Crane global anti-collision system based on real-time digital twinning and dynamic bounding box

Abstract

A kind of crane global anti-collision system based on real-time digital twinning and dynamic bounding box, the system includes the three-dimensional position model of crane and obstacle according to real-time pose data by three-dimensional modeling module, by high-fidelity three-dimensional digital twinning model constructed based on equipment geometric parameters;Through regional hierarchical structure for confirming the execution state area of crane and obstacles in the surrounding environment according to instantaneous nearest distance;And through the anti-collision decision module, the set anti-collision operation is executed, through warning zone→braking zone;After entering warning zone, first remind / slow down, then rigid brake, ensure to untangle quickly and reduce human operation risk;After entering braking zone, only allow reverse operation to form gradual safety control, the present scheme in high load rhythm is prone to excessive downtime or blind area of protection, which solves the problems of affecting production efficiency and increasing safety risk.

Description

Technical Field

[0001] This invention relates to the field of safety technology for lifting equipment in ports, shipbuilding, and other industries, and particularly to a crane global collision avoidance system based on real-time digital twins and dynamic bounding boxes. Background Technology

[0002] In shipbuilding and ship repair dock operations, large lifting equipment such as gantry cranes, shipbuilding gantry cranes, span gantry cranes, and parallel-rail gantry cranes undertake key processes such as steel structure segment transfer, assembly, and positioning. To ensure production cycle and assembly accuracy, continuous safety protection and linkage control of the lifting equipment are required to ensure stable operation even under complex marine engineering environments (sea fog, backlight, salt spray corrosion).

[0003] Currently, commonly used protection methods mainly include: fixed protection zones based on limit switches / limit switches, proximity alarms based on single-point sensors, and collaborative methods mainly based on driver observation and intercom dispatch. These methods generally suffer from the following problems: First, the static nature of the protection boundary and its conservatism make it difficult to meet cycle time requirements; second, environmental factors such as sea fog and backlighting can easily lead to false alarms or sensor failures; third, in scenarios where multiple cranes are operating on the same track or in close proximity, there is a lack of a unified global judgment and hierarchical handling mechanism, making it difficult to promptly arbitrate early warning and braking strategies; and fourth, there is a lack of event recording and traceability capabilities for engineering management.

[0004] Due to the aforementioned factors, existing solutions are prone to excessive downtime or blind spots under high-load cycles, affecting both production efficiency and increasing safety risks. Therefore, there is an urgent need for a comprehensive protection technology for various types of lifting equipment in shipbuilding docks. This technology should be able to achieve reliable early warning and braking control through unified spatial modeling and strategy arbitration without relying on complex speed estimation, while also considering real-time performance and feasibility in engineering environments. Summary of the Invention

[0005] This invention provides a crane global collision avoidance system based on real-time digital twin and dynamic bounding box to overcome the above-mentioned technical problems.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows:

[0007] A crane global collision avoidance system based on real-time digital twin and dynamic bounding box includes a 3D modeling module, a dynamic acquisition module for oriented bounding boxes, a nearest distance acquisition module, a region hierarchical construction module, and a collision avoidance decision module.

[0008] The 3D modeling module is used to acquire real-time pose data of the crane and obstacles in the surrounding environment after pre-installation and calibration under a preset coordinate system. The preset coordinate system includes a preset global coordinate system of the plant area and a local coordinate system about the crane's track, traveling mechanism and rotation center. At the same time, based on the 3D modeling software, the module acquires the corresponding digital twin 3D position models of the crane and obstacles according to the real-time pose data. The digital twin 3D position models include real-time 3D models of each component structure in the crane and obstacles.

[0009] The dynamic acquisition module is used to configure dynamic body orientation bounding boxes for each component structure according to the real-time 3D model, and to take the union of the body orientation bounding boxes of the corresponding crane as the anti-collision space; and to take the union of the body orientation bounding boxes of the corresponding obstacle as the obstacle space.

[0010] The closest distance acquisition module is used to acquire the instantaneous closest distance between the body-oriented bounding box in the collision avoidance space and the body-oriented bounding box in the obstacle space;

[0011] The regional hierarchical structure is used to determine the execution state region type of the crane and obstacles in the surrounding environment based on the instantaneous closest distance; and the execution state region type includes warning zone, braking zone and normal zone.

[0012] The collision avoidance decision module is used to execute the corresponding collision avoidance operation based on the type of execution state area where the crane is located.

[0013] Furthermore, the method for obtaining the instantaneous closest distance based on the anti-collision space and the obstacle space specifically includes the following steps:

[0014] S10: Get the current crane object With obstacle objects The collision avoidance space and obstacle space are respectively denoted as , , ;in Indicates the corresponding crane object The Middle The body-oriented bounding box of each substructure; Represents the corresponding obstacle object The Middle The body-oriented bounding box of each substructure;

[0015] S11: Obtain any pair of bounding boxes in the collision avoidance space and obstacle space, denoted as an orientation bounding box pair. , The separation axis theorem is used to determine whether the oriented bounding box pairs intersect.

[0016] If an intersection is confirmed, the distance between the current directed bounding box pairs is defined as 0; if no intersection is confirmed, a set of candidate bounding box pairs is obtained; and for all candidate directed bounding box pairs in the candidate bounding box pair set, the geometric minimum Euclidean distance is calculated. Simultaneously, based on the geometric minimum Euclidean distance... The expression for obtaining the instantaneous nearest distance at different times is:

[0017] ,

[0018] In the formula: express The instantaneous closest distance corresponds to the moment.

[0019] Furthermore, the specific method for obtaining the directional bounding boxes of each body of the crane is as follows:

[0020] S20: Define device base reference point and ;in This indicates the equipment base reference point in the pre-set global coordinate system of the plant area. The x-coordinate, y-coordinate, and vertical coordinate;

[0021] Define the key areas of the crane, including the end of the slewing boom, the tail end of the counterweight, the structural feature surfaces of the gantry crane's rigid / flexible legs, the trolley and lifting gear, and the track end area; and confirm the crane's real-time PLC inputs based on a preset PLC control system, wherein the real-time PLC inputs include at least the trolley travel at the current moment. Car journey Lifting height Portal corner and the amplitude of the gate seat ;

[0022] S21: Confirm the type of the current crane. If the crane type is confirmed to be a gantry crane, continue to S22; if the crane type is confirmed to be a portal crane, continue to S23.

[0023] S22: Defines a parameter matrix used to characterize the orientation of the body unit in a gantry crane;

[0024] Furthermore, the gantry crane body unit includes at least a trolley, a crane arm, and a hook;

[0025] And based on the parameter matrix, according to the gantry crane's trolley travel... Car journey and lifting height Combined with equipment base reference point Acquire real-time position data of the gantry crane, including the real-time position of the trolley, the real-time position of the crane arm, and the real-time position of the hook.

[0026] Real-time location of large vehicle The formula for obtaining it is:

[0027] Real-time location of the car The formula for obtaining it is:

[0028] ;

[0029] real-time position of hook The formula for obtaining it is:

[0030] ;

[0031] in: These represent the parameter matrices used to characterize the orientation of the trolley, the gantry, and the hook, respectively.

[0032] Based on the given position offset in the local coordinate system To obtain corrected real-time location data;

[0033] The correction formula for the real-time location data is: ,in Indicates the correction position; Indicates the location of the gantry crane body unit and = ; Represents the parameter matrix and = ;

[0034] Based on the bounding box algorithm, the directional body bounding box of the gantry crane body unit that is currently located and surrounds the key area is obtained by combining the corrected real-time position data of the gantry crane with the key area.

[0035] S23: Define the parameter matrix used to characterize the orientation of the body unit in a gantry crane;

[0036] Furthermore, the main body unit of the gantry crane includes at least a trolley, a slewing platform, a boom root hinge point, a slewing boom, and a hook position;

[0037] And based on the parameter matrix, according to the trolley travel of the gantry crane Lifting height Portal corner and the amplitude of the gate seat To obtain real-time position data of the gantry crane, including the real-time position of the trolley, the real-time position of the slewing platform, the real-time position of the boom root hinge point, the attitude data of the slewing boom, and the real-time position of the hook.

[0038] Real-time location of large vehicle The formula for obtaining it is:

[0039] Real-time position of slewing platform The formula for obtaining it is:

[0040] ; ;

[0041] ,

[0042] Real-time position of arm root hinge point The formula for obtaining it is:

[0043] , ;

[0044] Slewing boom attitude data The formula for obtaining it is:

[0045] ; ,

[0046] real-time position of hook The formula for obtaining it is:

[0047] ;

[0048] in, These represent the parameter matrices used to characterize the orientation of the main vehicle and the orientation of the slewing platform, respectively. Indicates the angle of rotation of the gate seat. rotation matrix; This indicates a fixed offset from the given rotary platform to the boom root; Indicates the amplitude of the gate seat. rotation matrix; Indicates arm length; This indicates the height of the hook on the z-axis;

[0049] Based on the given position offset in the local coordinate system To obtain corrected real-time location data;

[0050] The correction formula for the real-time location data is:

[0051] ;

[0052] ,

[0053] in Indicates the correction position; Indicates the location of the gantry crane body unit and = ; Represents the parameter matrix and = ;

[0054] Based on the bounding box algorithm, the directional body bounding box of the body unit in the gantry crane that is currently located and is surrounded by the key area is obtained by combining the real-time position data of the gantry crane with the key area.

