Clearing method and device for linkage of clearing machine and portal crane
By setting up GNSS differential positioning terminals and lidar SLAM mapping at the ship's hatch, a unified local coordinate system was established. Combined with V2X communication and collision detection, the problem of inconsistent coordinates caused by tidal changes was solved, enabling precise positioning and collaborative operation between the cleaning machine and the gantry crane, thus improving the automation and safety of port cleaning operations.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-29
- Publication Date
- 2026-03-31
AI Technical Summary
In existing technologies, ship cabin cleaning equipment cannot achieve precise positioning and collaborative operation due to the inconsistency of the coordinate system caused by tidal changes in ships. This can easily lead to blind spots or repetitive operations. Furthermore, collisions are prone to occur when the cleaning machine and the gantry crane work together. Existing systems lack an effective multi-device collaboration mechanism.
By setting up a GNSS differential positioning terminal at the hatch as a dynamic reference point, a unified local coordinate system with the hatch positioning terminal as the origin is established. Combined with lidar SLAM mapping and V2X communication, the precise positioning and real-time coordination between the cleaning machine and the gantry crane are realized, generating a spiral-shaped operation path, and collision avoidance control is performed through a collision detection algorithm.
It enables precise avoidance and efficient collaboration between the cleaning machine and the gantry crane, eliminates positioning deviations and blind spots caused by inconsistent coordinate systems, and improves the automation level and safety of ship hold cleaning operations.
Smart Images
Figure CN121757739A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of dock unloading control, and in particular to a cleaning method and device that links a cleaning machine with a gantry crane. Background Technology
[0002] With the rapid development of global trade and the continuous growth of port throughput, the automation and intelligentization of port loading and unloading technologies have become key factors in enhancing port competitiveness. In bulk carrier loading and unloading operations, hold cleaning is a crucial part of the entire process, and its efficiency directly affects ship turnaround time and port operating costs. Traditional hold cleaning operations mainly rely on manual operation of small loaders or bulldozers. This method is not only inefficient and labor-intensive, but also poses serious safety hazards in the confined environment of the hold. Especially when handling bulk cargoes such as coal and ore, the high dust environment poses a serious threat to the health of workers, while the narrow working space and irregular cargo stacking patterns further increase the difficulty of the operation.
[0003] Against this backdrop, automated cargo hold cleaning equipment has emerged, but it still faces numerous technical challenges in practical applications. First, the internal environment of a ship's hold is complex and variable, with irregularly stacked cargo, making it difficult for existing cleaning machinery to achieve efficient environmental perception and autonomous path planning. In high-dust environments, traditional visual sensors often fail, preventing the equipment from accurately acquiring information about its surroundings. Second, in actual operations, the cleaning machine needs to work in conjunction with the gantry crane's grab bucket—the cleaning machine pushes materials from the hold's edge to the central area, while the gantry crane's grab bucket retrieves materials from the hold. Both operating simultaneously in the same confined space greatly increases the risk of collisions. However, current technology lacks an effective multi-device coordination mechanism, failing to achieve real-time information sharing and dynamic obstacle avoidance between the cleaning machine and the gantry crane.
[0004] Existing technical solutions have significant shortcomings. For example, CN114620189A discloses an intelligent tracked ship hold cleaning machine. Although it has a material handling mechanism and a pusher mechanism, enabling it to move autonomously within the ship hold via a tracked walking mechanism and push cargo to the center of the hold, this solution only considers the independent operation of a single cleaning machine, completely neglecting the issue of coordination with the gantry crane grab. This patent neither addresses the optimized design of the cleaning machine's operating path nor provides an anti-collision mechanism between the cleaning machine and the gantry crane grab, failing to meet the needs of multi-equipment coordination in actual port operations. Furthermore, existing intelligent cleaning systems generally lack the ability to optimize unloading operation strategies, failing to dynamically adjust cleaning strategies based on different ship types, cargo characteristics, and the real-time operating status of the gantry crane grab, resulting in low overall operational efficiency and low equipment utilization.
[0005] It is particularly important to point out that in actual operations at port bulk cargo terminals, the berthing position of ships will dynamically move with the tides. The gantry crane, as a shore-based fixed equipment, uses the port's fixed coordinate system, while the cleaning machine moves within the hold and uses its own relative coordinate system. The inconsistency of the three coordinate systems makes it impossible to achieve precise positioning and effective coordination, which can easily lead to blind spots or repetitive operations. This technical bottleneck seriously restricts the development of automation in port cleaning operations. Summary of the Invention
[0006] To address the operational blind spots caused by the inconsistency in coordinates among the ship's hold, gantry crane, and in-hull cleaning machine due to tidal changes, this application provides a cleaning method and apparatus that links the cleaning machine and gantry crane. By setting a GNSS differential positioning terminal at the hatch opening as a dynamic reference point, a unified local coordinate system is established with the hatch positioning terminal as the origin. The pre-calibrated fixed coordinates of the gantry crane and the relative pose of the cleaning machine obtained through lidar SLAM mapping are uniformly transformed into this local coordinate system, thereby improving the linkage efficiency between the cleaning machine and the gantry crane.
