A port-oriented large-scale equipment intelligent networking technology

Abstract

The application discloses a kind of port-oriented large-scale equipment intelligent networking technology, specifically relates to wireless communication network technical field, including obtaining port equipment real-time running state and production scheduling data, constructs the three-dimensional digital twin model containing future dynamic barrier space-time information;Based on the model, the change trend of Fresnel zone residual gap is predicted, and a dynamic link quality weight sequence is generated;According to the weight sequence, weighted clustering clustering is used, and the main cluster head and shadow cluster head are elected in each cluster.The shadow cluster head synchronizes the main cluster head routing forwarding table in real time, but keeps data forwarding silent;When the link degradation is predicted to the switching condition, the main cluster head sends the activation instruction, the shadow cluster head enables the standby link, and the source node concurrently transmits data packets to the main and standby links during the transition period until the switching is completed after the main link fails.The application realizes the active prediction and seamless smooth switching of link quality, significantly reduces the data interruption delay, and is suitable for reliable networking in high-dynamic shielding environment of port.

Description

Technical Field

[0001] This invention relates to the field of wireless communication network technology, and more specifically, to a large-scale intelligent networking technology for port equipment. Background Technology

[0002] With the advancement of port automation and intelligent construction, various operating equipment such as quay cranes, rail-mounted gantry cranes, container trucks, and automated guided vehicles (AGVs) are gradually achieving unmanned and networked operation. The scale of wireless communication nodes in port operation areas continues to expand, placing stringent requirements on the reliability, handover continuity, and latency performance of large-scale equipment networks. Port scenarios are characterized by diverse equipment types, frequent mobile operations, and dense metal structures and container stacking. During operations, dynamic processes such as spreader lifting, boom rotation, container truck movement, and yard crane movement frequently obstruct the propagation path of wireless signals, leading to drastic fluctuations in link quality or even instantaneous interruptions. Traditional wireless networking solutions are insufficient to meet the high-reliability continuous communication requirements of port production control operations.

[0003] Existing large-scale port equipment networking technologies generally suffer from the following technical deficiencies when dealing with dynamic obstacle obstruction:

[0004] Excessive route reconstruction latency: Relying on reactive route repair after link interruption, in highly dynamic port obstruction scenarios, the link status may change from normal to interrupted within milliseconds. The path reconstruction process usually takes hundreds of milliseconds to several seconds, far exceeding the latency tolerance limit of port real-time control services.

[0005] Flooding leads to a waste of channel resources: Frequent routing message broadcasts can seriously occupy wireless channel resources in large-scale equipment scenarios, especially in environments with severe frequency band interference in ports. Control message overhead may exceed the effective data transmission bandwidth, exacerbating network congestion.

[0006] Cluster head election lacks link quality prediction basis: existing clustering schemes mostly rely on static node attributes or instantaneous metrics to elect cluster heads, without considering the changing trend of wireless link quality between nodes in the future time period; the elected cluster head may lose connection with cluster members within seconds of being elected due to dynamic occlusion, resulting in frequent cluster structure reconstruction and network topology oscillation;

[0007] The lack of a smooth transition mechanism during the switchover: When the routing path fails, the source node completely loses its communication capability before finding a new path. There is no packet caching or concurrent transmission mechanism during the switchover process, which leads to the loss of control commands and affects the continuity and safety of port operations.

[0008] In summary, existing technologies fail to effectively address the rapid degradation of link quality caused by dynamic obstacles in large-scale dynamic equipment networking scenarios in ports. They lack the ability to proactively predict link changes, and the switching process has a significant data interruption window, making it difficult to meet the high reliability and low latency requirements of port automation operations for wireless communication networks. Summary of the Invention

[0009] In order to overcome the above-mentioned defects of the prior art, the present invention provides a large-scale intelligent networking technology for port equipment, which solves the problems mentioned in the background art through the following scheme.

[0010] To achieve the above objectives, the present invention provides the following technical solution: a large-scale intelligent networking technology for port equipment, comprising:

[0011] S1: Obtain real-time operating status data of equipment and production scheduling plan data within the port operation area, and construct a three-dimensional digital twin model containing the spatiotemporal location information of dynamic obstacles in future continuous time slots;

[0012] S2: Based on the three-dimensional digital twin model, predict the Fresnel gap change trend of the wireless transmission path between devices to be networked in the future continuous time slots, and generate a link quality weight sequence that changes dynamically with time.

[0013] S3: Based on the link quality weight sequence, a weighted clustering strategy is used to dynamically cluster the device, and a master cluster head node and a shadow cluster head node are elected in each cluster; the shadow cluster head node synchronizes the routing table of the master cluster head node in real time, and always maintains the silent state of data forwarding function;

[0014] S4: When the link communication quality is predicted based on the link quality weight sequence, and the communication quality deteriorates to the preset switching condition, the main cluster head node sends an activation command to the shadow cluster head node to enable the backup transmission link corresponding to the shadow cluster head node. The source node is controlled to concurrently transmit data packets to the current main link and the backup transmission link during the switching transition until the main link can no longer maintain data carrying.