[0055] S24: Based on the definition of the oriented ontology bounding box, obtain the set of ontology bounding boxes;

[0056] And the expression for the set of the bounding boxes of the ontology is:

[0057] ,

[0058] In the formula: Represents the body unit; Represents the set of body units and for gantry cranes For gantry cranes ; Indicates the real-time location of the main body unit; This represents the parameter matrix used to characterize the current orientation of the body element; Indicates the half-side length of the oriented bounding box OBB and Indicates the length of the bounding box OBB; This indicates the width of the bounding box OBB; This indicates the height of the bounding box OBB.

[0059] Furthermore, it also includes a body bounding box optimization module;

[0060] The enclosure box optimization module is used to optimize the enclosure box corresponding to the slewing boom in the gantry crane body unit. The method is as follows:

[0061] S100: The single body enclosure corresponding to the slewing boom is randomly divided into two body enclosures according to a preset direction, and the union of the two body enclosures is used as the protective enclosure of the slewing boom. ;

[0062] The protective enclosure The distances from each side of the boom to the straight line containing the outer surface or outer edge of the boom are equal and are defined as follows: ;

[0063] S101: In the preset local coordinate system of the slewing boom, the outer corners of the single body bounding box far from the slewing boom's corresponding slewing center are chamfered / trimmed to obtain a chamfered bounding box; and the chamfered / trimmed position is the intersection of two body bounding boxes, specifically:

[0064] S1011: Defines the chamfer length to be chamfered at the chamfer / bevel position. With the width of the chamfer And set the chamfered bounding box hypotenuse length constraint as follows:

[0065] ,

[0066] ,

[0067] In the formula: Indicates the target value of the beveled bounding box hypotenuse length. ,and ; Indicates a fixed safety margin and ;

[0068] S1012: Solve and obtain the length of the chamfer to be chamfered based on the constraint of the hypotenuse length of the chamfered bounding box. With the width of the chamfer for:

[0069] ,

[0070] ,

[0071] S1013: Obtain the chamfered bounding box after chamfering / bevel processing according to S1012;

[0072] S102: Achieving the protective enclosure based on the chamfered enclosure box To optimize, its expression is:

[0073] ,

[0074] In the formula: Indicates the number of partitions in a single body's bounding box; Indicates the first A beveled surround box.

[0075] Furthermore, the method for determining the execution state region of the crane and obstacles in the surrounding environment based on the instantaneous nearest distance is as follows:

[0076] S31: Obtain the gear parameters of each body unit in the corresponding crane;

[0077] The gear parameters include, but are not limited to, the luffing gear value, slewing gear value, and trolley forward and backward travel gear value of the crane.

[0078] Based on the gear position parameters, the 3D position and orientation of the current body unit are determined, and the position and orientation information of multiple adjacent / following body units that are adjacent to or have a following relationship with the current body unit are obtained. Based on the motion direction, the directional bounding boxes of the corresponding adjacent / following body units are obtained according to the position and orientation information, and the union of these directional bounding boxes is defined as the potential collision avoidance direction set and denoted as . ;

[0079] S32: For any potential collision avoidance direction and Based on expert experience, two types of directional bounding boxes are set for the current ontology unit, including the early warning bounding box. and in Internal brake enclosure For all directions of the same body unit The union of these two sets is denoted as the warning boundary. ,all The union of the two sets is denoted as the braking boundary. And satisfy ;

[0080] S33: Based on a given fixed safety margin The equivalent distance used for early warning judgment is obtained based on the instantaneous nearest distance. The formula is:

[0081] ,

[0082] S34: Based on equivalent distance Formulas are used to obtain the collision avoidance space and warning boundary, respectively. Braking boundary The equivalent distance between them is used to confirm the intersection relationship, and the execution state area of ​​the crane and the obstacles in the surrounding environment is determined based on the intersection relationship; and the execution state area includes the warning area, the braking area and the normal area;

[0083] The method for confirming the intersection relationship is as follows:

[0084] Get the crane object One or more sub-organizations When the gear parameter value is not zero, its collision avoidance direction The corresponding warning boundary With braking boundary ;

[0085] Get another crane or obstacle object The collision avoidance space is based on the equivalent distance. Formula to get objects Collision avoidance space and warning boundary Braking boundary The real-time equivalent distance;

[0086] Objects are identified based on real-time equivalent distance. Collision avoidance space and warning boundary Braking boundary The intersection relationship of objects; if objects Collision avoidance space and warning boundary Intersecting and with braking boundary If they do not intersect, then the execution state area at this time is confirmed as a warning zone; if the objects Collision avoidance space and warning boundary Intersecting and with braking boundary If they intersect, the execution state area is confirmed to be the braking zone; otherwise, the execution state area is confirmed to be the normal zone.

[0087] Furthermore, the geometric minimum Euclidean distance described in S10 This includes: calculating the distance from any edge segment within one of the candidate oriented bounding boxes to an edge segment within the other oriented bounding box, specifically:

[0088] The nearest point parameter used to calculate the distance between edge segments is obtained as follows:

[0089] ,

[0090] ,

[0091]

[0092] ,

[0093] ,

[0094] In the formula: Represents the reference point of the edge segment P in a directional bounding box; The parametric equation point of the edge segment P in the oriented bounding box; Indicate design parameters; This represents the reference point of the edge segment Q in another oriented bounding box; Represents the parametric equation point of the edge segment Q in another oriented bounding box; , , Indicates intermediate parameters; This represents the direction vector of line segment P; This represents the direction vector of line segment Q; , This represents the parameter of the nearest point on the line corresponding to each of the two line segments;

[0095] Calculate the nearest point parameters , The Euclidean distance between them is used as a candidate distance for obtaining the instantaneous closest distance;

[0096] For each candidate bounding box, the distance from any point on one bounding box to the rectangular face of the bounding box in the other bounding box is calculated as follows:

[0097] Define the bounding box parameters for any rectangular face on one of the candidate oriented bounding boxes; and the bounding box parameters include the center of the rectangular face. Unit legal direction orthogonal basis in plane The length and width of half of the rectangle ;

[0098] Based on the bounding box parameters, the distance between any point on another oriented bounding box and any rectangular face on another oriented bounding box is obtained, and this distance is used as a candidate distance for obtaining the instantaneous nearest distance. The expression is as follows:

[0099] ,

[0100] , ,

[0101] ,

[0102] In the formula: Indicates the intermediate vector; Represents two mutually orthogonal unit direction axes within a rectangular plane; This represents the distance from any point on another orientation bounding box to any rectangular face on another orientation bounding box.

[0103] Beneficial Effects: This invention provides a crane global collision avoidance system based on real-time digital twins and dynamic bounding boxes. A 3D modeling module acquires 3D position models of the crane and obstacles based on real-time pose data. A high-fidelity 3D digital twin model, constructed based on equipment geometric parameters, is used to determine collisions in virtual space using dynamic geometric bounding boxes and nearest-neighbor distances. A dynamic acquisition module configures body-oriented bounding boxes for each substructure based on the real-time 3D model, and the union of the body-oriented bounding boxes for the corresponding crane is used as the collision avoidance space; the union of the body-oriented bounding boxes for the corresponding obstacles is used as the obstacle space. By parametrically modeling key mechanisms such as the boom end, counterweight tail, gantry crane rigid / flexible legs, trolley, and hook in the twin model, and according to real-time... The system generates dynamic bounding boxes based on attitude; it solves for the nearest gap of the bounding box using representative points / faces, sets thresholds for each sub-mechanism object of different mechanisms, and thus implements component-level collision avoidance logic at the model level; it constructs a region hierarchy to determine the execution state region of the crane and obstacles in the surrounding environment based on the instantaneous nearest distance; and it executes the set collision avoidance operation through the collision avoidance decision module, from the warning zone to the braking zone; upon entering the warning zone, it first provides a warning / deceleration, and then performs rigid braking to ensure rapid untangling and reduce the risk of human operation; upon entering the braking zone, it only allows reverse operation, forming a progressive safety control. This invention can achieve reliable warning and braking control through unified spatial modeling and strategy arbitration without relying on complex speed estimation, while taking into account real-time performance and feasibility in engineering environments. Attached Figure Description

[0104] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0105] Figure 1 System block diagram for inventing a crane full-domain collision avoidance system based on real-time digital twin and dynamic bounding box;

[0106] Figure 2 This is a schematic diagram of a single body enclosure box in this embodiment;