[0007] One aspect of this application provides a clearance method involving a clearance machine and a gantry crane, comprising: acquiring positioning data, including GNSS differential positioning coordinates of the hatch positioning terminal, pre-calibrated coordinates of the gantry crane, and the relative pose of the clearance machine obtained by lidar SLAM mapping; establishing a local coordinate system with the hatch positioning terminal as the origin through coordinate transformation; acquiring a three-dimensional point cloud of materials inside the hatch and generating an operating path for the clearance machine; wherein the operating path is a spiral-shaped path that pushes materials from the hatch side to the center area of the hatch; collecting joint angle data of the first and second robotic arms of the gantry crane and calculating the real-time position of the gantry crane grab bucket based on the joint angle data; wherein the first robotic arm is the main arm connecting the gantry crane operating compartment and the second robotic arm, and the second robotic arm is the auxiliary arm connecting the first robotic arm and the grab bucket cable; setting the real-time position of the gantry crane grab bucket as a high-cost area of the local cost map and correcting the operating path of the clearance machine; obtaining the minimum distance between the corrected operating path of the clearance machine and the real-time position of the gantry crane grab bucket through a collision detection algorithm, and locking the grab bucket lifting mechanism for emergency stop control when the minimum distance is less than a preset safety threshold.
[0008] Furthermore, a local coordinate system is established with the hatch positioning terminal as the origin through coordinate transformation, including: transforming the GNSS differential positioning coordinates of the hatch positioning terminal. Set the origin O of the local coordinate system; set the world coordinates of the gantry crane as the local coordinate system origin; Through the first coordinate transformation matrix Transform to the local coordinate system to obtain the position of the gantry crane in the local coordinate system. The position vector relative to the hatch positioning terminal is obtained from the SLAM mapping of the warehouse cleaning machine's lidar. and attitude angle Construct the second coordinate transformation matrix Transform the machine's own coordinate system to the local coordinate system; map any point in the world coordinate system to the local coordinate system using the rotation matrix R: ,in, This represents the coordinates of a point in the local coordinate system; R represents the coordinates of a point in the world coordinate system; R is a 3x3 rotation matrix, determined by the heading angle of the hatch positioning terminal, such that the X-axis of the local coordinate system points longitudinally to the cabin, the Y-axis points laterally to the cabin, and the Z-axis is vertically upward.
[0009] Furthermore, acquiring 3D point cloud data of materials inside the chamber and generating the cleaning machine's operating path includes: scanning the chamber environment using a lidar mounted on the cleaning machine to acquire 3D point cloud data of the material surface based on a local coordinate system. ; for 3D point cloud data The compartment is divided into m x n grid cells, and the average material height within each grid cell (i, j) is calculated. and material distribution density ; Material height Greater than the preset height threshold And the distance from the bulkhead is less than the preset distance. The grid markers are designated as areas to be cleaned; based on the distribution of these areas, a spiral or zigzag work path is generated.
[0010] Furthermore, based on the distribution of the area to be cleaned, a spiral-shaped working path is generated, including: determining the outline of the inner boundary of the compartment and generating the outermost rectangular path parallel to the compartment wall; setting the effective pushing width w of the cleaning machine bucket; starting from the outermost path, after each round of pushing is completed, shifting inward by a distance w to generate the inner path;
[0011] Furthermore, the joint angle data of the first and second robotic arms of the gantry crane are collected, and the real-time position of the gantry crane grab bucket is calculated based on the joint angle data, including: collecting the rotation angle of the first robotic arm relative to the gantry crane operating compartment. and pitch angle Collect the pitch angle of the second robotic arm relative to the first robotic arm. ; Collect the release length L of the grab bucket cable; Position the gantry crane in the local coordinate system Using this as a reference point, calculate the real-time position of the gantry crane grab bucket in the local coordinate system. .
[0012] Furthermore, the real-time position of the gantry crane grab is set as a high-cost area on the local cost map to correct the operation path of the clearing machine. This includes: constructing a local cost map based on grid cells and initializing all grid cost values to 0; receiving the real-time position of the gantry crane grab through the hatch relay communication unit. Real-time position of the gantry crane grab bucket Mapped to corresponding raster coordinates using raster coordinates Centered on, with radius All grid cells within the range are set to high generation value. Among them, high-value The corresponding area is a restricted area; retrieve the current location of the clearance machine. Heading angle and speed Within the dynamic constraints of the clearing machine, multiple predicted trajectories are generated; the minimum distance between each predicted trajectory and the high-cost area is calculated, and trajectories that enter the high-cost area are eliminated; the trajectory closest to the operation path is selected as the correction path; the correction path is converted into clearing machine control commands to avoid the gantry crane grab bucket in real time.
[0013] Furthermore, within the dynamic constraints of the clearing machine, multiple predicted trajectories are generated, including those based on the maximum linear velocity of the clearing machine. Maximum angular velocity and maximum linear acceleration Maximum angular acceleration Determine the boundaries of the speed window; based on the current speed of the clearing machine. and angular velocity Calculate the achievable speed range within the time interval Δt; uniformly sample N speed combinations within the achievable speed range. Where i = 1, 2, ..., N; for each velocity combination According to the kinematic model of the clearance machine, from the current position and heading angle First, generate a sequence of trajectory points for a predicted duration T; then discretize each trajectory into M trajectory points to form the predicted trajectory.