[0015] The technical effects and advantages of this invention are as follows:

[0016] 1. This invention constructs a three-dimensional digital twin model containing the spatiotemporal location information of dynamic obstacles through S1 to S2, and calculates the Fresnel gap value on a time-slot basis based on the model, thereby generating a link quality weight sequence that dynamically changes over time in future consecutive time slots; this technical means enables the network to perceive the occlusion trend in advance before the link actually deteriorates, providing a predictive basis for subsequent handover decisions, and overcoming the shortcomings of traditional reactive networking based on real-time measurement feedback in response lag in port highly dynamic scenarios;

[0017] 2. In the weighted agglomerative hierarchical clustering process of S3, this invention uses a weighted combination of spatial Euclidean distance between nodes and the similarity of link quality weight sequences as the dissimilarity measure, rather than relying solely on instantaneous signal strength or a single geographical distance for clustering. This technique enables the clustering results to take into account both physical spatial proximity and the similarity of time-varying link quality characteristics, effectively avoiding frequent cluster reconstruction caused by temporary occlusion or instantaneous fading, and improving the robustness of the cluster structure and the continuous stability of communication links within the cluster.

[0018] 3. In this invention, a shadow cluster head node is elected in S3. This node synchronizes the routing and forwarding table of the main cluster head node in real time through the synchronization control channel, but its data link layer forwarding engine always remains in a logically silent state and does not forward any non-control data packets. This technique enables the shadow cluster head to maintain a hot standby state of the control plane only during the standby period, while the data plane does not participate in service carrying. This ensures the consistency of routing information during handover and avoids the power consumption and channel resource waste caused by simultaneous forwarding by two cluster heads. It is especially suitable for port mobile equipment scenarios.

[0019] 4. In S4 of this invention, when it is predicted that the main link is about to deteriorate, the main cluster head node first activates the backup transmission link of the shadow cluster head node, and controls the source node to concurrently transmit data packets to the current main link and the backup transmission link during the handover transition. At the same time, the receiving end achieves deduplication and out-of-order reordering through a unified associated sequence number. This technical means ensures that before the main link is officially cut off, the backup link has already carried the same data stream, and the receiving end can remove duplicate packets according to the sequence number, ensuring that there is no data loss during the handover transition and that the upper layer services are not affected by out-of-order delivery. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the logical structure of the method of the present invention.

[0021] Figure 2 This is a schematic diagram of the three-dimensional digital twin model structure of the present invention S1.

[0022] Figure 3 This is a schematic diagram of the link quality weight sequence generation structure of S2 in this invention.

[0023] Figure 4 This is a schematic diagram of the weighted clustering structure of S3 in this invention. Detailed Implementation

[0024] 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, and 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.

[0025] Please see Figures 1-4 As shown, this embodiment of the invention provides a large-scale intelligent networking technology for port equipment, including:

[0026] S1: Obtain real-time operating status data of equipment and production scheduling plan data within the port operation area, and construct a three-dimensional digital twin model containing the spatiotemporal location information of dynamic obstacles in future continuous time slots;

[0027] In this embodiment, S1 is used to acquire real-time operating status data of equipment within the port operation area and production scheduling plan data, and to construct a three-dimensional digital twin model containing spatiotemporal location information of dynamic obstacles within future continuous time slots, specifically including:

[0028] S101: Multi-source heterogeneous data acquisition and timestamp synchronization

[0029] The Internet of Things (IoT) sensing layer deployed in the port operation area is used to acquire real-time operational status data. The sensing layer includes, but is not limited to, high-precision GNSS positioning terminals, inertial measurement units (IMUs), ultra-wideband (UWB) positioning base stations installed on the equipment, and operational status parameters output by the equipment controller, such as spreader lifting height, boom rotation angle, and trolley and crane operating speed.

[0030] By connecting to the port equipment scheduling management through a data interface, production scheduling plan data within a future time window is retrieved; the scheduling plan data includes: the work task sequence of each piece of equipment, target location coordinates, scheduled arrival time window, and travel path nodes.

[0031] The data preprocessing unit performs timestamp calibration on the aforementioned multi-source data, unifying the temporal resolution of all data to a preset base time slot step size. In this embodiment, the preferred method is... It can be adjusted according to the dynamic frequency of port operations to ensure timing consistency in subsequent fusion calculations.

[0032] S102: Dynamic Obstacle Target Extraction and Real-time Motion State Estimation

[0033] The collected real-time operational status data is input into the target tracking filter to perform identification and motion modeling on all mobile devices within the port operation area, specifically including:

[0034] For wheeled equipment such as container trucks, a uniform acceleration motion model combined with extended Kalman filtering is used to smooth and denoise their position, velocity, and acceleration.

[0035] For equipment with multi-joint rotation characteristics, such as quay cranes and yard cranes, a multi-rigid-body kinematic model is adopted to calculate the instantaneous spatial coordinates of the spreader and the end of the boom in real time based on the boom rotation angle and pitch angular velocity.

[0036] Generate the current time The motion state data of all dynamic obstacles is provided below, including: real-time pose snapshots and velocity vectors. Yaw angle .

[0037] S103: Spatiotemporal Trajectory Prediction Based on Production Scheduling Plan

[0038] The discrete path nodes in the production scheduling plan data are dynamically weighted and fused with the estimation results of real-time motion state data to predict the current time. The following series The time slot for each device, i.e., the future time window. The spatiotemporal position sequence of each dynamic obstacle within the area, and the specific prediction logic includes:

[0039] Path matching and intent recognition: Map the current real-time location to the planned path in the production scheduling plan to determine which operation phase the equipment is in: empty transfer, heavy transport, or fixed-point loading and unloading;

[0040] If the equipment is in an unloaded transfer or heavy-load transport phase, the planned speed in the scheduling plan is used as the baseline target value, and the real-time inertial measurement value is used as the correction value. A PID control algorithm is employed to calculate the equipment's position in the global coordinate system within each time slot. The predicted three-dimensional coordinates are shown below.