[0107] Figure 3 This is a schematic diagram of the beveled edge of a single body enclosure box in this embodiment;

[0108] Figure 4 This is a schematic diagram of a single body bounding box optimized by the body bounding box optimization module in this embodiment;

[0109] Figure 5This is a schematic diagram of bounding box intersection detection in this embodiment. Detailed Implementation

[0110] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0111] This embodiment provides a crane global collision avoidance system based on real-time digital twin and dynamic bounding box, such as Figure 1 As shown, it includes a 3D modeling module, a dynamic acquisition module for oriented bounding boxes, a nearest distance acquisition module, a region hierarchical construction module, and a collision avoidance decision module;

[0112] The 3D modeling module is used to acquire real-time pose data of the crane and obstacles in the surrounding environment after pre-installation and calibration in a preset coordinate system. The preset coordinate system includes a preset global coordinate system of the plant area and a local coordinate system about the crane's track, traveling mechanism, and rotation center. Simultaneously, based on 3D modeling software, a digital twin 3D position model of the crane and obstacles is acquired according to the real-time pose data. This digital twin 3D position model includes real-time 3D models of each component structure, i.e., substructure, in the crane and obstacles. In this embodiment, on-site modeling and calibration are completed in one step before deployment: geometric information of the crane on-site is acquired through UAV photogrammetry, photo / video reconstruction, total station / RTK measurement, and / or as-built drawings; registration to the world coordinate system is performed, and geometric constants and installation offsets are extracted; a mapping between PLC counters and geometric quantities (zero position and scaling factor) is established; accuracy acceptance is completed (position preferably at the centimeter level, attitude preferably ≤0.2°), and external parameters, template library, and mapping parameters are solidified into the operating configuration. This embodiment also includes reading the position / pose and related gear values ​​of the PLC in each sampling cycle, and simultaneously collecting the bypass switch status, data quality indicator, communication heartbeat and round-trip delay, and completing the communication health judgment in the same cycle: if the heartbeat is missing or the delay exceeds the threshold, the communication is set to abnormal and switch to conservative control; if it is normal, it enters other preset operations.

[0113] The dynamic acquisition module is used to configure dynamic body-oriented bounding boxes for each substructure model based on the real-time 3D model and the bounding box algorithm, and to use the union of the body-oriented bounding boxes of the corresponding crane as the anti-collision space; and to use the union of the body-oriented bounding boxes of the corresponding obstacle as the obstacle space. In this embodiment, a preset bounding box template is instantiated according to the current position and orientation of each submechanism to obtain the body bounding box set at the current moment. At the same time, the outer contour of the gantry rotating arm is preferably represented by the union of two oriented bounding boxes to achieve a chamfer / bevel. If necessary, the bounding box is expanded with a fixed safety margin to enhance robustness.

[0114] In this embodiment, the bounding box generally refers to a rectangular box that can completely contain a given geometry or set of points. There are two common types: (1) Axis-Aligned Bounding Box (AABB): parallel to the coordinate axes, simple to calculate, but often with significant redundancy; (2) Oriented Bounding Box (OBB): allows rotation and can fit the target more tightly. In this embodiment, each object will have a dynamic oriented bounding box (hereinafter referred to as the ontological oriented bounding box). This embodiment also includes the world coordinate system {W}: Along the direction the vehicle is traveling, Orthogonal to it (the direction of the gantry car). Upwards; simultaneously, it is agreed that the coordinate system of the main gantry crane / gantry crane is parallel to {W}; the coordinate system of the trolley crane is parallel to {W} and orthogonal to the direction of the main gantry crane. In this embodiment, the real-time position / pose data collected by the PLC is the main data source (optionally fused with other sensors). For various target objects in the factory area (including gantry cranes, shipbuilding gantry cranes, span gantry cranes, parallel-track gantry cranes, and fixed ground structures), corresponding geometric protection models are established for their movable mechanisms in the digital twin model. Specifically, based on 3D modeling, 3D models of each target object and its sub-mechanisms are obtained, and a local coordinate system and its translational / rotational degrees of freedom about a given axis are defined for each sub-mechanism. Based on this, one or more body-oriented bounding boxes (OBBs) or equivalent simplified models (such as the "equivalent collision radius" of a circle / cylinder / sphere, hereinafter referred to as body-oriented bounding boxes) are configured for each sub-mechanism. The body-oriented bounding boxes are associated with the pose of the sub-mechanism in real time and their spatial position and attitude are updated according to the sampling period in the plant's world coordinate system. The union of multiple body-oriented bounding boxes constitutes the anti-collision space of the target object, which is used for subsequent collision detection and alarm / braking strategies. The body-oriented bounding boxes will change with the attitude of the mechanism. For example, the boom angle, amplitude, and slewing angle of the gantry will change the space it occupies instantly. The current position of the trolley and the lifting device will change the motion boundary of the gantry crane. Based on the positional quantities such as stroke, slewing, amplitude, and lifting height reported by the PLC, the system locates these key components in the plant's global coordinate system in real time and generates all current body-oriented bounding boxes accordingly.

[0115] Specifically, the method for obtaining the directional bounding boxes of each body of the crane is as follows:

[0116] S20: Define device base reference point and ;in This indicates the equipment base reference point in the pre-set global coordinate system of the plant area. The x-coordinate, y-coordinate, and vertical coordinate;

[0117] Define the key areas of the crane, including the end of the slewing boom, the tail end of the counterweight, the structural feature surfaces of the gantry crane's rigid / flexible legs, the trolley and lifting gear, and the track end area; and confirm the crane's real-time PLC inputs based on a preset PLC control system, wherein the real-time PLC inputs include at least the trolley travel at the current moment. Car journey Lifting height (Positive direction is downward), gate seat rotation angle (around) ) and the portal pitch angle (Local coordinate system around the arm) (Axis, with the lifting arm as the positive axis).

[0118] S21: Confirm the type of the current crane. If the crane type is confirmed to be a gantry crane, continue to S22; if the crane type is confirmed to be a portal crane, continue to S23.

[0119] S22: Defines a parameter matrix used to characterize the orientation of the body unit in a gantry crane;

[0120] Furthermore, the gantry crane body unit includes at least a trolley, a crane arm, and a hook;

[0121] And based on the parameter matrix, according to the gantry crane's trolley travel... Car journey and lifting height Combined with equipment base reference point Acquire real-time position data of the gantry crane, including the real-time position of the trolley, the real-time position of the crane arm, and the real-time position of the hook.

[0122] Real-time location of large vehicle The formula for obtaining it is:

[0123] Real-time location of the car The formula for obtaining it is:

[0124] ;

[0125] real-time position of hook The formula for obtaining it is:

[0126] ;

[0127] in: These represent the parameter matrices used to characterize the orientation of the trolley, the gantry, and the hook, respectively.

[0128] Based on the given position offset in the local coordinate system To obtain corrected real-time location data;

[0129] The correction formula for the real-time location data is: ,in Indicates the correction position; Indicates the location of the gantry crane body unit and = ; Represents the parameter matrix and = ;

[0130] Based on the bounding box algorithm, the directional body bounding box of the gantry crane body unit that is currently located and surrounds the key area is obtained by combining the corrected real-time position data of the gantry crane with the key area.

[0131] S23: Define the parameter matrix used to characterize the orientation of the body unit in a gantry crane;

[0132] Furthermore, the main body unit of the gantry crane includes at least a trolley, a slewing platform, a boom root hinge point, a slewing boom, and a hook position;

[0133] And based on the parameter matrix, according to the trolley travel of the gantry crane Lifting height Portal corner and the amplitude of the gate seat To obtain real-time position data of the gantry crane, including the real-time position of the trolley, the real-time position of the slewing platform, the real-time position of the boom root hinge point, the attitude data of the slewing boom, and the real-time position of the hook.

[0134] Real-time location of large vehicle The formula for obtaining it is:

[0135] Real-time position of slewing platform The formula for obtaining it is:

[0136] ; ;

[0137] ,

[0138] Real-time position of arm root hinge point The formula for obtaining it is:

[0139] , ;

[0140] Slewing boom attitude data The formula for obtaining it is:

[0141] ; ,

[0142] real-time position of hook The formula for obtaining it is:

[0143] ;

[0144] in, These represent the parameter matrices used to characterize the orientation of the main vehicle and the orientation of the slewing platform, respectively. Indicates the angle of rotation of the gate seat. rotation matrix; This indicates a fixed offset from the given rotary platform to the boom root; Indicates the amplitude of the gate seat. rotation matrix; Indicates arm length; This indicates the height of the hook on the z-axis;

[0145] Based on the given position offset in the local coordinate system To obtain corrected real-time location data;

[0146] The correction formula for the real-time location data is:

[0147] ;

[0148] ,

[0149] in Indicates the correction position; Indicates the location of the gantry crane body unit and = ; Represents the parameter matrix and = ;

[0150] Based on the bounding box algorithm, the directional body bounding box of the body unit in the gantry crane that is currently located and is surrounded by the key area is obtained by combining the real-time position data of the gantry crane with the key area.