[0014] Furthermore, a collision detection algorithm is used to obtain the minimum distance between the corrected operating path of the clearing machine and the real-time position of the gantry crane grab bucket. When the minimum distance is less than a preset safety threshold, the grab bucket lifting mechanism is locked for emergency stop control, including: setting a cuboid bounding box according to the external dimensions of the clearing machine; setting a cylindrical bounding box according to the external dimensions of the gantry crane grab bucket; obtaining M predicted position points on the corrected path of the clearing machine within a future time T; calculating the minimum distance between the cuboid bounding box and the cylindrical bounding box for each predicted position point; and sending an emergency stop signal to the gantry crane through the hatch relay communication unit when the minimum distance is less than the preset safety threshold.
[0015] Another aspect of this application provides a warehouse clearing device that links a warehouse clearing machine with a gantry crane, comprising: a warehouse clearing machine installed inside the ship's hold, including: a lidar installed on top of the warehouse clearing machine to scan the environment inside the hold and acquire three-dimensional point cloud data; a positioning module to calculate the pose of the warehouse clearing machine relative to the hatch positioning terminal based on the SLAM mapping data of the lidar; a path planning module to generate an operation path based on the three-dimensional point cloud data and pose; and a first V2X communication module to send the position of the warehouse clearing machine and receive the position of the gantry crane grab bucket;
[0016] The gantry crane, located on the dock outside the ship's cabin, includes: a gantry crane operating cabin, located on top of the gantry crane; a first robotic arm, one end of which is hinged to the gantry crane operating cabin and capable of rotating and pitching relative to the gantry crane operating cabin; a second robotic arm, one end of which is hinged to the other end of the first robotic arm and capable of pitching relative to the first robotic arm; and a grab bucket, connected to the other end of the second robotic arm by a steel cable.
[0017] The second V2X communication module is located inside the gantry crane operating compartment; the hatch positioning terminal is located at a predetermined position on the edge of the hatch, including a GNSS differential positioning receiver.
[0018] The hatch relay communication unit (RSU) is set at a preset position in the hatch and establishes wireless communication connections with the first V2X communication module and the second V2X communication module respectively.
[0019] The cleaning machine obtains its position and orientation information relative to the hatch positioning terminal through LiDAR scanning, the gantry crane obtains the position information of the grab bucket through the joint angle sensors of the first and second robotic arms, and the hatch relay communication unit (RSU) forwards the position information and control commands between the cleaning machine and the gantry crane in real time, so as to realize the collaborative operation of the cleaning machine and the gantry crane.
[0020] Compared to existing technologies, the advantages of this application are:
[0021] To address the problems in existing technologies where the ship's cabin coordinate system changes dynamically due to tidal variations, shore-based gantry cranes use a fixed port coordinate system, and in-cabin cleaning machines use their own relative coordinate system—leading to inaccurate positioning and coordination, blind spots, or repetitive operations—this application provides a cleaning method and apparatus that links a cleaning machine with a gantry crane. By setting a GNSS differential positioning terminal at the hatch opening as a dynamic reference point, a unified local coordinate system is established with the hatch positioning terminal as the origin. The pre-calibrated fixed coordinates of the gantry crane and the relative pose of the cleaning machine obtained through LiDAR SLAM mapping are uniformly transformed into this local coordinate system. Based on V2X communication, the position information of each device in the unified coordinate system is synchronized in real time. This enables precise positioning and real-time coordination of dynamic ships, fixed gantry cranes, and mobile cleaning machines, eliminating positioning deviations and blind spots caused by inaccurate coordinate systems. It ensures precise avoidance and efficient collaboration between the cleaning machine and the gantry crane grab, improving the automation level and operational safety of ship cabin cleaning operations. Attached Figure Description
[0022] This application will be further described by way of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting; in these embodiments, the same reference numerals denote the same structures, wherein:
[0023] Figure 1 This is a schematic diagram of a clearance system in which the clearance machine and the gate machine are linked, according to some embodiments of this application;
[0024] Figure 2 This is an exemplary flowchart illustrating a clearance method that links a clearance machine with a door operator, according to some embodiments of this application;
[0025] Figure 3 This is an exemplary flowchart of the coordinate system establishment process according to some embodiments of this application;
[0026] Figure 4 These are schematic diagrams of the ship's cabin shown according to some embodiments of this application;
[0027] Figure 5 This is a schematic diagram of the basic clearing machine's U-shaped running route according to some embodiments of this application. Detailed Implementation
[0028] The methods and systems provided in the embodiments of this application will now be described in detail with reference to the accompanying drawings.
[0029] Example 1
[0030] like Figure 1 As shown, the clearing machine, installed inside the ship's hold, includes:
[0031] A lidar unit, mounted on top of the clearing machine, scans the environment inside the cabin to acquire three-dimensional point cloud data.
[0032] The positioning module calculates the pose of the clearing machine relative to the hatch positioning terminal based on the SLAM mapping data from the lidar.
[0033] The path planning module generates the operation path based on 3D point cloud data and pose.
[0034] The first V2X communication module sends the location of the clearing machine and receives the location of the gantry crane grab bucket;
[0035] The gantry crane, located on the dock outside the ship's hold, includes:
[0036] The gantry crane control compartment is located on top of the gantry crane.