[0041] If the equipment is in a fixed-point loading / unloading phase, the lifting or lowering time in the scheduling plan, combined with the hoisting speed of the spreader, is used to predict the vertical movement of the spreader. The trajectory of height change on the axis;

[0042] Reasonableness verification: Kinematic constraint verification is performed on the predicted spatiotemporal location sequence to remove abrupt changes caused by data noise, and cubic spline interpolation is used to smooth the trajectory.

[0043] Based on the above calculations, a set of spatiotemporal locations of all dynamic obstacles is generated within future consecutive time slots. ,in , This represents the device attitude angle, used for subsequent clearance calculations.

[0044] S104: Construct a 3D digital twin model and inject dynamic obstacle spatiotemporal tags.

[0045] Based on a static high-precision map of the port operation area, including buildings, fixed quay crane structures, container yard boundaries, and lighthouse pillars, a static 3D scene model is built in a digital twin engine; the static scene model uses an octree data structure for spatial partitioning.

[0046] Subsequently, the predicted spatiotemporal location set in S103 will be... Dynamic injection into a static model specifically includes:

[0047] For each dynamic obstacle, a dynamic blocking body is generated at the predicted coordinates of the corresponding time slot based on the three-dimensional parameters pre-stored according to its device type.

[0048] Add a timestamp index to each dynamic blocking body to construct a three-dimensional voxel change sequence oriented towards the time slot axis, i.e., a four-dimensional spatiotemporal dataset;

[0049] Between time slots, a linear interpolation algorithm is used to complete the continuous motion trajectory of the dynamic obstruction, forming a smooth time-dimensional animation sequence that covers the present moment to the future. A three-dimensional digital twin model containing complete spatiotemporal location information of dynamic obstacles within a time period is stored in a real-time database for S2 to access;

[0050] As a preferred embodiment, the future time window The length is dynamically set based on the maximum tolerable delay of the port's wireless link handover, typically ranging from 5 to 30 seconds. This ensures that the entire process of path obstruction is covered without affecting real-time performance due to excessive computation.

[0051] S2: Based on the three-dimensional digital twin model, predict the Fresnel gap change trend of the wireless transmission path between devices to be networked in the future continuous time slots, and generate a link quality weight sequence that changes dynamically with time.

[0052] In this embodiment, based on the three-dimensional digital twin model described in S1, the Fresnel zone clearance variation trend of the wireless transmission path between devices to be networked is predicted in future consecutive time slots, and a link quality weight sequence that dynamically changes over time is generated, specifically including:

[0053] S201: Determine the pairs of device nodes to be networked and their three-dimensional spatial motion trajectories.

[0054] Retrieves all wireless communication nodes currently in a pending network state from the equipment management list of the port operation area, including but not limited to vehicle-mounted access points on quay cranes, terminal nodes on yard cranes, mobile radios on container trucks, and fixed relay nodes deployed on lighthouses or building rooftops; for any two devices forming a node pair... Perform the following operations:

[0055] Based on the three-dimensional digital twin model constructed in S104, in future consecutive time slots Extract nodes and nodes The three-dimensional predicted coordinates of the phase center in the global coordinate system are stored in advance if the device antenna installation position has a fixed spatial offset relative to the device's centroid, and rigid body translation compensation is performed when extracting the coordinates.

[0056] For the node pairs formed by the devices to be networked Generate it in time slot The lower transmitter position vector and receiver position vector And calculate the straight-line distance between the two. .

[0057] S202: Calculate the Fresnel zone radius of the wireless transmission path in each time slot.

[0058] Determine the carrier wavelength based on the pre-set operating frequency band for port wireless communication. For node pairs In the time slot Transmission distance Calculate the Fresnel zone radius at any cross-section; the most crucial value is the Fresnel zone radius at the midpoint of the path. :

[0059]

[0060] Preferably, if it is necessary to consider the occlusion situation of different cross-sections, the Fresnel zone radius at the cross-section where the obstacle is located can also be calculated based on the location of the obstacle. In this embodiment, to simplify calculations and ensure engineering feasibility, the Fresnel zone radius at the midpoint of the path is used as the criterion.

[0061] S203: Calculate the real-time clearance value of dynamic obstacles on the transmission path.

[0062] The spatiotemporal location information of the dynamic obstacles generated in S104 is used as a spatial obstruction set, and the node pairs are targeted accordingly. For each time slot, the spatial linear propagation path is calculated sequentially, including spatial collision and gap calculations.

[0063] Spatial linear parameterization: time slot The propagation path under these conditions can be represented by parametric equations. ,in ;

[0064] Screening for potential occlusions: Traverse the 3D coordinates of all dynamic obstacles in the 3D digital twin model, and use the bounding box intersection test to filter out obstacles that intersect with line segments. For spatially unrelated obstacles that do not intersect, retain the set of candidate obstacles that may cause occlusion. ;

[0065] Calculate the minimum spatial clearance: for the candidate obstacle set Every obstacle in Calculate its three-dimensional coordinates to the line segment Shortest spatial distance The point-line distance projection method is used to calculate the perpendicular distance from the three-dimensional coordinates of the obstacle to the line segment, and the minimum value is taken as the characterization of the obstacle's penetration depth into the transmission path; if the three-dimensional coordinates of the obstacle intersect the line segment in space, it is determined that the obstacle completely blocks the Fresnel zone.