[0151] S24: Based on the definition of the oriented ontology bounding box, obtain the set of ontology bounding boxes;

[0152] And the expression for the set of the bounding boxes of the ontology is:

[0153] ,

[0154] In the formula: Represents the body unit; Represents the set of body units and for gantry cranes For gantry cranes ; Indicates the real-time location of the main body unit; This represents the parameter matrix used to characterize the current orientation of the body element; Indicates the half-side length of the oriented bounding box OBB and Indicates the length of the bounding box OBB; This indicates the width of the bounding box OBB; This indicates the height of the bounding box OBB.

[0155] In a specific embodiment, the system described in this embodiment also includes a body bounding box optimization module;

[0156] The bounding box optimization module is used to optimize the bounding box corresponding to the slewing boom in the main body unit of the gantry crane. This addresses the issue that when a fixed gantry crane is modeled using a single-directional bounding box (OBB) under trolley movement or slewing conditions, the furthest protection distance is equal to the diagonal of the bounding box space, leading to overly conservative judgments. This optimization is done without reducing the established safety margin. Without increasing computational complexity, this embodiment optimizes the surrounding area of ​​the rotating arm by chamfering / buttoning. The method is as follows:

[0157] S100: Because the outermost contour of a single body enclosure (OBB) is determined by its spatial diagonal when the slewing boom rotates around the rotation center, the far-end clearance is evaluated based on the diagonal length, resulting in an issue of "overly conservative assessment and increased probability of false triggering." Therefore, the single body enclosure corresponding to the slewing boom is randomly divided into two body enclosures according to a preset direction, and the union of the two divided body enclosures is used as the protective enclosure of the slewing boom. ;

[0158] The protective enclosure The distances from each side of the boom to the straight line containing the outer surface or outer edge of the boom are equal and are defined as follows: ; This indicates the specified collision avoidance distance, which is given by the project / standard; meanwhile, the outer contour of the rotating platform is considered fixed in this embodiment, and its rotation state is not considered.

[0159] S101: In the preset local coordinate system of the slewing boom, the outer corners of the single body bounding box far from the slewing boom's corresponding rotation center are chamfered / trimmed to obtain a chamfered bounding box; and the chamfered / trimmed position is the intersection of two body bounding boxes.

[0160] In this embodiment, the original single-body enclosure (OBB) of the boom is replaced with the union of two (or a small number of) OBBs to form a new protective enclosure. This union is geometrically equivalent to chamfering / trimming the outer corners of the original single-body bounding box OBB away from the center of rotation (i.e., "cutting out" a rectangular area from the original single-body bounding box OBB), as shown in the figure below. This causes the "farthest protection distance" to converge from the original diagonal length to a radial length consistent with the boom's rotation radius, while satisfying... At the same time, it reduces over-conservatism. The union representation preserves the union of enclosing units, so collision determination still remains the intersection detection of a finite number of OBBs. The computational complexity and implementation method remain unchanged, such as... Figures 2 to 3 As shown, the steps include:

[0161] S1011: Defines the chamfer length to be chamfered at the chamfer / bevel position. With the width of the chamfer Furthermore, a constraint is set on the length of the beveled bounding box hypotenuse. In this embodiment, the goal is to ensure that the most unfavorable straight path of the truncated outer contour to the platform side effectively satisfies the net distance requirement without increasing computational complexity. The hypotenuse length constraint is as follows:

[0162] ,

[0163] ,

[0164] In the formula: Indicates the target value of the beveled bounding box hypotenuse length. ,and ; Indicates a fixed safety margin and ; , ;

[0165] S1012: Solve and obtain the length of the chamfer to be chamfered based on the constraint of the hypotenuse length of the chamfered bounding box. With the width of the chamfer for:

[0166] ,

[0167] ,

[0168] S1013: Obtain the chamfered bounding box after chamfering / trimming according to S1012; this embodiment obtains the chamfered bounding box after chamfering / trimming. Adjustments can reduce over-conservatism, achieving similar collision avoidance distances along the rotation path;

[0169] S102: Achieving the protective enclosure based on the chamfered enclosure box To optimize, its expression is:

[0170] ,

[0171] In the formula: Indicates the number of partitions in a single body's bounding box; Indicates the first A beveled surround box.

[0172] In this embodiment, the original single body bounding box (OBB) is replaced with a union envelope of a small number of OBBs. (generally ),like Figure 4 The above is geometrically equivalent to applying a chamfer / bevel to the outer corner of the platform; collision detection still uses bounding box intersection detection, as shown. Figure 5As shown, the union rule of "intersection of any unit determines the overall intersection" is followed, keeping the computational complexity and refresh cycle unchanged. While maintaining the OBB intersection determination interface and refresh cycle, the truncated union enclosing achieves radial convergence of the boom outline, significantly reducing false alarms and false stops caused by diagonal length dominance. At the same time, it does not introduce high-overhead algorithms such as mesh-level collision or GJK / EPA, ensuring real-time performance and engineering feasibility.

[0173] In a specific embodiment, the nearest distance acquisition module is used to acquire the instantaneous nearest distance between the body-oriented bounding box in the collision avoidance space and the body-oriented bounding box in the obstacle space; the method for acquiring the instantaneous nearest distance specifically includes the following steps:

[0174] S10: Get the current crane object With obstacle objects The collision avoidance space and obstacle space are respectively denoted as , , ;in Indicates the corresponding crane object The Middle The body-oriented bounding box of each substructure; Represents the corresponding obstacle object The Middle The body-oriented bounding box of each substructure;

[0175] S11: Obtain any pair of bounding boxes in the collision avoidance space and obstacle space, denoted as an orientation bounding box pair. , The separation axis theorem is used to determine whether the oriented bounding box pairs intersect.

[0176] If an intersection is confirmed, the distance between the current directed bounding box pairs is defined as 0; if no intersection is confirmed, a set of candidate bounding box pairs is obtained; and for all candidate directed bounding box pairs in the candidate bounding box pair set, the geometric minimum Euclidean distance is calculated. Simultaneously, based on the geometric minimum Euclidean distance... The expression for obtaining the instantaneous nearest distance at different times is:

[0177] ,

[0178] In the formula: express The instantaneous closest distance corresponds to the moment of birth;

[0179] Specifically, the geometric minimum Euclidean distance This includes: calculating the distance from any edge segment within one of the candidate oriented bounding boxes to an edge segment within the other oriented bounding box, specifically:

[0180] The nearest point parameter used to calculate the distance between edge segments is obtained as follows:

[0181] ,

[0182] ,

[0183]

[0184] ,

[0185] ,

[0186] In the formula: Represents the reference point of the edge segment P in a directional bounding box; The parametric equation point of the edge segment P in the oriented bounding box; Indicate design parameters; This represents the reference point of the edge segment Q in another oriented bounding box; Represents the parametric equation point of the edge segment Q in another oriented bounding box; , , Indicates intermediate parameters; This represents the direction vector of line segment P; This represents the direction vector of line segment Q; , Represents the parameters of the nearest points on the lines corresponding to the two line segments respectively. , ;

[0187] Calculate the nearest point parameters , The Euclidean distance between them is used as a candidate distance for obtaining the instantaneous closest distance;

[0188] For each candidate bounding box, the distance from any point on one bounding box to the rectangular face of the bounding box in the other bounding box is calculated as follows:

[0189] Define the bounding box parameters for any rectangular face on one of the candidate oriented bounding boxes; and the bounding box parameters include the center of the rectangular face. Unit legal direction orthogonal basis in plane The length and width of half of the rectangle ;

[0190] Based on the bounding box parameters, the distance between any point on another oriented bounding box and any rectangular face on another oriented bounding box is obtained, and this distance is used as a candidate distance for obtaining the instantaneous nearest distance. The expression is as follows:

[0191] ,

[0192] , ,

[0193] ,

[0194] In the formula: Indicates the intermediate vector; Represents two mutually orthogonal unit direction axes within a rectangular plane; This represents the distance from any point on another orientation bounding box to any rectangular face on another orientation bounding box.

[0195] This embodiment also includes using AABB bounding boxes when each submechanism only has linear motion, and using OBB bounding boxes if there is rotational motion. and When the axis-aligned bounding box (AABB) is used, the geometric minimum Euclidean distance is:

[0196] ,

[0197] ,

[0198] In the formula: This represents the separation distance of the AABB bounding box along the x-axis (positive when the intervals do not intersect, and 0 when they intersect / overlap). This indicates the lower bound of the A bounding box in the AABB bounding box on the x-axis; This indicates that the bounding box B in the AABB bounding box is at the upper bound of the x-axis; This indicates the lower bound of the B-boundary box in the AABB-boundary box on the x-axis; This indicates that the bounding box A in the AABB bounding box is at the upper bound of the x-axis; d represents the minimum three-dimensional Euclidean distance between bounding boxes A and B in an AABB bounding box (d=0 if the two boxes intersect).