[0037] The first robotic arm is hinged at one end to the gantry crane operating compartment and can rotate and pitch relative to the gantry crane operating compartment.
[0038] The second robotic arm is hinged at one end to the other end of the first robotic arm and can perform pitching motion relative to the first robotic arm.
[0039] The grab bucket is connected to the other end of the second robotic arm via a steel cable;
[0040] The second V2X communication module is located inside the gantry crane control compartment;
[0041] The hatch positioning terminal is set at a predetermined position on the edge of the ship's hatch and includes a GNSS differential positioning receiver.
[0042] The hatch relay communication unit (RSU) is set at a preset position in the hatch and establishes wireless communication connections with the first V2X communication module and the second V2X communication module respectively.
[0043] The cleaning machine obtains its position and orientation information relative to the hatch positioning terminal through LiDAR scanning, the gantry crane obtains the position information of the grab bucket through the joint angle sensors of the first and second robotic arms, and the hatch relay communication unit (RSU) forwards the position information and control commands between the cleaning machine and the gantry crane in real time, so as to realize the collaborative operation of the cleaning machine and the gantry crane.
[0044] Coordinate system establishment process as follows: Figure 2 As shown:
[0045] The positioning terminal is fixed in a fixed position at the hatch, such as the lower right corner. It obtains its accurate current coordinates through satellite navigation differential positioning.
[0046] The gantry crane is a fixed device, and its accurate coordinates have been pre-calibrated.
[0047] The clearing machine scans the warehouse environment and creates a map to obtain the relative position of the warehouse entrance positioning terminal.
[0048] A local coordinate system is established by combining the positions of the gantry crane, the hatch positioning terminal, and the pose of the clearing machine relative to the hatch positioning terminal. Plocal is the local coordinate system, R is the rotation matrix, O is the origin of the local coordinate system, and Pworld is the world coordinate system. ;
[0049] Clearance machine workflow;
[0050] The cleaning machine in this application is an automated cleaning machine, capable of automatically planning paths to accumulate materials from the hold side towards the hatch opening. For example... Figure 3 As shown in the top view, the shaded areas are the parts that the gantry crane grab cannot grab. The automatic clearing machine mainly piles the material in the shaded (dark and light shaded) areas towards the middle so that the gantry crane grab can grab the material.
[0051] The automatic clearing machine plans the clearing path along the edge in a U-shape, such as... Figure 4 As shown, the material at the edge is gradually transported to the position where the gantry crane can grab it.
[0052] V2X-based clearance machine obstacle avoidance workflow:
[0053] After completing the path planning, the clearing machine performs its operations according to the path.
[0054] The gantry crane calculates the current robotic arm 1 ( Figure 1 Component 5) and gantry crane operating compartment ( Figure 1 The angles of component 7), robotic arm 1 and robotic arm 2 are used to calculate the real-time position of the grab in the local coordinate system.
[0055] The position of the grab bucket is transmitted to the cleaning machine in real time via the hatch relay communication unit (RSU).
[0056] The clearing machine obtains the real-time position of the grab bucket, designates this position as a no-passage area, sets virtual obstacles, and corrects the locally planned path to avoid the grab bucket's grabbing position. This ensures the continuity of the clearing machine's operation.
[0057] V2X-based gantry crane grab workflow:
[0058] The clearing machine senses the warehouse environment in real time and processes the warehouse interior into a grid.
[0059] The material accumulation height within the grid is transmitted to the hatch relay communication unit (RSU), which then transmits this data to the gantry crane operating compartment.
[0060] The gantry crane's control bay dynamically displays the information sensed by the clearing machine, enabling operators to prioritize grabbing materials from suitable areas. This information can also serve as a basis for decision-making in subsequent unmanned automated operations of the gantry crane.
[0061] V2X-based gantry crane emergency stop workflow:
[0062] The clearance machine obtains its current position in the ship's cabin in real time based on its sensors, and predicts its future location based on the planned route.
[0063] The cleaning machine transmits the location data to the hatch repeater RSU, which then transmits the data to the gantry crane control room.
[0064] The gantry crane calculates the current robotic arm 1 ( Figure 1 Component 5) and gantry crane operating compartment ( Figure 1 The angles of component 7), robotic arm 1 and robotic arm 2 are used to calculate the real-time position of the grab in the local coordinate system.
[0065] The gantry crane control bay calculates the collision risk between the grab bucket and the clearing machine in real time.
[0066] If there is a risk of collision, immediately stop lowering the grab bucket. Wait for the clearing machine to move away from the collision risk area before continuing operations.