[0066] Calculate Fresnel clearance Select the shortest spatial distance that has the greatest impact on transmission from all candidate obstacle spatial distances. Compare it with the Fresnel zone radius in S202 Comparison, defining Fresnel clearance for:

[0067]

[0068] when When the obstacle is completely outside the Fresnel zone, the path approximates free space propagation; when When this occurs, it indicates that the obstacle partially intrudes into the Fresnel zone, resulting in diffraction attenuation; when... This indicates that an obstacle has touched or crossed the line-of-sight connection between the transmitter and receiver, posing a risk of link interruption.

[0069] S204: Trend Prediction and Smoothing Filtering Based on Multi-Slot Residual Slot Sequences

[0070] Repeat steps S201 to S203 to calculate the future consecutive... Node pairs in each time slot The Fresnel gap values ​​are used to form the original gap time series. ;

[0071] To prevent drastic changes in the clearance value due to minute positioning jitter in the digital twin model, which could lead to misjudgments in subsequent handovers, the original clearance time series is subjected to sliding window mid-range filtering. The clearance value of each time slot is replaced by the median of its immediate and next-neighboring values, resulting in a smoothed clearance sequence. ;

[0072] Meanwhile, in order to capture the changing trend of the gap, the least squares method is used to perform first-order linear fitting on the end time slot of the smoothed sequence to calculate the slope of the gap change. If the slope of the gap change is less than 0, it indicates that the link quality is continuously deteriorating. The absolute value of the slope can be used as an auxiliary parameter for predicting and switching communication quality degradation in S4.

[0073] S205: Mapping gap values ​​to a link quality weight sequence

[0074] Establish a nonlinear mapping relationship between Fresnel zone clearance and link quality weight, and define node pairs. In the time slot Link quality weights The larger the value, the better the link quality. The specific mapping function adopts a piecewise continuous model:

[0075]

[0076] in and These are preset high-quality and low-quality thresholds, respectively. Based on historical values ​​from port wireless communication equipment, this embodiment preferably uses... , When the spare slot ratio is greater than 0.6, the link is considered a high-quality link; when the spare slot ratio drops below 0, the link is considered unusable.

[0077] The above mapping is applied time-slot by time slot to the smoothed residual gap sequence. Generate node pairs Link quality weight sequence that changes dynamically over time ;

[0078] Finally, S201 to S205 are repeatedly executed for all nodes of equipment to be networked in the port operation area to obtain a set of multidimensional weight sequences characterizing the time-varying characteristics of the link quality of the entire network topology. The data is stored in the database and used as input data for dynamic clustering and cluster head election in S3.

[0079] S3: Based on the link quality weight sequence, a weighted clustering strategy is used to dynamically cluster the device, and a master cluster head node and a shadow cluster head node are elected in each cluster; the shadow cluster head node synchronizes the routing table of the master cluster head node in real time, and always maintains the silent state of data forwarding function;

[0080] In this embodiment, S3 dynamically clusters the devices according to the link quality weight sequence using a weighted clustering strategy, and elects a master cluster head node and a shadow cluster head node within each cluster; the shadow cluster head node synchronizes the routing table of the master cluster head node in real time and always maintains a silent state for data forwarding function, specifically including:

[0081] S301: Construct a weighted network topology and set the clustering trigger period.

[0082] Read the multidimensional weight sequence set generated in S205 The set of all network-connected devices within the port operation area is used as the vertex set. The link quality weight between any two nodes in the current time slot. Construct an undirected weighted network topology graph with edge weights. ;

[0083] Set the execution cycle for dynamic clustering. Considering the relatively slow movement speed of port equipment and the regularity of scheduling plans, this embodiment preferably... In addition, to cope with sudden topology changes caused by sudden occlusion, an event triggering mechanism is added: when the slope of any link clearance change calculated in S204 exceeds the preset drastic change threshold, non-periodic emergency re-clustering is triggered to ensure the timeliness of the cluster structure.

[0084] S302: Dynamic Clustering Based on Agglomerative Hierarchical Clustering with Weighted Distance

[0085] Weighted agglomerative hierarchical clustering algorithm is used to cluster the node set. The division and clustering are based on the measure of dissimilarity between nodes. Taking into account both physical spatial proximity and the similarity of link quality weight sequences, it is specifically defined as:

[0086]

[0087] in Represents the spatial Euclidean distance between two nodes at the current time. Indicates the maximum coverage radius of port wireless communication. Indicates the total number of consecutive time slots in the future. This represents a weighting adjustment factor that balances physical distance and link quality.

[0088] The weighted agglomerative hierarchical clustering algorithm needs further explanation. Its specific process includes:

[0089] Initialization: Treat each device node as an independent cluster;

[0090] Iterative merging: Calculate the shortest weighted distance between all clusters, in order to... The average value is used as the inter-cluster distance, and the two clusters with the smallest distance are merged into a new cluster;

[0091] Termination condition: Merging is terminated when the distance between all clusters is greater than the preset cluster radius threshold, or when the number of clusters reaches the preset lower limit.

[0092] Through the above clustering process, the dynamic equipment in the port operation area is divided into relatively independent node clusters with stable internal link quality. ,in This indicates the final number of clusters generated.

[0093] S303: Elect a master cluster head node within the cluster.