[0199] The regional hierarchical structure is used to determine the execution state region type of the crane and obstacles in the surrounding environment based on the instantaneous closest distance; and the execution state region type includes warning zone, braking zone and normal zone.

[0200] Specifically, the method for determining the execution state region type of the crane and obstacles in the surrounding environment based on the instantaneous nearest distance is as follows:

[0201] S31: Obtain the gear position parameters of each body unit in the corresponding crane; and the gear position parameters include, but are not limited to, the luffing gear value, slewing gear value, and trolley forward and backward travel gear value of the crane; the actual value of the gear position parameter can be positive or negative, and the direction of movement of the mechanism can be determined by the positive or negative value of the gear position.

[0202] Based on the gear position parameters, the 3D position and orientation of the current body unit are determined, and the position and orientation information of multiple adjacent / following body units that are adjacent to or have a following relationship with the current body unit are obtained. Based on the motion direction, the directional bounding boxes of the corresponding adjacent / following body units are obtained according to the position and orientation information, and the union of these directional bounding boxes is defined as the potential collision avoidance direction set and denoted as . ;

[0203] S32: For any potential collision avoidance direction and Based on expert experience, two types of directional bounding boxes are set for the current ontology unit, including the early warning bounding box. and in Internal brake enclosure For all directions of the same body unit The union of these two sets is denoted as the warning boundary. ,all The union of the two sets is denoted as the braking boundary. And satisfy ;

[0204] S33: Based on a given fixed safety margin The equivalent distance used for early warning judgment is obtained based on the instantaneous nearest distance. The formula is:

[0205] ,

[0206] In this embodiment, the warning boundary can be obtained based on the kinematic characteristics and safety strategy of the object (e.g., different parts of the object have different ranges of motion or sizes, requiring different anti-collision ranges to be set for each part). and the braking boundary The specific values ​​include fixed safety margins for positioning errors, fixed network and execution link hysteresis, driver response time window, and specific working conditions (such as typical swing amplitude when hoisting large components). These fixed safety margins do not depend on on-site speed measurement or calculation, but are determined as constants during the project commissioning phase according to the equipment manual, acceptance standards, and on-site experience, and can be grouped and configured according to equipment type, track section, and operation mode.

[0207] S34: Based on equivalent distance Formulas are used to obtain the collision avoidance space and warning boundary, respectively. Braking boundary The equivalent distance between them is used to confirm the intersection relationship, and the execution state area of ​​the crane and the obstacles in the surrounding environment is determined based on the intersection relationship; and the execution state area includes the warning area, the braking area and the normal area;

[0208] The method for confirming the intersection relationship is as follows:

[0209] Get the crane object One or more sub-organizations When the gear parameter value is not zero, its collision avoidance direction The corresponding warning boundary With braking boundary ;

[0210] Get another crane or obstacle object The collision avoidance space is based on the equivalent distance. Formula to get objects Collision avoidance space and warning boundary Braking boundary The real-time equivalent distance;

[0211] Objects are identified based on real-time equivalent distance. Collision avoidance space and warning boundary Braking boundary The intersection relationship of objects; if objects Collision avoidance space and warning boundary Intersecting and with braking boundary If they do not intersect, then the execution state area at this time is confirmed as a warning zone; if the objects Collision avoidance space and warning boundary Intersecting and with braking boundary If they intersect, the execution state region is confirmed as the braking region; otherwise, the execution state region is confirmed as the normal region. Furthermore, this embodiment considers the object's rotation or translation during the process. or It is possible that the object will intersect with its own anti-collision space. If such self-intersection occurs, the system will determine it as an invalid risk and will not trigger a warning or braking. The specific mechanism for determining whether it is an invalid risk is that all the bounding boxes of the body with potential self-intersection will be monitored. If the current object intersects with the bounding box of the body, the system will determine whether the bounding box intersects with the bounding box of the body. or Intersecting bounding boxes that fall within the bounding box of the main body are considered invalid risks.

[0212] The collision avoidance decision module is used to execute the corresponding collision avoidance operation based on the type of execution state area where the crane is located.

[0213] Specifically, when the operating state area of ​​the crane is confirmed to be a warning zone, a significant warning is simultaneously issued on the HMI in the cab and the ground control terminal, and a warning is also sent to the crane target. The system issues deceleration / downshift commands to the relevant gears; when it confirms that the crane's operating state area is a braking area, the system determines that there is a collision risk and sends a deceleration / downshift command to the crane. The relevant gear position issues the braking command; in this state, only when The direction of the relevant gear shift command is opposite to that at the time of triggering (i.e., away from the trigger). Movement is only permitted when the direction of movement is indicated (i.e., when the risk is eliminated).

[0214] This embodiment also includes bypass and recording: for special crane operating conditions requiring temporary crossing of the warning limit, the operator can apply for short-term clearance via the "bypass" switch on the control panel. During the bypass period, the system continues to monitor any object using the same distance criterion. It tracks the intersection relationships between the collision avoidance space and other objects, and automatically records the start and end times, duration, and involved equipment and space range of the bypass. After the bypass ends or the risk is eliminated, the collision avoidance logic automatically returns to normal, completing closed-loop control without relying on speed or acceleration information throughout the entire process.

[0215] Control strategies and human-machine interaction:

[0216] (a) Strategy Priority and Arbitration Rules: To avoid command conflicts, the system sets a hierarchical handling priority: Braking Zone > Warning Zone > Normal Zone. When multiple triggering conditions exist at the same time, the priority arbitration module selects the highest level of handling according to the above order and blocks the lower level of handling. When the handling levels from different objects / organizations are the same, the "most unfavorable principle" (selecting the stricter handling) is adopted, and the most recent arrival timestamp is used for first-come-first-served handling.

[0217] (II) Execution of Control Command Set and Issuance Mechanism: The system issues the following atomic commands to the relevant gears (travel, slewing, luffing, and hoisting, etc.) of the target object. All commands are transmitted to the PLC for execution via fieldbus / industrial Ethernet:

[0218] Deceleration / Downshifting: Downshift the current gear by a preset ratio or discrete levels until the lower speed limit is reached;

[0219] Stop / Brake: Issues a stop or emergency stop command to the relevant gear;

[0220] Directional constraints: Only movement away from the risk object is allowed; any positive displacement in the direction of approach is prohibited.

[0221] Release timing control: When the risk is released or the bypass expires, restore normal control according to the unlocking sequence (first remove the directional constraint, then release the deceleration / stop);

[0222] Command hold and watchdog: Sets hold bit and watchdog heartbeat for braking / direction constraints, and enters a more stringent conservative state in case of communication failure.

[0223] (III) Bypass Control: When a situation arises where "extremely close-range hoisting or temporary exceeding of the warning limit" is required due to process needs, manual bypass is permitted:

[0224] Triggering condition: The operator of the cab control panel initiates a bypass request via the "bypass" knob / switch;

[0225] Confirmation mechanism: Bypass actions must be confirmed by on-site management personnel via walkie-talkie;

[0226] Monitoring and Recording: During the bypass, the system continues to monitor the intersection relationship with the protected space of adjacent objects according to the same distance criterion, and automatically records the start / end time, duration, identification of involved equipment, and controlled space range of the bypass;

[0227] Statistics and Synchronization: Automatically synchronizes bypass data and timestamps to the upper management platform in real time;

[0228] Exit and Return: After the bypass ends or the risk is eliminated, the control logic automatically returns to normal, and the priority and handling rules remain unchanged.

[0229] (iv) Fault and Degradation Strategies: When abnormalities occur in positioning quality, communication links, or execution feedback (out-of-bounds, packet loss, jitter):

[0230] 1. Expand the effective boundary and upgrade the handling level (if necessary, directly enter the braking zone);

[0231] 2. By default, displacement in the direction of approach is prohibited; only low-speed evacuation in the direction of departure is allowed.

[0232] 3. Record the cause of the anomaly and report it to the management platform for operation and maintenance traceability.

