A yard intelligent vehicle dispatching method and system based on parallel cooperation

Abstract

The present application relates to the technical field of port logistics, in particular to a yard intelligent vehicle scheduling method and system based on parallel cooperation, comprising the following steps: integrating coordinate parameters to generate scheduling interference topological graph data, constructing concurrent task queues and establishing running state data, mapping three-dimensional envelope surfaces to calculate intersection volume difference, mapping the volume difference to brake level injection motor output avoidance instructions, and combining swing angle acceleration and car acceleration to derive superimposed compensation current. In the present application, three-dimensional anti-collision logic is mapped as a graph structure priority topology to ensure concurrent task immunity to space deadlock conflicts, volume derivative is used to dynamically predict multi-machine displacement overlap rate to realize flexible sliding peak-shifting avoidance of rigid brake physical collision loss, and space cross-multiplication nonlinear coupling deviation is combined to extract feedforward compensation current to compensate for the potential energy of unbalanced load and ensure the stability of the vehicle handling process.

Description

Technical Field

[0001] This invention relates to the field of port logistics technology, and in particular to a method and system for intelligent yard traffic scheduling based on parallel collaboration. Background Technology

[0002] The port logistics technology field mainly involves the overall planning of container loading, unloading, handling, warehousing, and circulation, as well as the transmission of operational instructions. Its core aspects include berth allocation, yard space planning, crane task assignment, and horizontal transport equipment path setting. The overall technical system relies on the underlying physical loading and unloading machinery and the upper-level information management system. It maintains the operation of port cargo circulation through business order parsing and hardware equipment status interaction. Traditional intelligent crane scheduling methods for container yards involve the process allocation and operational control of multiple rail-mounted cranes simultaneously performing container lifting and handling tasks within the container yard. This typically employs a combination of centralized instruction distribution and rigid physical isolation. The process involves the central control unit reading the container bill of lading business data and extracting the corresponding three-dimensional coordinates of the lifting position and the unloading position. Then, using a fixed wireless base station, the extracted coordinate set is sent sequentially to the on-board industrial control computer of the trolley within the designated work grid. After receiving the coordinates, the on-board industrial control computer directly sends a drive current of a specified frequency to the trolley travel inverter and the trolley travel winch to execute the displacement action. When there are multiple trolleys on the same running track, contact-type limit switches and polyurethane buffers are installed at the ends of the main steel beams of each trolley. When two trolleys move relative to each other and touch the limit switches, the three-phase AC power supply circuit of the travel motor is directly cut off to trigger the physical brake action.

[0003] Traditional yard crane scheduling typically relies on a combination of centralized command distribution and rigid physical isolation when multiple devices are performing handling tasks together. This operating mode relies on contact limit switches and buffers to forcibly cut off the power supply circuit of the travel motor at the critical point when the devices are in relative contact, triggering the physical brake action. This makes the equipment prone to physical deadlock and flow interruption in densely populated work areas. At the same time, the rigid brake causes huge impact losses to the lifting mechanism and wire rope, exacerbating the complex torsion of the container and creating safety hazards under complex operating conditions. Summary of the Invention

[0004] To achieve the above objectives, the present invention adopts the following technical solution: a yard intelligent crane scheduling method based on parallel cooperation, comprising the following steps:

[0005] S1: Compare the vertical height coordinate parameters of the corresponding items in the container lifting three-dimensional coordinate set to generate vertically stacked directed edges, calculate the difference in the expected lifting action time parameters of the near-distance items to generate horizontal interference directed edges, and integrate and output the scheduling interference topology map data.

[0006] S2: Calculate the number of nodes pointed to by each node in the scheduling interference topology graph data to generate the node in-degree value, take the zero value item to construct a concurrent execution task queue and issue it, extract the first spatial dwell coordinate parameter, the second spatial dwell coordinate parameter, the first vector running speed parameter and the second vector running speed parameter, and generate parallel driving status data;

[0007] S3: Based on the parallel vehicle operation status data, the displacement of the first vector running speed parameter and the first spatial dwelling coordinate parameter, and the second vector running speed parameter and the second spatial dwelling coordinate parameter are mapped to the three-dimensional vertex matrix of the first displacement envelope surface and the three-dimensional vertex matrix of the second displacement envelope surface, and the three-dimensional intersection volume is calculated to derive the intersection volume change parameter.

[0008] S4: Calculate the difference between the intersection volume change parameter and the preset physical intrusion tolerance value and map it into a deceleration braking level signal. Compare the specific values ​​of the first vector running speed parameter and the second vector running speed parameter, inject the deceleration braking level signal into the motor controller, and output the anti-collision avoidance drive command.

[0009] S5: In response to the collision avoidance drive command, the vehicle deceleration and avoidance action is triggered. The real-time swing angle value is extracted by comparing the pixel changes of the continuous deflection image frame and the swing angle acceleration value is derived differentially. The coupling deviation with the actual moving acceleration parameter of the vehicle is calculated to generate the nonlinear coupling disturbance term value. The equivalent compensation current value is derived and superimposed with the drive stator current. The coordinated scheduling correction control data is output.

[0010] As a further embodiment of the present invention, the scheduling interference topology map data includes a node connection matrix, an edge weight set, and a directed layer-level table; the parallel driving operation status data specifically refers to a real-time azimuth coordinate set, an instantaneous vector speedometer, and a concurrent operation timestamp; the intersection volume change parameters specifically include the intrusion volume derivative value, the boundary intersection expansion rate, and the interference space reduction degree; the collision avoidance drive command includes a frequency converter braking pulse signal, a deceleration smoothness level code, and a forced current limiting threshold; and the collaborative scheduling correction control data specifically refers to the torque feedforward compensation value, the sliding mode anti-disturbance gain coefficient, and the reverse stator excitation quantity.

[0011] As a further aspect of the present invention, the step of obtaining the scheduling interference topology map data specifically includes:

[0012] S101: During the container scheduling and transfer process in the port yard, collect the three-dimensional coordinate set of container lifting, extract the horizontal coordinate parameters, vertical coordinate parameters and vertical height coordinate parameters in the three-dimensional coordinate set of container lifting, and perform comparison calculations on the corresponding vertical height coordinate parameters when the horizontal coordinate parameters and the vertical coordinate parameters are consistent at the same time, and establish vertically stacked directed edges.

[0013] S102: Call the container lifting three-dimensional coordinate set, perform absolute straight-line distance calculation on multiple independent data in the container lifting three-dimensional coordinate set, obtain the absolute straight-line distance value, extract the expected lifting action time parameter for the corresponding item whose absolute straight-line distance value is less than the physical width setting value, perform time difference calculation on multiple expected lifting action time parameters, and establish a horizontal interference directed edge.

[0014] S103: Call the vertically stacked directed edge and the horizontally interfering directed edge, perform node association and integration operations on the vertically stacked directed edge and the horizontally interfering directed edge, and establish scheduling interference topology graph data.

[0015] As a further aspect of the present invention, the step of acquiring the parallel driving operation status data specifically includes:

[0016] S201: Extract the vertically stacked directed edges and horizontally interfering directed edges from the scheduling interference topology graph data, and perform statistical calculations on the number of container lifting three-dimensional coordinates pointed to by the vertically stacked directed edges and horizontally interfering directed edges to generate node in-degree values.