[0094] For each cluster generated in S302 The process of electing the master cluster head node is executed, and the election is based on a comprehensive weighted score. The score is calculated by weighted summation of the following three indicators:

[0095]

[0096] in Represents a node The time average of the link quality weights between the node and all other nodes in the cluster over consecutive future time slots is used to characterize the communication stability of the node.

[0097] Represents a node Normalized topological degree centrality within the cluster, i.e., the relationship between the current time and the node The proportion of neighboring nodes with valid links to the total number of nodes in the cluster is used to characterize the connectivity breadth of a node.

[0098] Represents a node The remaining power is set to 1 for stationary power supply equipment and the current remaining power percentage for battery-powered mobile truck equipment, used to characterize the node's endurance reliability. This represents the corresponding weighting coefficient;

[0099] Calculate the comprehensive score of all nodes in the cluster, and select the one with the highest score as the main cluster head node of the cluster. The main cluster head node is responsible for aggregating the perception data of the member nodes in the cluster and conducting routing relay communication with the main cluster head of the adjacent cluster or the shore-based core gateway.

[0100] S304: Election of Shadow Cluster Head Nodes

[0101] In the same cluster Within the cluster, a shadow cluster head node is elected from the remaining nodes excluding the primary cluster head node. The election criteria for the shadow cluster head are similar to those for the primary cluster head, but an additional spatial difference between the shadow and primary cluster head nodes is introduced. As a punitive factor:

[0102]

[0103] in The spatial diversity encouragement factor prioritizes nodes that are geographically far from the main cluster head but have excellent communication capabilities as shadow cluster heads. This ensures that when the main cluster head fails due to strong local occlusion, the shadow cluster head can provide a reliable backup path in a spatial diversity sense that is physically different from the main cluster head. The node with the highest score after correction is selected as the shadow cluster head node. If there is only a single node in the cluster, that node serves as both the main cluster head and the shadow cluster head.

[0104] S305: Shadow cluster head nodes synchronize the routing and forwarding table of the primary cluster head node in real time.

[0105] Once the primary cluster head node is elected, it immediately establishes and dynamically maintains the routing table for this cluster. The routing table contains at least the following fields: destination node address, next-hop node address, outgoing interface identifier, hop count, and the real-time quality weight value of the corresponding link.

[0106] A dedicated in-band synchronization control channel is established between the shadow cluster head node and the primary cluster head node. Using the reliable transmission mechanism of the TCP / IP protocol, the real-time synchronization process of the routing table is performed, specifically including:

[0107] Full synchronization: At the initial moment after the shadow cluster head is selected, the master cluster head sends the complete routing table to the shadow cluster head through the synchronization control channel. After receiving it, the shadow cluster head loads it into the local backup routing table area.

[0108] Incremental synchronization: During subsequent operation, whenever the routing table of the primary cluster head node is updated due to topology changes, the primary cluster head node immediately encapsulates the updated entry into a synchronization message and sends it to the shadow cluster head node with an update cycle of no more than 50 milliseconds. This is to ensure that the routing information backed up by the shadow cluster head node remains consistent with that of the primary cluster head node.

[0109] S306: Shadow cluster head nodes maintain a silent state for data forwarding functionality.

[0110] While the shadow cluster head node performs normal routing table synchronization, its data link layer forwarding engine is configured to logical silent mode, specifically including:

[0111] The shadow cluster head node sets a forwarding enable flag at the media access control layer or network layer. Until an activation command is received, this flag will always be set to prevent forwarding.

[0112] When the upper-layer protocol stack of the shadow cluster head node receives a non-control class data packet from a member node or a neighboring node within the cluster, its forwarding plane processing procedure performs the following determination:

[0113] like If the data packet is directly discarded without any relay or routing forwarding operation, only a local reception log is recorded for auditing purposes; However, if the destination address of the data packet is the shadow cluster header itself, it will be normally delivered to the upper-layer protocol stack for processing;

[0114] In order to maintain basic time synchronization and neighbor awareness with the main cluster head and members within the cluster while in a silent state, the shadow cluster head node is only allowed to send and receive necessary link maintenance beacon frames to maintain synchronization of the physical layer and MAC layer, but does not respond to any data routing requests or participate in carrying inter-cluster data traffic.

[0115] Through the aforementioned silent mechanism, the shadow cluster head node only serves as a real-time update of the hot standby control plane during normal operation. All cluster master cluster head node and shadow cluster head node information, cluster member lists, and current routing tables of each cluster are aggregated in the port edge computing server for S4 to use for switching decisions.

[0116] S4: When the link communication quality is predicted based on the link quality weight sequence, and the communication quality deteriorates to the preset switching condition, the main cluster head node sends an activation command to the shadow cluster head node to enable the backup transmission link corresponding to the shadow cluster head node. The source node is controlled to concurrently transmit data packets to the current main link and the backup transmission link during the switching transition until the main link can no longer maintain data carrying.