[0233] (v) Control logic of the system described in this embodiment

[0234] S1 Data and Status Acquisition: Acquires real-time PLC pose, relevant gear values, bypass switch status, and data quality indicators; S2 Dynamic Boundary Calculation: Based on... Calculate the currently effective S3 Candidate Object Screening: Performs fast screening (region / track grouping) on ​​surrounding objects to obtain a set of objects to be judged; S4 Intersection Detection: Performs intersection detection on the protection space of each candidate object and... Perform 3D intersection relationship determination; S5 Priority Arbitration: Determine the global handling level based on the maximum handling level (Braking > Warning > Normal); S6 Bypass Determination: If the handling level is Braking and the bypass application is valid and confirmed, proceed to the bypass branch; otherwise, proceed to the normal handling branch; S7 Normal Handling - Normal Zone: Non-intersecting → Maintain normal control, disable alarms, clear constraints and proceed to S11; S8 Normal Handling - Warning Zone: Intersecting And they do not intersect. → HMI / ground terminal issues significant warning; issues deceleration / downshift and directional constraints to relevant gears (only allowing departure); S9 standard handling—braking zone: simultaneous intersection →Issue braking commands and directional constraints to the relevant gears; only allow slow, away-direction evacuation; S10 Bypass Branch: Record the start and end times and range of the bypass and maintain continuous monitoring; if the risk is eliminated or the bypass expires, cancel the bypass and return to S5; otherwise, maintain the bypass status; S11 Release Timing Judgment: if the risk has been eliminated and the data quality meets the threshold, cancel the constraints according to the unlock sequence; otherwise, maintain the current handling and return to the S1 loop; S12 Operation Log and Platform Synchronization: Structure and archive alarm / braking / bypass events and synchronize them to the management platform.

[0235] This embodiment also includes an operational procedure for adaptive system network and system anomaly handling:

[0236] (i) Communication Health Measurement: The system continuously monitors the communication health status between the vehicle and the ground collision avoidance controller. The main PLC, sub-PLC and the arithmetic controller are configured to exchange heartbeat messages at a preset period to quantify link reachability and measure end-to-end round-trip time (RTT).

[0237] (II) Abnormal Judgment Criteria: Communication is deemed abnormal when any of the following situations occur:

[0238] The heartbeat message was not received within the set time window; or the measured round-trip time exceeded the threshold. .

[0239] The above-mentioned triggering reasons may be caused by 5G signal attenuation, base station shutdown or other link failures, but the determination is based on heartbeat and latency measurements.

[0240] (III) Alarms and Event Logs: Once the alarm is confirmed, the HMI in the cab and the ground control terminal will issue a significant alarm simultaneously through sound, light and screen display. The system will also generate a complete event log, including the timestamp of the occurrence / recovery of the anomaly, the duration, the identification of the equipment involved, etc., and write it into the database to form searchable and statistical report entries. The event will be synchronized to the crane intelligent digital management platform according to the RabbitMQ protocol for operation and maintenance tracking and message push.

[0241] (iv) Control principles during abnormal periods: During the period of communication abnormality, the anti-collision logic remains in an alarm state and the control policy is issued and restricted.

[0242] (v) Recovery and smooth regression: When the heart rate stabilizes and the round-trip time is no greater than 1 / 3. When the alarm is cleared, the system automatically cancels the alarm and smoothly restores the control logic to the normal anti-collision closed loop in a predetermined sequence; the above process does not require manual intervention and does not introduce sudden changes in on-site operation.

[0243] The system described in this embodiment includes the following application examples:

[0244] Application Example 1: Collision Prevention for Two Gantry Cranes Meeting on the Same Track

[0245] Scenario and Objective: Two gantry cranes on track A in the West Zone may approach each other in the same or opposite directions during daily operations, posing a collision risk. The objective of this example is to determine the proximity and control deceleration / braking between the two cranes using only real-time position data from a PLC, without requiring speed / acceleration estimation.

[0246] Deployment and Parameters: Each crane is equipped with an industrial-grade 5G CPE and connected to the factory's 5G SA private network (end-to-end control link latency target less than 50ms). The vehicle-side HMI is a 22-inch display. Ground-based computing controllers and digital twin display controllers are installed. The dynamic bounding box of the gantry crane is generated according to the vehicle's shape and position. Two levels of fixed thresholds are pre-tuned: the outer layer is the warning boundary, and the inner layer is the braking boundary. Both contain fixed safety margins for positioning errors, execution link hysteresis, and model approximation.

[0247] Operation Process: During operation, the PLCs of the two gantry cranes transmit real-time position data, such as the trolley travel and the gantry position, to the computing controller via the 5G network. The controller generates dynamic bounding boxes for the two gantry cranes in the digital twin model and calculates the instantaneous closest distance between the two bounding boxes according to the representative point / representative surface set. When the closest distance is less than or equal to the warning boundary, the system immediately issues a significant warning on the HMIs of both gantry cranes and the ground end, and issues a deceleration command; if the closest distance continues to shrink and is less than or equal to the braking boundary, the system triggers forced braking and simultaneously restricts the driver's control input, allowing only reverse operations to separate the two gantry cranes. The entire process is written to the local database, including the trigger and release times and equipment information; the digital twin screen synchronously displays the proximity status and color-coded prompts. Once the closest distance between the two gantry cranes returns to outside the warning boundary, the system automatically releases the restriction and resumes normal control.

[0248] Effects and benefits: Real-time approach detection and hierarchical control of gantry cranes on the same track can be achieved without relying on speed or acceleration estimation, reducing the risk of collision and generating auditable event records, which facilitates post-event review and parameter optimization.

[0249] Application Example 2: Clearance and Collision Prevention between the Boom / Counterweight of a Gantry Crane and the Rigid / Flexible Legs of a Shipbuilding Gantry Crane

[0250] Scenario and Objective: During slewing and luffing operations, the boom end and counterweight tail of a gantry crane may come into contact with the rigid / flexible legs of a shipbuilding gantry crane. The objective is to avoid physical contact within a confined space using geometric bounding box clearance determination.

[0251] Deployment and parameters: Both the gantry crane and the portal crane are connected to the 5G SA private network; the computing controller generates the instantaneous bounding box of the gantry based on the slewing angle, amplitude angle and spreader height reported by the PLC; the rigid / flexible legs of the portal crane are approximated by feature surfaces; two fixed distance thresholds for warning and braking are set between the two types of equipment.

[0252] Operation: When the gantry crane begins to rotate and luff, the system updates the spatial positions of the boom end and counterweight end based on real-time position data; simultaneously, it acquires the current posture and position of the rigid and flexible legs of the gantry crane. The arithmetic controller calculates the minimum clearance between the aforementioned key components. When the clearance is less than or equal to the warning boundary, the system issues a deceleration warning to the gantry crane and limits the rotation / luffing speed (if the equipment provides this command channel); when the clearance further decreases to or within the braking boundary, the system applies braking to the gantry crane and restricts the operator to only perform reverse operations that increase the clearance. The system automatically recovers after the anomaly is resolved.

[0253] Effects and benefits: By calculating the clearance of the enclosure box in key areas, spatial isolation between the gantry and the gantry crane is achieved, avoiding interference between large components during close-range operations, while maintaining complete and traceable data records.

[0254] Application Example 3: Collision Prevention for Gantry Cranes of the Same Track / Different Track Types Meeting

[0255] Scenario and Objective: Two gantry cranes share the same rail for loading, unloading, and transfer operations. During slewing and luffing, the boom end and the counterweight tail end may approach or even interfere when they meet longitudinally. The objective is to determine and classify the geometric proximity using only real-time position data under the condition of parallel operation.

[0256] Deployment and parameters: Two gantry cranes are connected to a unified global coordinate digital twin platform; the system generates their respective dynamic bounding boxes in real time based on the rotation angle, amplitude angle and travel distance reported by the PLC, and sets an outer warning boundary and an inner braking boundary. The boundaries include fixed safety margins for measurement error, link hysteresis and structural deflection.

[0257] Operation Process: During operation, the computing controller continuously calculates the instantaneous closest clearance between key components of the two gantry cranes (boom end, counterweight tail, tower slewing radius, etc.) and the other vehicle's encirclement box. When the clearance is less than or equal to the warning boundary, the system immediately issues a warning on the HMI and at the ground level and sends a deceleration command to the vehicle in the encounter. If the clearance further becomes less than or equal to the braking boundary, forced braking is triggered, and the driver can only perform reverse maneuvers until the risk is eliminated. The event is written to the local database and the passing vehicles are marked with color-coded levels on the large screen.

[0258] Effects and benefits: To address the typical passing risks of vehicles on the same track, the system enables location-level clearance determination and graded control without introducing speed estimation, thereby reducing close-range interference between the boom, counterweight, and tower body.

[0259] Application Example 4: Collision Prevention When a Large Gantry Crane Crosses a Small Gantry Crane

[0260] Scenario and Objective: A large shipbuilding gantry crane needs to pass over a smaller gantry crane. The main risk in this scenario lies in the relative position of the gantry crane's lifting rope and hook: the gantry crane's hook / rope may sweep across the smaller gantry crane's beams, walkways, or railings; conversely, the smaller gantry crane may also collide with the gantry crane's lifting rope and hook. The objective is to determine and classify the geometric relationship between the lifting rope and hook using only real-time position data, without introducing speed / swing angle estimation, to ensure safe crossing.

[0261] Deployment and Parameters: Each hoisting rope of each device is modeled as a straight line segment from the upper pulley reference point to the hook connection point, and assigned a rectangular bounding box; the hook is modeled as a rectangular bounding box. The system sets two levels of fixed thresholds: a warning boundary and a braking boundary, both of which include fixed safety margins for measurement error, link hysteresis, structural deflection, and minor rope offsets not modeled.