[0067] Example 2
[0068] like Figure 5 As shown, a method for clearing a warehouse in conjunction with a gantry crane includes: acquiring positioning data, which includes GNSS differential positioning coordinates of the hatch positioning terminal, pre-calibrated coordinates of the gantry crane, and the relative pose of the clearing machine obtained by lidar SLAM mapping; establishing a local coordinate system with the hatch positioning terminal as the origin through coordinate transformation; acquiring a three-dimensional point cloud of materials inside the hatch and generating the operating path of the clearing machine; wherein the operating path is a spiral-shaped path that pushes materials from the hatch side to the center area of the hatch; collecting joint angle data of the first and second robotic arms of the gantry crane and calculating the real-time position of the gantry crane grab bucket based on the joint angle data; wherein the first robotic arm is the main arm connecting the gantry crane operating compartment and the second robotic arm, and the second robotic arm is the auxiliary arm connecting the first robotic arm and the grab bucket cable; setting the real-time position of the gantry crane grab bucket as a high-cost area of the local cost map and correcting the operating path of the clearing machine; obtaining the minimum distance between the corrected operating path of the clearing machine and the real-time position of the gantry crane grab bucket through a collision detection algorithm, and locking the grab bucket lifting mechanism for emergency stop control when the minimum distance is less than a preset safety threshold.
[0069] In S1, a local coordinate system with the hatch positioning terminal as the origin is established through coordinate transformation, including: receiving GNSS differential positioning data output by the hatch positioning terminal and extracting longitude. ,latitude and elevation Transformed into planar coordinates using Gaussian projection Set this coordinate as the origin of the local coordinate system O(0, 0, 0); read the pre-calibrated world coordinates stored in the gantry crane control system. Calculate the translation vector Through the first coordinate transformation matrix Transform the gantry crane coordinates to the local coordinate system to obtain ;in, It is a 3×3 identity matrix, and T is a 3×1 translation vector; this conversion provides the gantry crane reference point for calculating the grab position in S3, so that the grab position calculated by the joint angle can be unified to the local coordinate system.
[0070] Receive pose data output from the LiDAR SLAM module of the clearance machine, including relative position vectors. Euler angle posture Construct the second coordinate transformation matrix: ;in, , , These are rotation matrices about the X, Y, and Z axes, respectively; The transformation is a translation vector; this transformation enables the clearing machine to accurately position itself in the local coordinate system, providing a starting position for the path planning of S2 and a real-time position update reference for the dynamic obstacle avoidance of S4.
[0071] Obtain the ship's heading angle θ output from the electronic compass of the hatch positioning terminal, and construct the local coordinate system alignment rotation matrix: ;
[0072] For any world coordinate point Perform coordinate transformation: The unified coordinate system enables obstacle avoidance calculations for the grab bucket position and the clearing machine path in S4 to be performed in the same reference system, providing an accurate spatial positional relationship for collision detection in S5.
[0073] In particular, the berthing position of a ship changes with the tides, and the ship's cabin coordinate system is dynamic; the gantry crane is a shore-based fixed device and uses the port's fixed coordinate system; the cleaning machine moves and operates inside the cabin and uses its own relative coordinate system; the inconsistency of the three coordinate systems makes it impossible to accurately position and coordinate, which can easily lead to blind spots or repetitive operations.
[0074] The hatch positioning terminal is fixed to the edge of the ship's hatch, making it an ideal reference point for connecting dynamic vessels with fixed shore-based equipment. For cleaning machines, the hatch is the necessary passage for entering and exiting the ship's hold, and the LiDAR SLAM mapping will inevitably scan this location, forming a natural positioning reference benchmark. For gantry cranes, the hatch is precisely the center of the target area for grab operations, and all grabbing actions revolve around this point. More importantly, the hatch positioning terminal moves with the ship's tidal motion, automatically compensating for dynamic changes in the ship's position and fundamentally solving the problem of inconsistent coordinate systems.
[0075] The coordinates of the gantry crane are obtained after transformation It is directly used to calculate the real-time position of the grab bucket, ensuring that the cleaning machine can accurately identify the areas that need to be avoided; the cleaning machine obtains not only its own position through the T2 transformation matrix, but more importantly, its attitude information. This is crucial for executing a spiral feeding path—only by accurately knowing its own orientation can the cleaning machine correctly push the material to the center area of the hatch; simultaneously... The rotation matrix aligns the X-axis of the local coordinate system with the longitudinal direction of the ship's hold, matching the zigzag operation mode of the cleaning machine, which greatly simplifies the computational complexity of path planning.
[0076] S2, acquire the 3D point cloud of materials inside the chamber and generate the operation path of the cleaning machine, including: scanning the chamber environment with the lidar mounted on the cleaning machine to acquire 3D point cloud data of the material surface based on the local coordinate system. ; for 3D point cloud data The compartment is divided into m x n grid cells, and the average material height within each grid cell (i, j) is calculated. and material distribution density ; Material height Greater than the preset height threshold And the distance from the bulkhead is less than the preset distance. The grid marks are designated as areas to be cleaned; based on the distribution of areas to be cleaned, a spiral-shaped working path is generated, including: determining the outline of the inner boundary of the compartment and generating the outermost rectangular path parallel to the compartment wall; setting the effective pushing width w of the cleaning machine bucket; starting from the outermost path, after each round of pushing is completed, the inner path is shifted inward by a distance w.
[0077] Specifically, in port bulk cargo clearing operations, gantry crane grabs, limited by their mechanical structure, can only cover 60%–70% of the central area of the hold, leaving 30%–40% of the material at the edges as blind spots. While the original 3D point cloud data is highly accurate, its massive volume and discrete distribution make it difficult to directly use for operational decisions. This application addresses this by dividing the hold area into m×n grid cells and calculating the average material height in each grid cell. and distribution density The system can quickly identify which areas need to be cleaned—that is, those areas whose height exceeds the threshold. and the distance from the bulkhead is less than The grid-like representation not only facilitates path planning for the clearing machine, but more importantly, it enables efficient transmission to the gantry crane control bay via V2X communication, allowing operators to intuitively understand the material distribution within the bay and providing a basis for grabbing decisions.