[0117] In this embodiment, S4 predicts the link communication quality based on the link quality weight sequence. When the communication quality deteriorates to a preset switching condition, the primary cluster head node sends an activation command to the shadow cluster head node to activate the backup transmission link corresponding to the shadow cluster head node. This controls the source node to concurrently transmit data packets to both the current primary link and the backup transmission link during the switching transition, until the primary link can no longer maintain data carrying capacity. Specifically, this includes:

[0118] S401: Prediction and handover triggering based on link quality weight sequence

[0119] During each dynamic clustering cycle, the master cluster head node reads the link quality weight sequence generated in S205, which specifies the link quality weights between itself and each member node within the cluster for future consecutive time slots. And the slope of the clearance change calculated in S204; the main cluster head node performs prediction and executes the trigger judgment logic:

[0120] Define the preset switching trigger condition as the condition that either of the following two conditions is met:

[0121] Condition for reaching the absolute value of the weight: the existence of a future time slot. This makes the predicted link quality weights ,in This represents a preset handover trigger threshold value. In this embodiment, it is preferred based on the upper limit of the port's wireless communication tolerance for packet loss rate. ;

[0122] Accumulated deterioration trend conditions: Over consecutive future time slots, the moving average slope of the link quality weight sequence remains negative, and the single-step decrease in the weight of the current time slot compared to the weight of the previous time slot is significant. This indicates that the link quality is undergoing a rapid and irreversible degradation.

[0123] When either of the above two conditions is met, the main cluster head node determines that the current main link will be unable to effectively carry service data in a short period of time and immediately activates the S402 handover preparation process; at the same time, in order to prevent frequent false triggers due to small fluctuations in the weight sequence, a de-jitter timer is introduced into the judgment logic, and the handover trigger signal is output only after the triggering condition is continuously met for more than the timer duration.

[0124] S402: The primary cluster head node sends an activation command to the shadow cluster head node and completes the signaling handshake.

[0125] After the switching trigger signal is output, the primary cluster head node encapsulates and generates a switching activation instruction frame. The instruction frame contains at least the following fields: current primary link identifier, target data stream identifier to be switched, network address of the shadow cluster head node, timestamp of the activation time, and digital signature verification code for authentication.

[0126] The primary cluster head node sends the instruction to the shadow cluster head node via the dedicated synchronization control channel established in S305 using a three-way handshake confirmation mechanism:

[0127] First handshake: The primary cluster head sends a command frame to the shadow cluster head and starts a retransmission timer;

[0128] Second handshake: After the shadow cluster head node successfully receives and verifies the instruction, it immediately replies to the main cluster head with an activation preparation confirmation frame and caches the activation parameters in the verification instruction locally;

[0129] The third handshake: After receiving the preparation confirmation frame, the primary cluster head sends an activation execution confirmation frame to the shadow cluster head, completing the entire signaling handshake process;

[0130] If the primary cluster head does not receive a ready confirmation frame after the retransmission timer expires, it will automatically retransmit the command, retrying a maximum of 3 times. If all 3 retries fail, the primary cluster head determines that the shadow cluster head node is disconnected and triggers an emergency broadcast reselection process within the cluster to select the suboptimal node as the new shadow cluster head and re-execute.

[0131] S403: The shadow cluster head node enables the backup transmission link and switches to active forwarding state.

[0132] After successfully completing the third handshake of S402, the shadow cluster head node automatically performs a state transition operation when the timestamp of the activation time arrives:

[0133] Modify the forwarding enable flag: Change the forwarding enable flag maintained in S306. Switch from logical value 0 to logical value 1;

[0134] Activate backup wireless transceiver channel: If the shadow cluster head node is equipped with a multi-radio multi-channel wireless interface, immediately switch the preset backup frequency point or backup antenna link from low power standby mode to full power working mode, and establish a backup routing forwarding fast table for the target data flow in the local data plane according to the routing forwarding table synchronized in real time in S305, which includes the corresponding next hop out interface and link layer address mapping.

[0135] Broadcast a backup link ready notification to member nodes within the cluster: The shadow cluster head node sends a link ready announcement to all member nodes within the cluster via a single-hop broadcast, informing each member node that the backup transmission link has been activated, and carrying the MAC address of the shadow cluster head and the backup link channel parameters.

[0136] After completing the above operations, the shadow cluster head node officially transitions from hot standby silent state to active hot standby state, and its data plane forwarding channel is fully open. However, at this time, it has not yet actually carried service traffic and is only in a standby receiving state at the physical and link layers.

[0137] S404: The source node concurrently transmits data packets to both the primary and backup transport links during the handover transition.

[0138] After the shadow cluster head sends the announcement, the handover transition period officially begins. During this transition period, the primary cluster head node, acting as the control center, sends concurrent dual-transmission commands to all source nodes within the cluster that need to report or relay data. Upon receiving the commands, the source nodes execute the following concurrent transmission logic:

[0139] Packet duplication and sequence number marking: When the source node sends protocol data units, it assigns a unified primary and backup associated sequence number to each data packet to be sent in the extended field of the network layer or application layer header of the protocol stack, to ensure that the same data packet has the same payload content and sorting index on the primary and backup paths; a path identifier field is added to the header of the data packet, marked as 0 and 1 respectively, for the receiving end to perform deduplication processing;

[0140] Dual-path parallel transmission: The source node submits the copied data packets to both the primary link transmission queue and the backup link transmission queue simultaneously. The two links transmit in parallel on the physical channel, sharing the same time slot resources and frequency domain resources.

[0141] Deduplication and out-of-order reordering at the receiving end: After receiving data packets from the primary and backup links, the target gateway or core receiving node performs deduplication buffering based on the source address in the packet header; whenever a new data packet is received, the receiving end checks if a duplicate record exists.