[0017] S202: Call the node in-degree value, filter the data items whose node in-degree value is equal to zero to construct a concurrent execution task queue, send the concurrent execution task queue to the parallel driving node, output the first spatial dwell coordinate parameter, the second spatial dwell coordinate parameter, the first vector running speed parameter and the second vector running speed parameter and integrate them to establish parallel driving status data.

[0018] As a further aspect of the present invention, the process of distributing the concurrent execution task queue to the parallel driving node, outputting the first spatial dwell coordinate parameters, the second spatial dwell coordinate parameters, the first vector running speed parameters, and the second vector running speed parameters, and integrating them, specifically comprises:

[0019] Parse the concurrent execution task queue to obtain the physical track code;

[0020] Based on the physical track encoding, the concurrent execution task queue is split into a first scheduling instruction and a second scheduling instruction;

[0021] The first scheduling instruction is sent to the first parallel train node, and the second scheduling instruction is sent to the second parallel train node;

[0022] The positioning coordinate feedback value and moving speed feedback value of the bottom controller of the first parallel trolley node are collected. The positioning coordinate feedback value is mapped to the three-dimensional stockpile position coordinate to establish the first spatial dwelling coordinate parameter. The moving speed feedback value is converted into a spatial direction speed vector to establish the first vector running speed parameter.

[0023] The positioning coordinate feedback value and moving speed feedback value of the bottom controller of the second parallel trolley node are collected. The positioning coordinate feedback value is mapped to the three-dimensional stockpile position coordinate to establish the second spatial dwelling coordinate parameter. The moving speed feedback value is converted into a spatial direction speed vector to establish the second vector running speed parameter.

[0024] Perform an association and splicing operation on the first spatial dwell coordinate parameter, the second spatial dwell coordinate parameter, the first vector running speed parameter, and the second vector running speed parameter to establish parallel driving status data.

[0025] As a further aspect of the present invention, the step of obtaining the intersection volume change parameter specifically includes:

[0026] S301: Call the parallel vehicle operation status data, collect the predicted time step value, perform forward displacement mapping operation on the first vector running speed parameter and the first spatial dwell coordinate parameter based on the predicted time step value, establish the three-dimensional vertex matrix of the first displacement envelope surface, perform forward displacement mapping operation on the second vector running speed parameter and the second spatial dwell coordinate parameter based on the predicted time step value, establish the three-dimensional vertex matrix of the second displacement envelope surface, and output the vertex matrix dataset;

[0027] S302: Based on the vertex matrix dataset, calculate the size of the spatial overlap region between the three-dimensional vertex matrix of the first displacement envelope surface and the three-dimensional vertex matrix of the second displacement envelope surface, and generate a three-dimensional intersection volume value.

[0028] S303: Call the three-dimensional intersecting volume values, perform volume change difference calculation on the three-dimensional intersecting volume values ​​within the continuous control cycle, and derive the intersecting volume change parameters.

[0029] As a further aspect of the present invention, the step of obtaining the collision avoidance drive command specifically includes:

[0030] S401: Obtain a preset physical intrusion tolerance value, call the intersection volume change parameter, perform an intersection value difference calculation on the intersection volume change parameter and the preset physical intrusion tolerance value, and obtain the intersection value difference.

[0031] S402: Call the intersecting numerical difference, acquire the reference level, and perform an inverse proportional mapping operation between the intersecting numerical difference and the reference level to obtain the deceleration braking level signal;

[0032] S403: Call the deceleration and braking level signal, the first vector running speed parameter and the second vector running speed parameter, compare the values ​​of the first vector running speed parameter and the second vector running speed parameter, inject the deceleration and braking level signal into the bottom-level walking motor controller with the highest speed value, and output the anti-collision and avoidance drive command.

[0033] As a further aspect of the present invention, the process of obtaining the preset physical intrusion tolerance value specifically includes:

[0034] Read the underlying crane equipment parameter configuration library to obtain the spreader buffer compression thickness parameter, container metal shell deformation depth parameter, braking safety reserve distance parameter and crane rated mass benchmark value;

[0035] The initial unidirectional intrusion depth value is established by performing an accumulation operation on the spreader buffer compression thickness parameter, the container metal shell deformation depth parameter, and the braking safety reserved distance parameter.

[0036] Extract the actual load mass parameters of the currently hoisted container from the hoisting task details list;

[0037] Calculate the ratio of the actual load mass parameter to the rated load reference value of the vehicle, and establish an inertial offset compensation coefficient;

[0038] The initial unidirectional intrusion depth value is multiplied by the inertial offset compensation coefficient to establish the maximum unidirectional intrusion depth correction value;

[0039] Read the three-dimensional dimension contour data of the container currently being hoisted, and extract the geometric cross-sectional area parameters of the intersecting contact surfaces within the three-dimensional dimension contour data;

[0040] The maximum unidirectional intrusion depth correction value is multiplied by the geometric cross-sectional area parameter of the intersection contact surface to generate the preset physical intrusion tolerance value.

[0041] As a further aspect of the present invention, the step of acquiring the collaborative scheduling correction control data specifically includes:

[0042] S501: In response to the collision avoidance drive command, the vehicle deceleration and avoidance action is triggered. During the movement of the physical trolley of the crane in the port yard, the continuous deflection image frames captured by the crane spreader machine vision sensor are collected. The pixel edge coordinate changes within the continuous deflection image frames are compared, and the real-time swing angle value is extracted.

[0043] S502: Perform a second differential operation on the real-time swing angle values ​​for multiple control cycles to derive the swing angle acceleration values, collect the actual moving acceleration parameters of the trolley, perform nonlinear coupling deviation calculation on the swing angle acceleration values ​​and the actual moving acceleration parameters of the trolley, and generate nonlinear coupling disturbance term values.

[0044] S503: Perform error convergence calculation on the value of the nonlinear coupling disturbance term, derive the equivalent compensation current value, collect the drive stator current of the trolley travel inverter, perform superposition calculation on the equivalent compensation current value and the drive stator current, and output the coordinated scheduling correction control data.

[0045] As a further aspect of the present invention, the process of performing nonlinear coupling deviation calculation on the swing angle acceleration value and the actual moving acceleration parameter of the trolley to generate a nonlinear coupling disturbance term value specifically involves:

[0046] Read the current sling length parameter from the vehicle status control library;

[0047] The angular acceleration value is calculated by multiplying the angular acceleration value with the current sling length parameter to establish the linear acceleration value of the swing.

[0048] A spatial cross product operation is performed on the swing linear acceleration value and the actual moving acceleration parameter of the vehicle to generate the nonlinear coupled disturbance term value.

[0049] The process of performing error convergence calculations on the numerical value of the nonlinear coupling disturbance term to derive the equivalent compensation current value is as follows:

[0050] Read the sliding mode gain configuration parameters, steady-state approach law coefficients, and torque-current conversion constants from the underlying trolley frequency converter controller.

[0051] Extract the absolute value characteristic variables and directional polarity identification data of the nonlinear coupling perturbation term;

[0052] Perform a multiplication calculation on the absolute value characteristic variable and the steady-state convergence law coefficient to establish a convergence compensation component;

[0053] The initial torque compensation component is established by performing an algebraic summation operation on the convergence compensation component and the sliding mode gain configuration parameter in combination with the directional polarity identification data.