[0078] Furthermore, traditional reciprocating or random cleaning paths can cause disordered material accumulation, which actually increases the difficulty of the gantry crane's gripping. The spiral path starts from the outermost circle and advances parallel to the bulkhead. After each circle, it shifts inward by the effective width w of the bucket, forming a continuous push from the outside in.
[0079] Finally, the gridded data and the spiral path form a complete information loop. While the cleaning machine performs its spiral push, it updates the material height information of each grid in real time. This information is synchronized to the gantry crane via V2X communication, enabling the gantry crane operator to accurately determine which area has sufficient material accumulation and prioritize its grabbing. This grid-based dynamic information sharing mechanism tightly connects the previously independent cleaning machine and gantry crane, achieving true collaborative operation.
[0080] S3 collects joint angle data of the first and second robotic arms of the gantry crane, and calculates the real-time position of the gantry crane's grab bucket based on the joint angle data. This includes: real-time acquisition of the rotation angle of the first robotic arm relative to the gantry crane's operating compartment via a rotary encoder installed on the first robotic arm's rotation axis. and pitch angle The pitch angle of the second robotic arm relative to the first robotic arm is collected in real time by an angle sensor mounted on the connecting shaft of the second robotic arm. The release length L of the grab bucket cable is collected in real time by a cable length encoder installed on the grab bucket lifting mechanism; a kinematic model of the gantry crane is established based on the DH parameter method, according to the position of the gantry crane in the local coordinate system. Using this as a reference point, calculate the real-time position of the grab bucket in the local coordinate system: ;in, The length of the first robotic arm. The length of the second robotic arm. Here is the rotation matrix of the first robotic arm. This is the rotation matrix for the second robotic arm;
[0081] S4, set the real-time position of the gantry crane grab to a high-cost area on the local cost map, and correct the operation path of the clearing machine, including: constructing a local cost map based on grid cells and initializing all grid cost values to 0; receiving the real-time position of the gantry crane grab through the hatch relay communication unit. Real-time position of the gantry crane grab bucket Mapped to corresponding raster coordinates using raster coordinates Centered on, with radius All grid cells within the range are set to high generation value. Among them, high-value The corresponding area is a restricted area; retrieve the current location of the clearance machine. Heading angle and speed ;
[0082] Within the dynamic constraints of the clearing machine, multiple predicted trajectories are generated, including those based on the machine's maximum linear velocity. Maximum angular velocity and maximum linear acceleration Maximum angular acceleration Determine the boundaries of the speed window; based on the current speed of the clearing machine. and angular velocity Calculate the achievable velocity range within the time interval Δt: ; .
[0083] N sets of velocity combinations are uniformly sampled within the achievable velocity range. , where i = 1, 2, ..., N; for each velocity combination According to the kinematic model of the clearance machine, from the current position and heading angle To begin, generate a sequence of trajectory points for a predicted duration T: ; ; Each trajectory is discretized into M trajectory points to form a set of predicted trajectories for subsequent evaluation.
[0084] The time interval Δt and the prediction duration T are determined based on the control cycle of the clearing machine and the local planning requirements.
[0085] Calculate the minimum distance between each predicted trajectory and the high-cost area, and eliminate trajectories that enter the high-cost area; select the trajectory closest to the operation path as the correction path; convert the correction path into a clearing machine control command to avoid the gantry crane grab bucket in real time.
[0086] In particular, in traditional port clearance operations, when the clearance machine encounters the gantry crane grab bucket being lowered, it usually adopts a simple stop and wait strategy. Although this method is safe, it seriously affects the efficiency of operation. Frequent start-stop operations will also increase equipment wear and energy consumption.
[0087] This application constructs a dynamic cost map to display the real-time location of the gantry crane grab bucket. Map to a raster coordinate system and set the radius around it. In high-cost regions, the avoidance problem in three-dimensional space is simplified to a path planning problem in two-dimensional plane. The setting does not absolutely prohibit passage, but rather incurs a significant penalty in the path optimization algorithm, causing the cleaning machine to prioritize detours. This flexible avoidance strategy ensures that the cleaning machine can dynamically adjust its path to avoid the grab bucket's operating area while maintaining operational continuity.
[0088] Furthermore, as heavy-duty construction machinery, warehouse clearing machines possess significant inertia and limited maneuverability, preventing them from making sharp turns or stopping instantaneously like smaller robots. This is addressed by setting a maximum linear speed. Maximum angular velocity And corresponding acceleration limits, the system within the speed window The system generates physically feasible velocity combinations. This trajectory generation method based on dynamic constraints ensures that each candidate path is actually executable by the clearing machine, avoiding theoretically optimal but physically infeasible path planning results.
