module plans the optimal return path based on the AGV's current position, warehouse coordinates, and real-time environment; Navigation Return: The AGV travels along the return path, with lidar and inertial navigation systems ensuring driving accuracy and obstacle avoidance. Finally, it stops at the warehouse location and reports back to the scheduling system that it has returned to the warehouse location, completing the transfer process.

9. The automatic loading method for a van as described in claim 4, characterized in that, In step S3, after the AGV accurately docks with the rear of the van, the binocular camera and lidar ranging and acquisition module are activated. The two-dimensional image of the van is converted into three-dimensional spatial coordinates through an image coordinate transformation algorithm, and the following core operations are performed: Carriage parameter detection: Accurately measure the length, width, and height of the carriage interior, determine the flatness of the ground through point cloud density analysis, and identify fixed obstacles inside the carriage; Loading condition calculation: Detect the height of the steps at the entrance of the truck bed, and combine the height of the AGV's own chassis and climbing ability to deduce the climbing length required for the AGV to move to the first stacking position, so as to ensure that the path is compatible with the entrance of the truck bed; Abnormal alarm control: If an abnormality in flatness is detected, an obstacle is not removed, or the height of a step exceeds the AGV's climbing ability, the AGV will send an alarm message containing the abnormality type and location coordinates to the remote control tablet via wireless communication, suspend the stacking operation, and wait for manual handling or skipping the abnormality before the robotic arm can enter the predetermined stacking position.

10. An automated guided vehicle, characterized in that, The method includes a memory storing executable program code and a processor coupled to the memory; wherein the processor invokes the executable program code stored in the memory to perform the method as described in any one of claims 1 to 9.

11. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 9.

Images