[0054] The equivalent compensation current value is derived by multiplying the initial torque compensation component with the torque current conversion constant.

[0055] A parallel and collaborative intelligent yard traffic scheduling system includes:

[0056] The spreader interference analysis module compares the vertical height coordinate parameters of the corresponding terms in the container lifting three-dimensional coordinate set to generate vertically stacked directed edges, calculates the difference in the expected lifting action time parameters of the near-distance terms to generate horizontal interference directed edges, and integrates and outputs scheduling interference topology map data.

[0057] The concurrent task parsing module counts the number of nodes pointed to in the scheduling interference topology graph data to generate the node in-degree value, takes the zero value item to construct the concurrent execution task queue and distributes it, extracts the first spatial dwell coordinate parameter, the second spatial dwell coordinate parameter, the first vector running speed parameter and the second vector running speed parameter, and generates parallel vehicle running status data.

[0058] The envelope intersection deduction module, based on the parallel vehicle operation status data, maps the displacement of the first vector running speed parameter and the first spatial dwell coordinate parameter, and the second vector running speed parameter and the second spatial dwell coordinate parameter to the three-dimensional vertex matrix of the first displacement envelope surface and the three-dimensional vertex matrix of the second displacement envelope surface, and calculates the three-dimensional intersection volume value to derive the intersection volume change parameter;

[0059] The intrusion tolerance mapping module calculates the difference between the intersection volume change parameter and the preset physical intrusion tolerance value and maps it into a deceleration braking level signal. It compares the specific values ​​of the first vector running speed parameter and the second vector running speed parameter, injects the deceleration braking level signal into the motor controller, and outputs the anti-collision avoidance drive command.

[0060] The sway angle coupling compensation module responds to the vehicle deceleration and avoidance action triggered by the anti-collision avoidance drive command, compares the pixel changes of continuous deflection image frames to extract real-time sway angle values ​​and differentially derives sway angle acceleration values, calculates the coupling deviation with the actual moving acceleration parameters of the vehicle to generate nonlinear coupling disturbance term values, derives equivalent compensation current values ​​and superimposes them with the drive stator current, and outputs coordinated scheduling correction control data.

[0061] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0062] In this invention, the three-dimensional physical collision avoidance logic is mapped to a graph structure time priority topology scheduling mechanism to extract the concurrent task queue. Based on the time step, the three-dimensional intersection volume of the multi-machine displacement envelope is dynamically predicted. The difference in the intersection volume change is converted into a deceleration braking level signal and injected into the motor controller. The nonlinear coupling disturbance term is generated by combining the visual swing angle acceleration and the actual moving acceleration of the trolley to derive the compensation current and superimpose it on the inverter drive stator current. The hybrid topology stripping ensures that the concurrent tasks are immune to spatial conflicts. The intersection volume derivative calculation is used to realize the flexible sliding peak shifting of the equipment to replace the rigid brake. The potential energy of the off-center load sway is smoothed from the electrical bottom layer to ensure smooth handling. Attached Figure Description

[0063] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0064] Figure 1 This is a schematic diagram of the steps of the present invention;

[0065] Figure 2 This is a detailed schematic diagram of S1 of the present invention;

[0066] Figure 3 This is a detailed schematic diagram of S2 of the present invention;

[0067] Figure 4 This is a detailed schematic diagram of S3 of the present invention;

[0068] Figure 5 This is a detailed schematic diagram of S4 of the present invention;

[0069] Figure 6 This is a detailed schematic diagram of S5 of the present invention;

[0070] Figure 7 This is a system module diagram of the present invention. Detailed Implementation

[0071] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0072] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0073] Please see Figure 1 This invention provides a parallel collaborative intelligent yard traffic scheduling method, comprising the following steps:

[0074] S1: Compare the vertical height coordinate parameters of the corresponding items in the container lifting three-dimensional coordinate set to generate vertically stacked directed edges, calculate the difference in the expected lifting action time parameters of the near-distance items to generate horizontal interference directed edges, and integrate and output the scheduling interference topology map data.

[0075] S2: Statistically calculate the number of nodes pointed to by each node in the interference topology graph data to generate the node in-degree value, take the zero value to construct a concurrent execution task queue and issue it, extract the first spatial dwell coordinate parameter, the second spatial dwell coordinate parameter, the first vector running speed parameter and the second vector running speed parameter, and generate parallel vehicle running status data;

[0076] S3: Based on the parallel vehicle operation status data, the displacement of the first vector running speed parameter and the first spatial dwell coordinate parameter, and the second vector running speed parameter and the second spatial dwell coordinate parameter are mapped to the three-dimensional vertex matrix of the first displacement envelope surface and the three-dimensional vertex matrix of the second displacement envelope surface. The three-dimensional intersection volume is calculated and the intersection volume change parameter is derived.

[0077] S4: Calculate the difference between the intersection volume change parameter and the preset physical intrusion tolerance value and map it into a deceleration braking level signal. Compare the specific values ​​of the first vector running speed parameter and the second vector running speed parameter, inject the deceleration braking level signal into the motor controller, and output the anti-collision avoidance drive command.

[0078] S5: Responds to the collision avoidance drive command, compares the pixel changes of continuous deflection image frames to extract the real-time swing angle value and differentially derives the swing angle acceleration value, calculates the coupling deviation with the actual moving acceleration parameter of the vehicle to generate the nonlinear coupling disturbance term value, derives the equivalent compensation current value and superimposes it with the drive stator current, and outputs the coordinated scheduling correction control data.

[0079] The scheduling interference topology data includes node connection matrix, edge weight set, directed layer level table, parallel driving operation status data specifically refers to real-time orientation coordinate set, instantaneous vector speed table, operation concurrency timestamp, intersection volume change parameters specifically refer to intrusion volume derivative value, boundary intersection expansion rate, interference space reduction degree, collision avoidance drive command includes inverter braking pulse signal, deceleration smoothness level code, forced current limiting threshold, and collaborative scheduling correction control data specifically refer to torque feedforward compensation value, sliding mode disturbance rejection gain coefficient, and reverse stator excitation quantity.

[0080] Please see Figure 2 The specific steps for obtaining the scheduling interference topology map data are as follows:

[0081] S101: During the container scheduling and transfer process in the port yard, collect the three-dimensional coordinate set of container lifting, extract the horizontal coordinate parameters, vertical coordinate parameters and vertical height coordinate parameters in the three-dimensional coordinate set of container lifting, and perform comparison calculations on the corresponding vertical height coordinate parameters when the horizontal coordinate parameters and the vertical coordinate parameters are consistent at the same time, and establish vertically stacked directed edges.