[0089] In S5, a collision detection algorithm is used to obtain the minimum distance between the corrected operating path of the cleaning machine and the real-time position of the gantry crane's grab bucket. When the minimum distance is less than a preset safety threshold, the grab bucket lifting mechanism is locked for emergency stop control. This includes: simplifying the cleaning machine into a rectangular box with dimensions of length... ,Width ,high The center of the enclosure is located at the current position of the clearance machine. Based on the physical dimensions of the gantry crane grab bucket, the grab bucket is simplified to a cylinder with a radius of [missing information]. The height is The center of the cylinder is located at the real-time position of the grab bucket. ; Obtain the M predicted location points on the generated clearing machine correction path within the next T time period;
[0090] For each predicted location point, calculate the minimum distance between the clearing machine enclosure and the grab bucket cylinder: Project the clearing machine enclosure onto the horizontal plane to obtain a rectangular area; project the grab bucket cylinder onto the horizontal plane to obtain a circular area; calculate the minimum distance between the rectangle and the circle. ;when At that time, among them, To preset a safety threshold, an emergency stop signal is sent to the gantry crane via the hatch relay communication unit; upon receiving the emergency stop signal, the gantry crane immediately locks the winch motor of the grab bucket lifting mechanism, stops the release of the steel cable, and maintains the current height of the grab bucket; the distance between the cleaning machine and the grab bucket is continuously monitored, and when... At that time, among them, To restore the buffer distance, the emergency stop status is lifted, and normal operations can resume.
[0091] In particular, in actual operation, although the S4 provides intelligent obstacle avoidance for the clearing machine, there are still several risk scenarios where obstacle avoidance may fail: the clearing machine may be unable to avoid obstacles in time due to sensor malfunction, communication delays, or complex terrain limitations; the operator may also fail to detect the abnormal approach of the clearing machine in time due to blind spots or distraction. In such cases, relying solely on a one-way obstacle avoidance mechanism will lead to serious safety accidents.
[0092] Furthermore, a collision risk assessment is conducted based on M predicted locations of the clearing machine within a future timeframe T. This predictive protection considers two key factors: first, the clearing machine's inertia—even with immediate braking, the machine will continue to slide for a distance; and second, the response delay of the gantry crane's grab—a certain amount of time is required from issuing an emergency stop signal to the hoist motor fully locking. By predicting and assessing in advance, the system can trigger protective actions before a real collision risk materializes, ensuring sufficient safety margin.
[0093] The foregoing illustrative description of the present application and its embodiments is not restrictive and can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application. The accompanying drawings are only one embodiment of the present application, and the actual structure is not limited thereto. Therefore, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the present application, such designs should fall within the scope of protection of this application. Furthermore, the word "comprising" does not exclude other elements or steps, and the word "a" preceding an element does not exclude the inclusion of "a plurality" of that element. Terms such as "first," "second," etc., are used to indicate names and do not indicate any specific order.
Claims
1. A method of operating a stock unloading machine in conjunction with a door machine, the method comprising: The method comprises the following steps: Obtain positioning data, including GNSS differential positioning coordinates of the hatch positioning terminal, pre-calibration coordinates of the gantry, and relative poses obtained by the laser radar SLAM mapping of the stocker; establish a local coordinate system with the hatch positioning terminal as the origin through coordinate transformation; Obtain the three-dimensional point cloud of the materials in the cabin, and generate the working path of the stocker; wherein the working path is a spiral-shaped and back-shaped path for pushing the materials at the cabin edge to the center area of the hatch; Collect the joint angle data of the first mechanical arm and the second mechanical arm of the gantry, and calculate the real-time position of the gantry grab bucket according to the joint angle data; wherein the first mechanical arm is the main arm connecting the operation cabin of the gantry and the second mechanical arm, and the second mechanical arm is the auxiliary arm connecting the first mechanical arm and the grab bucket cable; Set the real-time position of the gantry grab bucket as a high-cost area of the local cost map, and correct the working path of the stocker; Obtain the minimum distance between the corrected working path of the stocker and the real-time position of the gantry grab bucket through a collision detection algorithm, and when the minimum distance is less than a preset safety threshold, lock the grab bucket lifting mechanism for emergency stop control.
2. The stocker and gantry linkage stocker cleaning method according to claim 1, wherein: A local coordinate system with the hatch positioning terminal as the origin is established through coordinate transformation, comprising: GNSS differential positioning coordinates of the hatch positioning terminal is set as the origin O of the local coordinate system; Pre-calibrated world coordinates of the door machine by the first coordinate transformation matrix Transforming to the local coordinate system, obtaining the position of the door machine in the local coordinate system ; a position vector relative to the hatch positioning terminal obtained according to a laser radar SLAM mapping of the clearance machine and an attitude angle , a second coordinate transformation matrix is constructed to convert the clearance machine own coordinate system to the local coordinate system; Mapping any point in the world coordinate system to the local coordinate system through a rotation matrix R: wherein, represents the coordinates of the point in the local coordinate system; represents the coordinates of the point in the world coordinate system; R is a 3x3 rotation matrix determined by the heading angle of the hatch positioning terminal, such that the X axis of the local coordinate system points in the longitudinal direction of the ship hatch, the Y axis points in the transverse direction of the ship hatch, and the Z axis is vertically upward.