[0142] If it does not exist, the data packet is delivered to the upper-layer protocol stack and the record is added to the cache; if it already exists, the newly arrived duplicate packet is directly discarded and the cache is updated according to the arrival time; the unsorted data packets in the cache are sorted in ascending order to offset the out-of-order phenomenon that may be caused by the difference in propagation delay between the two links.

[0143] Throughout the concurrent transmission phase, the primary link and the backup link simultaneously carry the exact same data traffic. Since the two links have diversity characteristics in the physical space path, any single local blockage will only affect the reception quality of one of the links, ensuring that the data packet delivery success rate is close to 1 during the transition period.

[0144] S405: Dynamic monitoring of main link failure, main link release and switching

[0145] During the concurrent dual-transition period, the main cluster head node continuously monitors the carrying capacity of the current main link in real time. When the main link cannot maintain data carrying capacity, a dual-criteria joint decision mechanism is adopted:

[0146] Physical layer criterion: Fresnel gap value predicted in the next adjacent time slot in S204 This means that the obstacle has been completely wedged into the line-of-sight propagation path of the transceiver, and the wireless diffraction loss tends to be infinite.

[0147] Link layer criterion: Continuous connection on the main link If a data packet is sent but no MAC layer acknowledgment frame is received from the receiving end, the actual measured equivalent packet loss rate exceeds the preset tolerance limit.

[0148] When any of the above criteria is met, the master cluster head node determines that the main link has completely lost its effective data carrying capacity, and the master cluster head node performs the following operations:

[0149] Send a main link release announcement: Broadcast a main link release message to the shadow cluster head node and all relevant source nodes, informing them to stop sending any new data to the main link from the next time slot;

[0150] Source node terminates concurrent dual transmission: After receiving continuous timeouts on the primary link, the source node immediately stops the data packet replication operation, queues all data packets to be sent to the backup link transmission queue, and switches to single-path transmission mode.

[0151] The shadow cluster head is officially promoted to the new primary cluster head: After receiving the information, the shadow cluster head node automatically raises the priority of its local routing table to the primary cluster head level, and broadcasts a cluster head handover completion announcement to the member nodes in the cluster, declaring itself to become the new primary cluster head node. The original primary cluster head node is downgraded to an ordinary cluster member node or enters a state of pending recovery detection.

[0152] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other.

[0153] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A large-scale intelligent networking technology for port equipment, characterized in that, include: S1: Obtain real-time operating status data of equipment and production scheduling plan data within the port operation area, and construct a three-dimensional digital twin model containing the spatiotemporal location information of dynamic obstacles in future continuous time slots; S2: Based on the three-dimensional digital twin model, predict the Fresnel gap change trend of the wireless transmission path between devices to be networked in the future continuous time slots, and generate a link quality weight sequence that changes dynamically with time. S3: Based on the link quality weight sequence, a weighted clustering strategy is used to dynamically cluster the device, and a master cluster head node and a shadow cluster head node are elected in each cluster; the shadow cluster head node synchronizes the routing table of the master cluster head node in real time, and always maintains the silent state of data forwarding function; S4: When the link communication quality is predicted based on the link quality weight sequence, and the communication quality deteriorates to the preset switching condition, the main cluster head node sends an activation command to the shadow cluster head node to enable the backup transmission link corresponding to the shadow cluster head node. The source node is controlled to concurrently transmit data packets to the current main link and the backup transmission link during the switching transition until the main link can no longer maintain data carrying.

2. The intelligent networking technology for large-scale equipment in ports according to claim 1, characterized in that, S1 specifically includes: S101: Obtain real-time operating status data of equipment through the IoT sensing layer deployed in the port operation area, pull production scheduling plan data from port equipment through the data interface, and perform timestamp calibration on the acquired multi-source data. S102: Input the real-time operating status data into the target tracking filter to identify and model the motion of all mobile devices in the port operation area, and generate motion status data of all dynamic obstacles at the current moment. S103: Dynamically weight and fuse discrete path nodes in production scheduling plan data with motion status data to predict the spatiotemporal position sequence of each dynamic obstacle in multiple consecutive time slots after the current moment; S104: Based on the static high-precision map of the port operation area, build a static three-dimensional scene model, and dynamically inject the predicted spatiotemporal location sequence into the static three-dimensional scene model to generate a three-dimensional digital twin model covering the current time to the future time window.

3. The intelligent networking technology for large-scale equipment in ports according to claim 1, characterized in that, S2 specifically includes: S201: Based on the three-dimensional digital twin model, extract the transmitter position vector and receiver position vector of any two nodes formed by the devices to be networked in the future continuous time slots, and calculate the spatial straight-line distance between them. S202: Determine the carrier wavelength according to the pre-set operating frequency band of port wireless communication, and calculate the Fresnel zone radius at the midpoint of the path for the transmission distance of the node pair in each time slot; S203: Using the spatiotemporal location information of the dynamic obstacle as a spatial obstruction set, for the spatial straight-line propagation path of the node pair in each time slot, calculate the shortest spatial distance from each dynamic obstacle to the spatial straight-line propagation path in each time slot, and calculate the Fresnel clearance value based on the shortest spatial distance and the Fresnel zone radius; S204: Repeat S201 to S203 to calculate the Fresnel residual slot values ​​of the node pairs in multiple consecutive time slots in the future, form the original residual slot time series, and perform smoothing filtering on the original residual slot time series. S205: Establish a nonlinear mapping relationship between Fresnel zone clearance and link quality weight, and map the smoothed and filtered clearance sequence to the link quality weight sequence that dynamically changes with time for the node pair.