[0082] Based on high-precision real-time dynamic differential positioning base stations deployed in the port yard area and displacement encoders at the spreader heads of container cranes, spatial positioning data during the container scheduling and transfer process in the port yard is continuously collected at a sampling frequency of 100 Hz. This constructs a 3D coordinate set for container lifting, including the unique identification code of each container and its corresponding spatial 3D coordinate data. The acquired 3D coordinate set undergoes data cleaning and outlier removal, filtering out invalid coordinate points with positioning error drift exceeding 50 mm. The lateral, longitudinal, and vertical height coordinate parameters corresponding to each container in the cleaned coordinate set are extracted. A coordinate overlap tolerance threshold of 30 mm is set, and all data items in the coordinate set are traversed. The lateral and longitudinal coordinate parameters of any two containers are compared. When the absolute values ​​of the differences in the lateral and longitudinal coordinate parameters of two containers are simultaneously less than or equal to 30 mm, the two containers are confirmed to be in the same stack column. Subsequently, the vertical height coordinate parameters of these two containers are extracted and compared to determine their hierarchical relationship in vertical space. The container with the larger vertical height coordinate parameter is defined as the source node, and the container with the smaller vertical height coordinate parameter is defined as the target node. A unidirectional association record from the source node to the target node is created in the graph database. This establishes a vertical stacked directed edge representing the physical obstruction relationship of the vertical stack, indicating that the upper high-level container must be lifted and moved before the lower low-level container below.

[0083] Table 1. Example of a 3D coordinate set for container lifting.

[0084]

[0085] Table 1 lists sample data of the three-dimensional coordinate set for container lifting. The horizontal and vertical coordinates of identification codes 10001 and 10002 are consistent, and a clear vertical pointing relationship can be established by comparing the height parameters.

[0086] S102: Call the container lifting three-dimensional coordinate set, perform absolute straight-line distance calculation on multiple independent data in the container lifting three-dimensional coordinate set, obtain the absolute straight-line distance value, extract the expected lifting action time parameter for the corresponding item whose absolute straight-line distance value is less than the physical width setting value, perform time difference calculation on multiple expected lifting action time parameters, and establish a horizontal interference directed edge.

[0087] The system retrieves the complete 3D coordinate set for container lifting from memory and extracts any two independent data points representing different containers using nested loops. It then reads the lateral, longitudinal, and vertical height coordinates of these two containers, calculates the sum of squares of the differences in these corresponding dimensions, and takes the square root of this sum to determine the straight-line distance between the two containers in 3D physical space. A safe physical width of 3200 mm is set for crane operations, and the calculated straight-line distance is compared to this value. For data pairs with a straight-line distance less than 3200 mm, a risk of horizontal collision interference between the spreader or container during the same lifting period is identified. For pairs with a collision interference risk, the vertical height coordinates of the two containers are extracted. The rated lifting speed constant of the hoisting mechanism motor is set to 0.5 m / s. The vertical height coordinates, initially in millimeters, are converted to meters and then divided by the rated lifting speed constant to derive the expected lifting time. For two containers with horizontal collision interference risk, the time difference is calculated by subtracting the expected lifting action time parameters. The polarity of the time difference is determined, and the container with the larger expected lifting action time parameter is defined as the source node, while the container with the smaller expected lifting action time parameter is defined as the target node. A one-way pointer from the high-time-consuming node to the low-time-consuming node is written into the graph memory area, thus establishing a directed edge representing the order of horizontal interference avoidance. This ensures that time-consuming operations are executed first to optimize parallel efficiency.

[0088] S103: Call the vertically stacked directed edges and the horizontally interfering directed edges, perform node association and integration operations on the vertically stacked directed edges and the horizontally interfering directed edges, and establish scheduling interference topology graph data;

[0089] The system retrieves all vertically stacked directed edges and all horizontally interfering directed edges already generated in the storage medium. It initializes a full hollow directed graph structure in memory, with container identification codes as vertices. The unidirectional pointer relationships contained in the vertically stacked directed edges are written one by one into the edge set of this hollow directed graph structure, mapping the vertical job dependency constraints. Simultaneously, the unidirectional pointer relationships contained in the horizontally interfering directed edges are appended to the same edge set, mapping the horizontal peak avoidance constraints. A depth-first search operation is performed on the written edge set to detect the existence of closed loops in the graph. If a closed loop caused by conflicting horizontal and vertical constraints is detected, the connection of the horizontally interfering directed edges in the loop is automatically broken, while the hard physical constraints of the vertically stacked directed edges are forcibly preserved. After completing the graph structure loop elimination and node association integration operation, the graph structure containing the final vertex topology mapping and all edge weight relationships is serialized, fully packaged according to the standard data exchange format, and a scheduling interference topology graph data that can accurately reflect the collision avoidance dependency logic of the three-dimensional space of the storage yard is established. This data is then stored in the high-speed cache of the scheduling control terminal for dispatch and circulation.

[0090] Please see Figure 3 The specific steps for obtaining parallel vehicle operation status data are as follows:

[0091] S201: Extract the vertically stacked directed edges and horizontally interfering directed edges from the scheduling interference topology graph data, perform statistical calculations on the number of container lifting 3D coordinates pointed to by the vertically stacked directed edges and horizontally interfering directed edges, and generate node in-degree values.

[0092] Extract scheduling interference topology graph data from the control terminal's cache, parse the graph data, and separate the edge set containing all associated mappings. Traverse this edge set, extracting vertically stacked directed edges representing vertical obstruction and horizontally interfering directed edges representing horizontal collision risk. Initialize a one-dimensional integer array in memory with a length equal to the total number of containers to be lifted, where the array indices correspond one-to-one with the node indices of the container lifting 3D coordinates, and initialize all elements to 0. Perform an accumulation and statistical calculation on the number of container lifting 3D coordinates pointed to by vertically stacked directed edges and horizontally interfering directed edges as target nodes in the topology graph. Specifically, whenever a directed edge arrow is detected pointing to a container node, increment the value at the corresponding index in the one-dimensional array by 1. After traversing all directed edges, end the accumulation operation; the non-negative integer values ​​stored in the one-dimensional array represent the node in-degree value representing the pre-obstruction conditions for each container lifting task.

[0093] S202: Call the node in-degree value, filter the data items whose node in-degree value is equal to zero, construct the concurrent execution task queue, send the concurrent execution task queue to the parallel driving node, output the first spatial dwell coordinate parameter, the second spatial dwell coordinate parameter, the first vector running speed parameter and the second vector running speed parameter and integrate them to establish the parallel driving status data.

[0094] A full scan and comparison operation is performed using the node in-degree values ​​stored in a one-dimensional array in memory. A judgment benchmark of 0 is established, and data items with in-degree values ​​equal to zero are filtered out using conditional statements. These zero-degree data items physically indicate that there are currently no upper container covers and no horizontal interference conflicts in the surrounding area, thus possessing reliable lifting conditions. All filtered zero-degree data items are packaged and extracted to construct a first-in-first-out concurrent execution task queue. The concurrent execution task queue is parsed to obtain the physical track code; based on the physical track code, the concurrent execution task queue is split into a first scheduling instruction and a second scheduling instruction; the first scheduling instruction is sent to the first parallel trolley node and the second scheduling instruction is sent to the second parallel trolley node via the industrial communication bus. The positioning coordinate feedback value and movement speed feedback value of the underlying controller of the first parallel trolley node are collected. The positioning coordinate feedback value is mapped to three-dimensional yard position coordinates to establish the first spatial dwelling coordinate parameter, and the movement speed feedback value is converted into a spatial direction velocity vector to establish the first vector running speed parameter; similarly, the corresponding feedback value of the underlying controller of the second parallel trolley node is collected and mapped to establish the second spatial dwelling coordinate parameter and the second vector running speed parameter. The first spatial dwell coordinate parameters, the second spatial dwell coordinate parameters, the first vector running speed parameters, and the second vector running speed parameters are associated and spliced. Field alignment and data frame packet integration are performed according to the predetermined timestamp to establish parallel driving status data that characterizes the physical motion benchmark of the two vehicles.