3. The stocker and gantry linkage stocker cleaning method according to claim 2, wherein: Obtaining the three-dimensional point cloud of the materials in the cabin to generate the working path of the stocker, comprising: The laser radar scans the cabin environment, and obtains three-dimensional point cloud data of a material surface based on a local coordinate system ; Griding the three-dimensional point cloud data Griding the three-dimensional point cloud data and the material distribution density ; Material height Greater than the preset height threshold And the distance from the bulkhead is less than the preset distance. The grid markers are designated as areas to be cleaned; Based on the distribution of the area to be cleaned, a spiral-shaped and back-shaped working path is generated.
4. The stocker and gantry linkage stocker cleaning method according to claim 3, wherein: Based on the distribution of the area to be cleaned, a spiral-shaped and back-shaped working path is generated, comprising: Determine the boundary contour in the cabin to generate the outermost circle rectangular path parallel to the cabin wall; Set the effective pushing width w of the stocker bucket; Starting from the outermost circle path, after completing a circle of pushing, offset inward by a distance w to generate an inner circle path.
5. The stocker and gantry linkage stocker cleaning method according to claim 3, wherein: Calculating the real-time position of the gantry grab bucket according to the joint angle data, comprising: Collecting a rotation angle of the first robot arm relative to the door machine operating bin and a pitch angle ; collecting a pitch angle of the second robot arm relative to the first robot arm ; Collecting the release length L of the grab bucket cable; Position of the gantry in the local coordinate system As a reference point, the real-time position of the gantry grab in the local coordinate system is calculated .
6. The stocker and gantry linkage stocker cleaning method according to claim 5, wherein: Correcting the working path of the stocker, comprising: Constructing a local cost map according to the grid unit, and initializing all grid cost values to 0; Receiving the real-time position of the grab bucket of the portal crane through the hatchway relay communication unit ; Real-time position of door machine grab mapped to corresponding grid coordinates , grid coordinates centered, all grids within a radius are set to high generation value ; wherein, high generation value table corresponding area is forbidden area Acquire current position of the destocker , heading angle and speed ; Generating multiple predicted trajectories within the kinematic constraints of the stocker; Calculating the minimum distance between each predicted trajectory and the high-cost area, and eliminating trajectories that enter the high-cost area; Selecting the trajectory closest to the working path as the corrected path; Convert the corrected path into stocker control instructions to avoid the gantry grab bucket in real time.
7. The stocker and gantry linkage stocker cleaning method according to claim 6, wherein: Generating multiple predicted trajectories, comprising: determining boundaries of a velocity window in dependence on a maximum linear velocity , a maximum angular velocity , and a maximum linear acceleration , a maximum angular acceleration of the destacker based on the current speed of the destocker and angular velocity a range of speeds reachable in a time interval Δt is calculated; Uniformly sampling N groups of speed combinations in a reachable speed range where i = 1, 2,..., N; for each combination of speed , from the current position and heading angle , generate a sequence of trajectory points for a prediction horizon T, according to the sweeper kinematic model; Discretize each trajectory into M trajectory points to form a predicted trajectory.
8. The stocker and gantry linkage stocker cleaning method according to claim 7, wherein: The minimum distance between the corrected operation path of the cleaning machine and the real-time position of the grab bucket of the portal crane is obtained by a collision detection algorithm, including: A cuboid bounding box is set according to the size of the cleaning machine; A cylindrical bounding box is set according to the size of the grab bucket of the portal crane; M predicted position points in T time in the future on the corrected path of the cleaning machine are obtained; The minimum distance between the cuboid bounding box and the cylindrical bounding box is calculated for each predicted position point; When the minimum distance is less than a preset safety threshold, an emergency stop signal is sent to the portal crane through the hatch relay communication unit.
9. A stock unloading device in which a stock unloader is linked to a gate, characterized by It comprises: The cleaning machine is arranged in the cabin and comprises: The laser radar is arranged at the top of the cleaning machine and scans the environment in the cabin to obtain three-dimensional point cloud data; The positioning module calculates the pose of the cleaning machine relative to the hatch positioning terminal according to the SLAM mapping data of the laser radar; The path planning module generates an operation path according to the three-dimensional point cloud data and the pose; The first V2X communication module sends the position of the cleaning machine and receives the position of the grab bucket of the portal crane; The portal crane is arranged on the wharf outside the cabin and comprises: The portal crane operating bin is arranged at the top of the portal crane; The first mechanical arm is hingedly connected to the portal crane operating bin at one end and can rotate and pitch relative to the portal crane operating bin; The second mechanical arm is hingedly connected to the other end of the first mechanical arm at one end and can pitch relative to the first mechanical arm; The grab bucket is connected to the other end of the second mechanical arm by a steel cable; The second V2X communication module is arranged in the portal crane operating bin; The hatch positioning terminal is arranged at a predetermined position on the edge of the hatch and comprises a GNSS differential positioning receiver; The hatch relay communication unit RSU is arranged at a predetermined position on the hatch and is wirelessly connected to the first V2X communication module and the second V2X communication module; The cleaning machine obtains the pose information relative to the hatch positioning terminal by scanning with the laser radar, the portal crane obtains the grab bucket position information by the joint angle sensors of the first mechanical arm and the second mechanical arm, and the hatch relay communication unit RSU forwards the position information and control instructions between the cleaning machine and the portal crane in real time, so as to realize the cooperative operation of the cleaning machine and the portal crane.