[0262] Operation Process: The computing controller calculates the closest clearance between the large gantry crane rope and the hook enclosure and the small gantry enclosure, using the track coordinates as a reference, and considering clearances in the front-to-back, left-to-right, and up-down directions. When the clearance enters the warning boundary, the system issues an alarm on the HMI and decelerates the moving vehicle; if the clearance reaches the braking boundary, the system brakes the relevant gantry and restricts forward movement, allowing operation only in the direction that increases the clearance. Automatic recovery occurs after the risk is eliminated.

[0263] Effects and benefits: By modeling the sling and hook as independent dynamic bounding boxes and using them as the primary criteria for crossing working conditions, the system can achieve deterministic and interpretable safety control without considering sling swing, relying on fixed safety margins and two-level distance thresholds; alarm events are fully logged for easy review and parameter optimization.

[0264] Application Example 5: Collision Prevention When a Gantry Crane Crosses a Small Gantry Crane

[0265] Scenario and Objective: A large gantry crane needs to pass over a small portal crane. The main risks in this situation are threefold: First, the gantry crane's hook / rope may sweep against the portal crane's boom, tower, etc.; second, if the portal crane is in a position with excessively high luffing, its boom may collide with the lower crossbeam of the gantry crane; third, if the portal crane's slewing angle is inappropriate, its boom or counterweight may interfere with the rigid / flexible legs of the gantry crane. The objective is to determine and classify the above geometric relationships using only real-time position data, without introducing speed or sway angle estimation, to ensure safe crossing.

[0266] Deployment and Parameters: The lifting ropes of both devices are modeled as straight line segments from the upper pulley reference point to the hook connection point, and assigned a cuboid bounding box; the hook is also modeled as a cuboid bounding box. Cuboid / polyhedral bounding boxes that can be updated in real time with rotation and luffing are created for the boom and counterweight of the small portal frame; bounding boxes for each mechanism are created for the gantry crane. The system sets two levels of fixed thresholds: a warning boundary and a braking boundary, both of which include fixed safety margins required for measurement errors, link hysteresis, structural deflection, and minor unmodeled rope offsets.

[0267] Operation process: The arithmetic controller calculates three types of closest clearances based on the track coordinates:

[0268] The system identifies three key clearance categories: 1) the closest clearance between the portal crane's rope / hook enclosure and the enclosures of various mechanisms on the small portal (boom end, tower body, walkway, etc.); 2) the vertical clearance between the small portal crane's boom enclosure and the portal crane's lower crossbeam enclosure (considering both vertical and horizontal height restrictions); and 3) the closest planar / spatial clearance between the small portal crane's boom / counterweight enclosure and the portal crane's rigid / flexible leg enclosure. When any of these closest clearances enters the warning boundary, the system issues a significant alarm on the relevant equipment's HMI and issues a deceleration command to the moving vehicle. When any of these closest clearances reaches the braking boundary, the system applies braking to the equipment on the side closer to the risk and restricts the driver's control input to only allow operations in the direction of increasing clearance (e.g., lifting / lowering, slewing / luffing reversal, trolley / tower reversal, etc.). When all three types of closest clearances return to outside the warning boundary, the process automatically unlocks, and the equipment returns to normal. The trigger / release times and involved equipment throughout the entire process are written to a local database, supporting visual playback and statistical reports.

[0269] Effects and benefits: By unifying the three types of risks—rope / hook sweep, boom-beam height limit, and boom / counterweight-rigid / flexible leg interference—into a geometric decision framework of "dynamic bounding box-nearest clearance-two-level threshold," the system achieves deterministic and interpretable safety control without considering rope swing; the "graded control" strategy avoids passive braking only when approaching the critical point, significantly reducing the probability of collisions during the crossing process; all events are fully traced, facilitating review and parameter optimization.

[0270] The beneficial effects of the method described in this embodiment are as follows:

[0271] (1) A high-fidelity three-dimensional digital twin model based on the geometric parameters of the equipment is constructed. Collision determination is performed in the virtual space using dynamic geometric bounding boxes and the nearest distance. The determination does not rely on external sensors such as optical / ultrasonic sensors that are easily affected by the environment. Instead, it is calculated based on the real-time position data of the PLC / encoder and a fixed safety margin is superimposed on the threshold. Therefore, it is not sensitive to environmental disturbances.

[0272] (2) Parametric modeling of key mechanisms such as boom end, counterweight tail, rigid / flexible legs of gantry crane, trolley, and hook is performed in the pre-set twin model, and dynamic bounding boxes are generated according to real-time attitude; the nearest gap of the bounding box is solved by representative points / representative surfaces, and thresholds are set for different mechanisms to realize component-level anti-collision logic from the model level.

[0273] (3) Construct a digital twin platform with a unified global coordinate system, centrally integrate the real-time position data of multiple cranes, uniformly calculate proximity relationships and visualize them in a three-dimensional interface / large screen; when a collision risk occurs, provide early warning and hierarchical control, and push messages to the management terminal.

[0274] (4) Two-level distance boundary: warning zone → braking zone; after entering the braking zone, only reverse operation is allowed, forming a progressive safety control: first, a warning / deceleration, then rigid braking, to ensure rapid untangling and reduce the risk of human operation. In addition, the computing power of the system described in this embodiment, such as the calculation cycle time, can be improved to 1ms.

[0275] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A crane full-domain collision avoidance system based on real-time digital twin and dynamic bounding box, characterized in that, It includes a 3D modeling module, a dynamic acquisition module for oriented bounding boxes, a nearest distance acquisition module, a region hierarchical construction module, and a collision avoidance decision module; The 3D modeling module is used to acquire real-time pose data of the crane and obstacles in the surrounding environment after pre-installation and calibration under a preset coordinate system. The preset coordinate system includes a preset global coordinate system of the plant area and a local coordinate system about the crane's track, traveling mechanism and rotation center. At the same time, based on the 3D modeling software, the module acquires the digital twin 3D position model of the crane and obstacles according to the real-time pose data. The digital twin 3D position model includes the real-time 3D model of each component structure in the crane and obstacles. The dynamic acquisition module is used to configure dynamic body orientation bounding boxes for each component structure according to the real-time 3D model, and to take the union of the body orientation bounding boxes of the corresponding crane as the anti-collision space; and to take the union of the body orientation bounding boxes of the corresponding obstacle as the obstacle space. The closest distance acquisition module is used to acquire the instantaneous closest distance between the body-oriented bounding box in the collision avoidance space and the body-oriented bounding box in the obstacle space; The regional hierarchical structure is used to determine the execution state region type of the crane and obstacles in the surrounding environment based on the instantaneous closest distance; and the execution state region type includes warning zone, braking zone and normal zone; The collision avoidance decision module is used to execute the corresponding collision avoidance operation based on the type of execution state area where the crane is located.

2. The crane global collision avoidance system based on real-time digital twin and dynamic bounding box as described in claim 1, characterized in that, The method for obtaining the instantaneous closest distance between the body-oriented bounding box in the collision avoidance space and the body-oriented bounding box in the obstacle space specifically includes the following steps: S10: Get the current crane object With obstacle objects The collision avoidance space and obstacle space are respectively denoted as , , ;in Indicates the corresponding crane object The Middle The body-oriented bounding box of each substructure; Represents the corresponding obstacle object The Middle The body-oriented bounding box of each substructure; S11: Obtain any pair of bounding boxes in the collision avoidance space and obstacle space, denoted as an orientation bounding box pair. , ; The separation axis theorem is used to determine whether the oriented bounding box pairs intersect; If an intersection is confirmed, the distance between the current directed bounding box pairs is defined as 0; if no intersection is confirmed, a set of candidate bounding box pairs is obtained; and for all candidate directed bounding box pairs in the candidate bounding box pair set, the geometric minimum Euclidean distance is calculated. Simultaneously, based on the geometric minimum Euclidean distance... The expression for obtaining the instantaneous nearest distance at different times is: In the formula: express The instantaneous closest distance corresponds to the moment.