[0095] Please see Figure 4 The specific steps for obtaining the intersection volume change parameters are as follows:

[0096] S301: Call the parallel vehicle operation status data, collect the predicted time step value, perform forward displacement mapping operation on the first vector running speed parameter and the first spatial dwell coordinate parameter based on the predicted time step value, establish the three-dimensional vertex matrix of the first displacement envelope surface, perform forward displacement mapping operation on the second vector running speed parameter and the second spatial dwell coordinate parameter based on the predicted time step value, establish the three-dimensional vertex matrix of the second displacement envelope surface, and output the vertex matrix dataset;

[0097] The system retrieves real-time generated parallel trolley operation status data and extracts the first vector running speed parameter, the first spatial dwell coordinate parameter, the second vector running speed parameter, and the second spatial dwell coordinate parameter. A predetermined prediction time step of 3000 milliseconds is collected. Based on the prediction time step, a scalar multiplication operation is performed on the first vector running speed parameter in the three-dimensional spatial vector domain to calculate the three-dimensional spatial displacement increment of the first trolley within the next 3000 milliseconds. This displacement increment is then vector-added with the first spatial dwell coordinate parameter to obtain the predicted coordinates of the first trolley's future center point. Using the physical dimensions of a standard container and spreader (length, width, height set to 12192 mm, width 2438 mm, height 2591 mm), eight outer boundary points are generated outwards from the predicted center point coordinates to establish a three-dimensional vertex matrix of the first displacement envelope surface that encloses the future motion limit boundary of the first trolley. Using the same displacement increment multiplication and vector addition forward displacement mapping operation logic, the second vector running speed parameter and the second spatial dwell coordinate parameter are calculated based on the predicted time step value. A three-dimensional vertex matrix of the second displacement envelope surface that wraps the limit boundary of the future motion of the second vehicle is established. The two sets of matrices are integrated and output as a vertex matrix dataset.

[0098] S302: Based on the vertex matrix dataset, calculate the size of the spatial overlap region between the three-dimensional vertex matrix of the first displacement envelope surface and the three-dimensional vertex matrix of the second displacement envelope surface, and generate a three-dimensional intersection volume value.

[0099] Based on the output vertex matrix dataset, two three-dimensional spatial geometries defined by the three-dimensional vertex matrices of the first and second displacement envelope surfaces are extracted. Using the three-dimensional spatial separation axis theorem and Boolean intersection logic, the geometric volume is calculated for the overlapping region of these two three-dimensional vertex matrices in the spatial coordinate system. The minimum and maximum boundary intersection lengths of the two cuboids along the horizontal, vertical, and coordinate axes are calculated respectively. If the intersection length along any coordinate axis is less than or equal to 0, no spatial overlap is determined, and the output volume is 0. If the intersection length along all three coordinate axes is greater than 0, the horizontal, vertical, and coordinate overlap lengths are multiplied together to generate a three-dimensional intersection volume value representing the severity of potential future collision intrusion between the two vehicles. This value, measured in cubic millimeters, directly reflects the objective equivalent of spatial conflict.

[0100] S303: Call the three-dimensional intersecting volume values, perform volume change difference calculation on the three-dimensional intersecting volume values ​​within the continuous control cycle, and derive the intersecting volume change parameters;

[0101] The system invokes the 3D intersection volume value output from the previous calculation step to initiate a time-series-based dynamic change monitoring mechanism. A circular queue register of length 2 is created in the underlying microcontroller to record the 3D intersection volume values ​​within two consecutive 100-millisecond control cycles, labeled as the current cycle volume value and the previous cycle volume value, respectively. A subtraction operation is performed on the 3D intersection volume values ​​within consecutive control cycles, subtracting the previous cycle volume value from the current cycle volume value to obtain the volume change difference. This volume change difference is used as an intersection volume change parameter characterizing the shrinkage trend of the collision space between the two vehicles. Its dimension is consistent with the 3D intersection volume value, uniformly measured in cubic millimeters. A positive value indicates that the collision overlap volume is expanding, and its magnitude (i.e., the intrusion volume derivative) reflects the rate of increase in collision risk. The current cycle's 3D intersection volume value (absolute quantity) is used for direct comparison with the downstream preset physical intrusion tolerance value, providing a feedforward benchmark for accurately formulating flexible deceleration strategies to determine the risk level.

[0102] Please see Figure 5 The specific steps for obtaining collision avoidance drive commands are as follows:

[0103] S401: Obtain the preset physical intrusion tolerance value, call the intersection volume change parameter, perform the intersection value difference calculation on the intersection volume change parameter and the preset physical intrusion tolerance value, and obtain the intersection value difference.

[0104] The system reads the underlying crane equipment parameter configuration library, retrieves the spreader buffer compression thickness parameter and assigns it a value of 200 mm, the container metal shell deformation depth parameter and assigns it a value of 50 mm, the braking safety reserve distance parameter and assigns it a value of 150 mm, and the crane's rated mass baseline value and assigns it a value of 40,000 kg. It then performs an additive operation on the spreader buffer compression thickness parameter, container metal shell deformation depth parameter, and braking safety reserve distance parameter (200 + 50 + 150) to establish an initial unidirectional intrusion depth value of 400 mm. The system extracts the actual load mass parameter of the currently being lifted container from the lifting task details list and assigns it a value of 30,000 kg. It calculates the ratio of the actual load mass parameter to the crane's rated mass baseline value (30,000 divided by 40,000, yielding 0.75) to establish an inertial offset compensation coefficient. Finally, it multiplies the initial unidirectional intrusion depth value of 400 mm with the inertial offset compensation coefficient of 0.75 to establish a maximum unidirectional intrusion depth correction value of 300 mm. The system reads the 3D dimensional contour data of the currently being hoisted container, extracts the geometric cross-sectional area parameter of the intersecting contact surface, and obtains this parameter as 6,316,858 square millimeters by multiplying the container's cross-sectional width of 2438 mm and height of 2591 mm. It then performs a product operation on the maximum unidirectional intrusion depth correction value of 300 mm and the geometric cross-sectional area parameter of 6,316,858 square millimeters to generate a preset physical intrusion tolerance value of 1,895,057,400 cubic millimeters. The system then calls the 3D intersection volume value derived from the microcontroller's memory for the current period and performs a subtraction operation between this 3D intersection volume value and the preset physical intrusion tolerance value. Subtracting the preset physical intrusion tolerance value from the 3D intersection volume value yields a difference in intersection values ​​representing the degree of over-limit danger if the result is greater than 0; otherwise, the value is forcibly reset to zero.