4. The intelligent networking technology for large-scale equipment in ports according to claim 3, characterized in that, The calculation of the Fresnel clearance value in S203 specifically includes: The spatial straight-line propagation path is expressed as a parametric equation. The three-dimensional coordinates of all dynamic obstacles in the three-dimensional digital twin model are traversed. The spatial bounding box intersection test is used to screen the set of candidate obstacles that may cause occlusion. For each obstacle in the candidate obstacle set, calculate the shortest spatial distance from its three-dimensional coordinates to the straight-line propagation path in space, and select the minimum value of the shortest spatial distance among all candidate obstacles; The minimum value is compared with the Fresnel zone radius, and the Fresnel zone clearance value is defined as the difference between the minimum value and the Fresnel zone radius.

5. The intelligent networking technology for large-scale equipment in ports according to claim 3, characterized in that, The nonlinear mapping relationship in S205 is as follows: A piecewise continuous model is adopted. When the ratio of the Fresnel clearance value to the Fresnel radius is greater than the preset high quality threshold, the link quality weight is 1; when the ratio is less than the preset low quality threshold, the link quality weight is 0; when the ratio is between the high quality threshold and the low quality threshold, the link quality weight is calculated by linear interpolation.

6. The intelligent networking technology for large-scale equipment in ports according to claim 1, characterized in that, S3 specifically includes: S301: Using all the nodes of the devices to be networked as the vertex set, and the link quality weight between any two nodes in the current time slot as the edge weight, construct an undirected weighted network topology graph, and set the execution period of dynamic clustering. S302: The weighted agglomerative hierarchical clustering algorithm is used to divide the vertex set. The clustering is based on the dissimilarity between nodes, which combines the spatial Euclidean distance between nodes and the similarity of the link quality weight sequence. The dynamic equipment in the port operation area is divided into node clusters with stable internal link quality. S303: For each node cluster, calculate a comprehensive score by weighted summation of the node's communication stability index, topology centrality index, and remaining power index, and select the node with the highest comprehensive score as the main cluster head node of the cluster. S304: Within the same node cluster, a shadow cluster head node is elected from the remaining nodes other than the main cluster head node, based on a modified score that includes a spatial location difference penalty factor. S305: A synchronization control channel is established between the shadow cluster head node and the main cluster head node to perform real-time synchronization of the routing table; S306: The data link layer forwarding engine of the shadow cluster head node is configured to a logical silent mode, prohibiting the forwarding of non-control data packets until an activation instruction is received.

7. The intelligent networking technology for large-scale equipment in ports according to claim 6, characterized in that, In S304, when electing the shadow cluster head node, the spatial location difference penalty factor is positively correlated with the spatial distance between the candidate node and the main cluster head node. Nodes that are geographically far from the main cluster head node and have excellent communication capabilities are preferentially selected as shadow cluster head nodes.

8. The intelligent networking technology for large-scale equipment in ports according to claim 1, characterized in that, S4 specifically includes: S401: The main cluster head node reads the link quality weight sequence between itself and each member node in the cluster in the future consecutive time slots. When the predicted link quality weight of the future time slot is lower than the preset handover trigger threshold, or the sliding average slope of the link quality weight sequence is negative and the single step decrease exceeds the preset threshold, it is determined that the current main link will soon be unable to effectively carry service data and the handover preparation process is triggered. S402: The primary cluster head node encapsulates and generates a switching activation instruction frame, which is then sent to the shadow cluster head node via a three-way handshake confirmation mechanism through the synchronization control channel. S403: After successfully completing the three-way handshake confirmation, the shadow cluster head node switches the forwarding enable flag to logical true, activates the backup wireless transceiver channel, and broadcasts a backup link ready notification to the member nodes in the cluster. S404: The master cluster head node sends a concurrent dual-transmission command to all source nodes in the cluster. The source nodes copy the data packets to be sent and submit them to the main link transmission queue and the backup link transmission queue for parallel transmission. S405: During the concurrent dual-transmission transition, the primary cluster head node continuously monitors the carrying capacity of the current primary link. When it is determined that the primary link cannot maintain data carrying, it sends a primary link release announcement, the source node stops the data packet replication operation and switches to single-path sending mode, and the shadow cluster head node is promoted to the new primary cluster head node.

9. The intelligent networking technology for large-scale equipment in ports according to claim 8, characterized in that, The concurrent data packet transmission in S404 specifically includes: The source node assigns a unified primary / backup association sequence number in the protocol stack header extension field of the data packet to be sent, and adds a path identifier field to the header of the data packet; After receiving data packets from the primary link and the backup link, the receiving node at the target end performs deduplication based on the source address and primary / backup association sequence number in the data packet header, discards duplicate data packets, and sorts the unsorted data packets in the buffer in ascending order to offset the out-of-order order caused by the difference in propagation delay between the two links.

10. A large-scale intelligent networking technology for port equipment according to claim 8, characterized in that, In S405, a dual-criteria joint decision mechanism is used to determine that the main link cannot maintain data transmission: Physical layer criterion: The predicted Fresnel gap value of the next adjacent time slot is less than or equal to zero; Link layer criterion: Multiple consecutive data packets sent on the main link fail to receive acknowledgment frames from the receiver, and the actual measured value of the equivalent packet loss rate exceeds the preset tolerance limit. If any criterion is met, the main cluster head node determines that the main link has lost its effective data carrying capacity.

Images