3. A crane global collision avoidance system based on real-time digital twin and dynamic bounding box as described in claim 2, characterized in that, The specific methods for obtaining the directional bounding boxes of each body of the crane are as follows: S20: Define device base reference point and ;in This indicates the equipment base reference point in the pre-set global coordinate system of the plant area. The x-coordinate, y-coordinate, and vertical coordinate; Define the key areas of the crane, including the end of the slewing boom, the tail end of the counterweight, the structural feature surfaces of the gantry crane's rigid / flexible legs, the trolley and lifting gear, and the track end area; and confirm the crane's real-time PLC inputs based on a preset PLC control system, wherein the real-time PLC inputs include at least the trolley travel at the current moment. Car journey Lifting height Portal corner and the amplitude of the gate position ; S21: Confirm the type of the current crane. If the crane type is confirmed to be a gantry crane, continue to S22; if the crane type is confirmed to be a portal crane, continue to S23. S22: Defines a parameter matrix used to characterize the orientation of the body unit in a gantry crane; Furthermore, the gantry crane body unit includes at least a trolley, a crane arm, and a hook; And based on the parameter matrix, according to the gantry crane's trolley travel... Car journey and lifting height Combined with equipment base reference point Acquire real-time position data of the gantry crane, including the real-time position of the trolley, the real-time position of the crane arm, and the real-time position of the hook. Real-time location of large vehicle The formula for obtaining it is: Real-time location of the car The formula for obtaining it is: ； real-time position of hook The formula for obtaining it is: ； in: These represent the parameter matrices used to characterize the orientation of the trolley, the gantry, and the hook, respectively. Based on the given position offset in the local coordinate system To obtain corrected real-time location data; The correction formula for the real-time location data is: ,in Indicates the correction position; Indicates the location of the gantry crane body unit and = ; Represents the parameter matrix and = ; Based on the bounding box algorithm, the directional body bounding box of the gantry crane body unit that is currently located and surrounds the key area is obtained by combining the corrected real-time position data of the gantry crane with the key area. S23: Define the parameter matrix used to characterize the orientation of the body unit in a gantry crane; Furthermore, the main body unit of the gantry crane includes at least a trolley, a slewing platform, a boom root hinge point, a slewing boom, and a hook position; And based on the parameter matrix, according to the trolley travel of the gantry crane Lifting height Portal corner and the amplitude of the gate position To obtain real-time position data of the gantry crane, including the real-time position of the trolley, the real-time position of the slewing platform, the real-time position of the boom root hinge point, the attitude data of the slewing boom, and the real-time position of the hook. Real-time location of large vehicle The formula for obtaining it is: Real-time position of slewing platform The formula for obtaining it is: ； ； Real-time position of arm root hinge point The formula for obtaining it is: , ； Slewing boom attitude data The formula for obtaining it is: ； real-time position of hook The formula for obtaining it is: ； in, These represent the parameter matrices used to characterize the orientation of the main vehicle and the orientation of the slewing platform, respectively. Indicates the angle of rotation of the gate seat. The rotation matrix; This indicates a fixed offset from the given rotary platform to the boom root; Indicates the amplitude of the gate seat. The rotation matrix; Indicates arm length; This indicates the height of the hook on the z-axis; Based on the given position offset in the local coordinate system To obtain corrected real-time location data; The correction formula for the real-time location data is: ； in Indicates the correction position; Indicates the location of the gantry crane body unit and = ; Represents the parameter matrix and = ; Based on the bounding box algorithm, the directional body bounding box of the body unit in the gantry crane that is currently located and is surrounded by the key area is obtained by combining the real-time position data of the gantry crane with the key area. S24: Based on the definition of the oriented ontology bounding box, obtain the set of ontology bounding boxes; And the expression for the set of the bounding boxes of the ontology is: In the formula: Represents the body unit; Represents the set of body units and for gantry cranes For gantry cranes ; Indicates the real-time location of the main body unit; This represents the parameter matrix used to characterize the current orientation of the body element; Indicates the half-side length of the oriented bounding box OBB and Indicates the length of the bounding box OBB; Indicates the width of the bounding box OBB; This indicates the height of the bounding box OBB.

4. A crane full-domain collision avoidance system based on real-time digital twin and dynamic bounding box as described in claim 3, characterized in that, It also includes a body bounding box optimization module; The enclosure box optimization module is used to optimize the enclosure box corresponding to the slewing boom in the gantry crane body unit. The method is as follows: S100: The single body enclosure corresponding to the slewing boom is randomly divided into two body enclosures according to a preset direction, and the union of the two body enclosures is used as the protective enclosure of the slewing boom. ; The protective enclosure The distances from each side of the boom to the straight line containing the outer surface or outer edge of the boom are equal and are defined as follows: ; S101: In the preset local coordinate system of the slewing boom, the outer corners of the single body bounding box far from the slewing boom's corresponding slewing center are chamfered / trimmed to obtain a chamfered bounding box; and the chamfered / trimmed position is the intersection of two body bounding boxes, specifically: S1011: Defines the chamfer length to be chamfered at the chamfer / bevel position. width of the chamfer And set the chamfered bounding box hypotenuse length constraint as follows: In the formula: Indicates the target value of the beveled bounding box hypotenuse length. ,and ; Indicates a fixed safety margin and ; S1012: Solve and obtain the length of the chamfer to be chamfered based on the constraint of the hypotenuse length of the chamfered bounding box. width of the chamfer for: S1013: Obtain the chamfered bounding box after chamfering / bevel processing according to S1012; S102: Achieving the protective enclosure based on the chamfered enclosure box To optimize, its expression is: In the formula: Indicates the number of partitions in a single body's bounding box; Indicates the first A beveled surround box.

5. A crane full-domain collision avoidance system based on real-time digital twin and dynamic bounding box as described in claim 4, characterized in that, The method for determining the execution state region of the crane and obstacles in the surrounding environment based on the instantaneous nearest distance is as follows: S31: Obtain the gear parameters of each body unit in the corresponding crane; The gear parameters include, but are not limited to, the luffing gear value, slewing gear value, and trolley forward and backward travel gear value of the crane. Based on the gear position parameters, the 3D position and orientation of the current body unit are determined. Position and orientation information of multiple adjacent / following body units that are adjacent to or have a following relationship with the current body unit are obtained. Based on the motion direction and position / orientation information, the directional bounding boxes of the corresponding adjacent / following body units are obtained. The union of these directional bounding boxes is defined as the potential collision avoidance direction set and denoted as... ; S32: For any potential collision avoidance direction and Based on expert experience, two types of directional bounding boxes are set for the current ontology unit, including the early warning bounding box. and in Internal brake enclosure This will apply to all directions of the same body unit. The union of these two sets is denoted as the warning boundary. ,all The union of the two sets is denoted as the braking boundary. And satisfy ; S33: Based on a given fixed safety margin The equivalent distance used for early warning judgment is obtained based on the instantaneous nearest distance. The formula is: S34: Based on equivalent distance Formulas are used to obtain the collision avoidance space and warning boundary, respectively. Braking boundary The equivalent distance between them is used to determine the intersection relationship, and the execution state area of ​​the crane and the obstacles in the surrounding environment is determined based on the intersection relationship; and the execution state area includes the warning area, the braking area and the normal area; The method for confirming the intersection relationship is as follows: Get the crane object One or more sub-organizations When the gear parameter value is not zero, its collision avoidance direction The corresponding warning boundary With braking boundary ; Get another crane or obstacle object The collision avoidance space is based on the equivalent distance. Formula to get objects Collision avoidance space and warning boundary Braking boundary The real-time equivalent distance; Objects are identified based on real-time equivalent distance. Collision avoidance space and warning boundary Braking boundary The intersection relationship of objects; if objects Collision avoidance space and warning boundary Intersecting and with braking boundary If they do not intersect, then the execution state area at this time is confirmed as a warning zone; if the objects Collision avoidance space and warning boundary Intersecting and with braking boundary If they intersect, the execution state area is confirmed to be the braking zone; otherwise, the execution state area is confirmed to be the normal zone.

6. A crane full-domain collision avoidance system based on real-time digital twin and dynamic bounding box according to claim 5, characterized in that, The geometric minimum Euclidean distance described in S10 This includes: calculating the distance from any edge segment within one of the candidate oriented bounding boxes to an edge segment within the other oriented bounding box, specifically: The nearest point parameter used to calculate the distance between edge segments is obtained as follows: In the formula: Represents the reference point of the edge segment P in a directional bounding box; The parametric equation point of the edge segment P in the oriented bounding box; Indicate design parameters; This represents the reference point of the edge segment Q in another oriented bounding box; Represents the parametric equation point of the edge segment Q in another oriented bounding box; , , Indicates intermediate parameters; This represents the direction vector of line segment P; This represents the direction vector of line segment Q; , This represents the parameter of the nearest point on the line corresponding to each of the two line segments; Calculate the nearest point parameters , The Euclidean distance between them is used as a candidate distance for obtaining the instantaneous closest distance; For each candidate bounding box, the distance from any point on one bounding box to the rectangular face of the bounding box in the other bounding box is calculated as follows: Define the bounding box parameters for any rectangular face on one of the candidate oriented bounding boxes; and the bounding box parameters include the center of the rectangular face. Unit legal direction orthogonal basis in plane The length and width of half a side corresponding to the rectangle ; Based on the bounding box parameters, the distance between any point on another oriented bounding box and any rectangular face on another oriented bounding box is obtained, and this distance is used as a candidate distance for obtaining the instantaneous nearest distance. The expression is as follows: , In the formula: Indicates the intermediate vector; Represents two mutually orthogonal unit direction axes within a rectangular plane; This represents the distance from any point on another orientation bounding box to any rectangular face on another orientation bounding box.

Images