[0105] S402: Call the intersecting numerical difference, acquire the reference level, and perform an inverse proportional mapping operation between the intersecting numerical difference and the reference level to obtain the deceleration braking level signal;

[0106] The system retrieves the full-load reference level on the drive side via the inverter's analog-to-digital conversion channel using a cross-sectional area difference (CBD) greater than 0. This reference level is calibrated as a 10,000 mV rated DC voltage signal. A nonlinear inverse proportional mapping operation is then performed between the CBD and the reference level. Specifically, the operation involves: matching the corresponding inverse proportional mapping coefficient based on the preset interval into which the CBD falls; using this coefficient as the denominator adjustment factor, adding it to the integer 1 to form the closed-loop feedback denominator; and then dividing the 10,000 mV reference level by this denominator. The higher the interval the CBD falls into, the larger the obtained inverse proportional mapping coefficient, resulting in an amplified closed-loop feedback denominator and a synchronously decreasing output deceleration braking level signal. This produces a deceleration braking level signal that suppresses vehicle speed. This signal value is inversely proportional to the severity of the collision; the higher the severity, the lower the output level signal, thus suppressing swaying caused by rigid mechanical brakes.

[0107] Table 2 Deceleration and Braking Level Mapping Control Parameter Table

[0108]

[0109] Table 2 lists the level mapping calculation results corresponding to the intersection value difference in three different intervals. For example, when the intersection value difference is 1000, the feedback denominator generated by the coefficient operation is 1.05, and the reference level divided by 1.05 results in a smooth output deceleration braking level signal of 9523 millivolts.

[0110] S403: Calls the deceleration braking level signal, the first vector running speed parameter and the second vector running speed parameter, compares the values ​​of the first vector running speed parameter and the second vector running speed parameter, injects the deceleration braking level signal into the bottom-level walking motor controller with the highest speed value, and outputs the anti-collision avoidance drive command;

[0111] The system invokes the deceleration and braking level signal, the first vector running speed parameter, and the second vector running speed parameter generated by the underlying control loop. It calculates the magnitude of the first and second vector running speed parameters in the three-dimensional coordinate system, i.e., performs a square root operation on the sum of the squares of the velocity components in each direction to obtain the scalar rate. It then compares the values ​​of the scalar rates corresponding to the first and second vector running speed parameters. The system locks onto the target crane node with the highest scalar rate value. Through the fieldbus control protocol, the generated deceleration and braking level signal is precisely injected into the analog input register of the underlying travel motor controller of the crane with the highest speed value. By reducing the stator voltage-to-frequency ratio of the target motor, the high-speed crane is prompted to perform a flexible deceleration avoidance operation, ultimately outputting an anti-collision avoidance drive command containing level instructions and device address codes.

[0112] Please see Figure 6 The specific steps for obtaining coordinated scheduling correction control data are as follows:

[0113] S501: Responds to the deceleration and avoidance action triggered by the collision avoidance drive command. During the movement of the physical trolley in the port yard, it collects continuous deflection image frames captured by the machine vision sensor of the crane spreader, compares the pixel edge coordinate changes within the continuous deflection image frames, and extracts the real-time swing angle value.

[0114] In response to the alternating deceleration and acceleration / deceleration of the gantry crane triggered by collision avoidance drive commands, a high-frame-rate machine vision sensor mounted vertically downwards at the bottom of the crane trolley is activated during the physical movement of the gantry crane in the port yard. Its sampling rate is set to 120 frames per second. Continuous image frames of the crane's spreader and the load-bearing wire rope deflection are acquired. Grayscale conversion and edge high-pass filtering enhancement are performed on the continuous image frames. The Hough line detection algorithm is used to accurately locate the pixel contour edges of the wire rope in the two-dimensional image plane. The specific displacement changes of the wire rope pixel edge coordinates in the current deflection image frame are compared with those in the previous time reference frame. Based on the camera's fixed focal length parameters and calibration matrix, the changes in the two-dimensional pixel edge coordinates are converted into physical angle values ​​through trigonometric geometric mapping relationships, extracting and generating real-time swing angle values ​​with an accuracy of 0.1 degrees.

[0115] S502: Perform a second differential calculation on the real-time swing angle values ​​for multiple control cycles to derive the swing angle acceleration values, collect the actual moving acceleration parameters of the trolley, perform nonlinear coupling deviation calculation on the swing angle acceleration values ​​and the actual moving acceleration parameters of the trolley, and generate nonlinear coupling disturbance term values.

[0116] A second-order difference operation is performed on the real-time swing angle values ​​within three consecutive control cycles in the buffer memory, with each control cycle fixed at 150 milliseconds. The angular velocity is obtained by subtracting the real-time swing angle values ​​from adjacent cycles and dividing by 150 milliseconds. Then, the angular velocities from adjacent cycles are subtracted and divided again by 150 milliseconds to derive the swing angular acceleration value reflecting the swing trend. The current sling length parameter issued by the hoisting drum is read from the crane status control library via the encoder communication interface and assigned a value of 15000 mm. A scalar multiplication calculation is performed between the swing angular acceleration value and the current sling length parameter of 15000 mm to establish the swing linear acceleration value that converts angular acceleration into physical linear displacement acceleration. The actual movement acceleration parameters of the crane are collected by the crane's onboard inertial measurement unit. Based on the dynamic model of a pendulum with a suspended load, the actual acceleration parameters of the trolley are used as external excitations. The coupling relationship between the trolley and the linear acceleration is analyzed, and the nonlinear inertial force caused by the acceleration and deceleration of the trolley on the load is solved. The nonlinear coupling disturbance term reflecting the nonlinear swaying kinetic energy of the load is generated.

[0117] S503: Performs error convergence calculation on the nonlinear coupling disturbance term, derives the equivalent compensation current value, collects the drive stator current of the trolley travel inverter, performs superposition calculation on the equivalent compensation current value and the drive stator current, and outputs coordinated scheduling correction control data.

[0118] Error convergence calculation based on sliding mode variable structure control theory is initiated for the numerical nonlinear coupled disturbance term. The sliding mode gain configuration parameter is read from the register of the underlying trolley inverter controller via the internal communication bus and assigned a value of 15, a steady-state reaching law coefficient of 0.80, and a motor-specific torque-current conversion constant of 2.50 amperes per Newton-meter. The absolute value characteristic variables of the nonlinear coupled disturbance term and the positive and negative polarity identification data for characterizing the sway direction are extracted. Multiplication is performed on the absolute value characteristic variables and the steady-state reaching law coefficient to establish a convergence compensation component to guide the control trajectory close to the sliding surface. Combining the polarity identification data, an algebraic summation operation with directional attributes is performed on the convergence compensation component and the sliding mode gain configuration parameter to establish an initial torque compensation component to eliminate sway kinetic energy. A constant product operation is performed on the initial torque compensation component and the torque-current conversion constant of 2.50 amperes per Newton-meter to derive the equivalent compensation current value in amperes. The target value of the base drive stator current currently output by the trolley travel inverter is collected. An electrical-level superposition calculation is performed between the derived equivalent compensation current value and this target drive stator current value to generate a new total target current command. This total target current command is then sent to the inverter's current closed-loop regulator, which dynamically generates a corresponding pulse width modulation (PWM) signal and drives the inverter's power switching transistors, outputting coordinated scheduling and correction control data to smooth out swaying.

[0119] Table 3. Sliding Mode Disturbance Immunity Control Loop Parameter Configuration Table

[0120]

[0121] Table 3 shows the various disturbance rejection configuration parameters under different control state variables. According to the calculation logic, when in the steady-state operating range and assuming the absolute value of the nonlinear coupling disturbance is 10, multiplying by the approach coefficient 0.80 yields a component of 8. After accumulating the gain of 15, the initial torque is 23. Multiplying by the current conversion constant 2.50 yields the accurate compensation current of 57.5 amperes. The overall logic strictly follows the electromagnetic torque derivation law.

[0122] Please see Figure 7 A parallel collaborative intelligent yard traffic scheduling system includes:

[0123] The spreader interference analysis module compares the vertical height coordinate parameters of the corresponding terms in the container lifting three-dimensional coordinate set to generate vertically stacked directed edges, calculates the difference in the expected lifting action time parameters of the near-distance terms to generate horizontal interference directed edges, and integrates and outputs scheduling interference topology map data.

[0124] The concurrent task parsing module counts the number of nodes pointed to in the scheduling interference topology data to generate the node in-degree value, takes the zero value item to construct the concurrent execution task queue and distributes it, extracts the first spatial dwell coordinate parameter, the second spatial dwell coordinate parameter, the first vector running speed parameter and the second vector running speed parameter, and generates parallel vehicle running status data.

[0125] The envelope intersection deduction module, based on the parallel vehicle operation status data, maps the displacement of the first vector running speed parameter and the first spatial dwell coordinate parameter, and the second vector running speed parameter and the second spatial dwell coordinate parameter to the three-dimensional vertex matrix of the first displacement envelope surface and the three-dimensional vertex matrix of the second displacement envelope surface, and calculates the three-dimensional intersection volume to derive the intersection volume change parameters.

[0126] The intrusion tolerance mapping module calculates the difference between the intersection volume change parameter and the preset physical intrusion tolerance value and maps it into a deceleration braking level signal. It compares the specific values ​​of the first vector running speed parameter and the second vector running speed parameter, injects the deceleration braking level signal into the motor controller, and outputs the anti-collision avoidance drive command.

[0127] The sway angle coupling compensation module responds to the collision avoidance drive command, compares the pixel changes of continuous deflection image frames to extract the real-time sway angle value and differentially derives the sway angle acceleration value, calculates the coupling deviation with the actual moving acceleration parameter of the vehicle to generate the nonlinear coupling disturbance term value, derives the equivalent compensation current value and superimposes it with the drive stator current, and outputs the coordinated scheduling correction control data.

[0128] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A yard intelligent crane scheduling method based on parallel cooperation, characterized in that, Includes the following steps: S1: Compare the vertical height coordinate parameters of the corresponding items in the container lifting three-dimensional coordinate set to generate vertically stacked directed edges, calculate the difference in the expected lifting action time parameters of the near-distance items to generate horizontal interference directed edges, and integrate and output the scheduling interference topology map data. S2: Calculate the number of nodes pointed to by each node in the scheduling interference topology graph data to generate the node in-degree value, take the zero value item to construct a concurrent execution task queue and issue it, extract the first spatial dwell coordinate parameter, the second spatial dwell coordinate parameter, the first vector running speed parameter and the second vector running speed parameter, and generate parallel driving status data; S3: Based on the parallel vehicle operation status data, the displacement of the first vector running speed parameter and the first spatial dwelling coordinate parameter, and the second vector running speed parameter and the second spatial dwelling coordinate parameter are mapped to the three-dimensional vertex matrix of the first displacement envelope surface and the three-dimensional vertex matrix of the second displacement envelope surface, and the three-dimensional intersection volume is calculated to derive the intersection volume change parameter. S4: Calculate the difference between the intersection volume change parameter and the preset physical intrusion tolerance value and map it into a deceleration braking level signal. Compare the specific values ​​of the first vector running speed parameter and the second vector running speed parameter, inject the deceleration braking level signal into the motor controller, and output the anti-collision avoidance drive command. The specific steps for obtaining the collision avoidance drive command are as follows: S401: Obtain a preset physical intrusion tolerance value, call the intersection volume change parameter, perform an intersection value difference calculation on the intersection volume change parameter and the preset physical intrusion tolerance value, and obtain the intersection value difference. S402: Call the intersecting numerical difference, acquire the reference level, and perform an inverse proportional mapping operation between the intersecting numerical difference and the reference level to obtain the deceleration braking level signal; S403: Call the deceleration braking level signal, the first vector running speed parameter and the second vector running speed parameter, compare the values ​​of the first vector running speed parameter and the second vector running speed parameter, inject the deceleration braking level signal into the bottom-level walking motor controller with the highest speed value, and output the anti-collision avoidance drive command. S5: In response to the collision avoidance drive command, the vehicle deceleration and avoidance action is triggered. The real-time swing angle value is extracted by comparing the pixel changes of the continuous deflection image frame and the swing angle acceleration value is derived differentially. The coupling deviation with the actual moving acceleration parameter of the vehicle is calculated to generate the nonlinear coupling disturbance term value. The equivalent compensation current value is derived and superimposed with the drive stator current. The coordinated scheduling correction control data is output.

2. The intelligent yard traffic scheduling method based on parallel cooperation according to claim 1, characterized in that, The scheduling interference topology data includes a node connection matrix, an edge weight set, and a directed layer-level table. The parallel driving status data specifically refers to a real-time azimuth coordinate set, an instantaneous vector speedometer, and a concurrent operation timestamp. The intersection volume change parameters specifically include the intrusion volume derivative value, the boundary intersection expansion rate, and the interference space reduction degree. The collision avoidance drive commands include the inverter braking pulse signal, the deceleration smoothness level code, and the forced current limiting threshold. The collaborative scheduling correction control data specifically refers to the torque feedforward compensation value, the sliding mode disturbance rejection gain coefficient, and the reverse stator excitation quantity.

3. The intelligent yard traffic scheduling method based on parallel cooperation according to claim 1, characterized in that, The specific steps for obtaining the scheduling interference topology map data are as follows: S101: During the container scheduling and transfer process in the port yard, collect the three-dimensional coordinate set of container lifting, extract the horizontal coordinate parameters, vertical coordinate parameters and vertical height coordinate parameters in the three-dimensional coordinate set of container lifting, and perform comparison calculations on the corresponding vertical height coordinate parameters when the horizontal coordinate parameters and the vertical coordinate parameters are consistent at the same time, and establish vertically stacked directed edges. S102: Call the container lifting three-dimensional coordinate set, perform absolute straight-line distance calculation on multiple independent data in the container lifting three-dimensional coordinate set, obtain the absolute straight-line distance value, extract the expected lifting action time parameter for the corresponding item whose absolute straight-line distance value is less than the physical width setting value, perform time difference calculation on multiple expected lifting action time parameters, and establish a horizontal interference directed edge. S103: Call the vertically stacked directed edge and the horizontally interfering directed edge, perform node association and integration operations on the vertically stacked directed edge and the horizontally interfering directed edge, and establish scheduling interference topology graph data.

4. The intelligent yard traffic scheduling method based on parallel cooperation according to claim 3, characterized in that, The specific steps for obtaining the parallel driving operation status data are as follows: S201: Extract the vertically stacked directed edges and horizontally interfering directed edges from the scheduling interference topology graph data, and perform statistical calculations on the number of container lifting three-dimensional coordinates pointed to by the vertically stacked directed edges and horizontally interfering directed edges to generate node in-degree values. S202: Call the node in-degree value, filter the data items whose node in-degree value is equal to zero to construct a concurrent execution task queue, send the concurrent execution task queue to the parallel driving node, output the first spatial dwell coordinate parameter, the second spatial dwell coordinate parameter, the first vector running speed parameter and the second vector running speed parameter and integrate them to establish parallel driving status data.

5. The intelligent yard traffic scheduling method based on parallel cooperation according to claim 4, characterized in that, The process of distributing the concurrent execution task queue to the parallel driving nodes, outputting the first spatial dwell coordinate parameters, the second spatial dwell coordinate parameters, the first vector running speed parameters, and the second vector running speed parameters, and integrating them, is specifically as follows: Parse the concurrent execution task queue to obtain the physical track code; Based on the physical track encoding, the concurrent execution task queue is split into a first scheduling instruction and a second scheduling instruction; The first scheduling instruction is sent to the first parallel train node, and the second scheduling instruction is sent to the second parallel train node; The positioning coordinate feedback value and moving speed feedback value of the bottom controller of the first parallel trolley node are collected. The positioning coordinate feedback value is mapped to the three-dimensional stockpile position coordinate to establish the first spatial dwelling coordinate parameter. The moving speed feedback value is converted into a spatial direction speed vector to establish the first vector running speed parameter. The positioning coordinate feedback value and moving speed feedback value of the bottom controller of the second parallel trolley node are collected. The positioning coordinate feedback value is mapped to the three-dimensional stockpile position coordinate to establish the second spatial dwelling coordinate parameter. The moving speed feedback value is converted into a spatial direction speed vector to establish the second vector running speed parameter. Perform an association and splicing operation on the first spatial dwell coordinate parameter, the second spatial dwell coordinate parameter, the first vector running speed parameter, and the second vector running speed parameter to establish parallel driving status data.

6. The intelligent yard traffic scheduling method based on parallel cooperation according to claim 4, characterized in that, The specific steps for obtaining the intersection volume change parameters are as follows: S301: Call the parallel vehicle operation status data, collect the predicted time step value, perform forward displacement mapping operation on the first vector running speed parameter and the first spatial dwell coordinate parameter based on the predicted time step value, establish the three-dimensional vertex matrix of the first displacement envelope surface, perform forward displacement mapping operation on the second vector running speed parameter and the second spatial dwell coordinate parameter based on the predicted time step value, establish the three-dimensional vertex matrix of the second displacement envelope surface, and output the vertex matrix dataset; S302: Based on the vertex matrix dataset, calculate the size of the spatial overlap region between the three-dimensional vertex matrix of the first displacement envelope surface and the three-dimensional vertex matrix of the second displacement envelope surface, and generate a three-dimensional intersection volume value. S303: Call the three-dimensional intersecting volume values, perform volume change difference calculation on the three-dimensional intersecting volume values ​​within the continuous control cycle, and derive the intersecting volume change parameters.

7. The intelligent yard traffic scheduling method based on parallel cooperation according to claim 6, characterized in that, The process of obtaining the preset physical intrusion tolerance value is as follows: Read the underlying crane equipment parameter configuration library to obtain the spreader buffer compression thickness parameter, container metal shell deformation depth parameter, braking safety reserve distance parameter and crane rated mass benchmark value; The initial unidirectional intrusion depth value is established by performing an accumulation operation on the spreader buffer compression thickness parameter, the container metal shell deformation depth parameter, and the braking safety reserved distance parameter. Extract the actual load mass parameters of the currently hoisted container from the hoisting task details list; Calculate the ratio of the actual load mass parameter to the rated load reference value of the vehicle, and establish an inertial offset compensation coefficient; The initial unidirectional intrusion depth value is multiplied by the inertial offset compensation coefficient to establish the maximum unidirectional intrusion depth correction value; Read the three-dimensional dimension contour data of the container currently being hoisted, and extract the geometric cross-sectional area parameters of the intersecting contact surfaces within the three-dimensional dimension contour data; The maximum unidirectional intrusion depth correction value is multiplied by the geometric cross-sectional area parameter of the intersection contact surface to generate the preset physical intrusion tolerance value; The specific steps for acquiring the coordinated scheduling correction control data are as follows: S501: In response to the collision avoidance drive command, the vehicle deceleration and avoidance action is triggered. During the movement of the physical trolley of the crane in the port yard, the continuous deflection image frames captured by the crane spreader machine vision sensor are collected. The pixel edge coordinate changes within the continuous deflection image frames are compared, and the real-time swing angle value is extracted. S502: Perform a second differential operation on the real-time swing angle values ​​for multiple control cycles to derive the swing angle acceleration values, collect the actual moving acceleration parameters of the trolley, perform nonlinear coupling deviation calculation on the swing angle acceleration values ​​and the actual moving acceleration parameters of the trolley, and generate nonlinear coupling disturbance term values. S503: Perform error convergence calculation on the value of the nonlinear coupling disturbance term, derive the equivalent compensation current value, collect the drive stator current of the trolley travel inverter, perform superposition calculation on the equivalent compensation current value and the drive stator current, and output the coordinated scheduling correction control data.

8. A yard intelligent crane dispatching system based on parallel collaboration, characterized in that, The system is used to implement the intelligent yard traffic scheduling method based on parallel cooperation as described in any one of claims 1-7, and the system includes: The spreader interference analysis module compares the vertical height coordinate parameters of the corresponding terms in the container lifting three-dimensional coordinate set to generate vertically stacked directed edges, calculates the difference in the expected lifting action time parameters of the near-distance terms to generate horizontal interference directed edges, and integrates and outputs scheduling interference topology map data. The concurrent task parsing module counts the number of nodes pointed to in the scheduling interference topology graph data to generate the node in-degree value, takes the zero value item to construct the concurrent execution task queue and distributes it, extracts the first spatial dwell coordinate parameter, the second spatial dwell coordinate parameter, the first vector running speed parameter and the second vector running speed parameter, and generates parallel vehicle running status data. The envelope intersection deduction module, based on the parallel vehicle operation status data, maps the displacement of the first vector running speed parameter and the first spatial dwell coordinate parameter, and the second vector running speed parameter and the second spatial dwell coordinate parameter to the three-dimensional vertex matrix of the first displacement envelope surface and the three-dimensional vertex matrix of the second displacement envelope surface, and calculates the three-dimensional intersection volume value to derive the intersection volume change parameter; The intrusion tolerance mapping module calculates the difference between the intersection volume change parameter and the preset physical intrusion tolerance value and maps it into a deceleration braking level signal. It compares the specific values ​​of the first vector running speed parameter and the second vector running speed parameter, injects the deceleration braking level signal into the motor controller, and outputs the anti-collision avoidance drive command. The sway angle coupling compensation module responds to the vehicle deceleration and avoidance action triggered by the anti-collision avoidance drive command, compares the pixel changes of continuous deflection image frames to extract real-time sway angle values ​​and differentially derives sway angle acceleration values, calculates the coupling deviation with the actual moving acceleration parameters of the vehicle to generate nonlinear coupling disturbance term values, derives equivalent compensation current values ​​and superimposes them with the drive stator current, and outputs coordinated scheduling correction control data.

Images