An automated N-to-N control system and method for rail-mounted gantry cranes based on a resource pooling architecture.

By using an N-to-N control system based on a resource pooling architecture, the control console and the rail-mounted gantry are dynamically bound together, solving the problems of low resource utilization, system silos, and weak fault recovery capabilities in existing technologies. This enables efficient and flexible equipment scheduling and collaborative operation, improving the reliability and safety of the system.

CN123113575APending Publication Date: 2026-11-17SHANGHAI INTERNATIONAL PORT
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611451236.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-18
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing remote control systems for automated rail-mounted gantry cranes suffer from low resource utilization, system silos, poor scalability, and weak fault recovery capabilities. Furthermore, they cannot effectively address issues such as ultra-low latency, synchronization of video streams and control commands, and heterogeneity of industrial communication protocols, leading to inflexible equipment utilization and safety hazards.

Method used

The system adopts an N-to-N control system based on a resource pooling architecture. Through the core scheduling engine, it realizes dynamic binding and decoupling between the operator console and the rail-mounted gantry. Combined with dynamic task allocation, exception handling and industrial protocol adaptation, it breaks down physical and logical isolation and realizes flexible scheduling and unified control across batches and regions.

Benefits of technology

It improved resource utilization, eliminated information silos, enhanced fault recovery capabilities, reduced expansion costs, ensured high system reliability and security, and enabled flexible scheduling and collaborative operation of equipment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123113575A_ABST
    Figure CN123113575A_ABST
Patent Text Reader

Abstract

This invention discloses an N-to-N automated rail-mounted gantry control system and method based on a resource pooling architecture, belonging to the field of port automation control technology. The system includes a remote operation terminal layer, a core network layer, a service scheduling layer, and an equipment execution layer. The core lies in the ROCS server, which, through a built-in dynamic task allocation engine, incorporates N operator consoles and M rail-mounted gantry cranes into the operator console resource pool and rail-mounted gantry crane resource pool respectively, achieving logical decoupling and dynamic binding between the operator consoles and rail-mounted gantry cranes. The system supports dynamic switching between system-assisted mode and manual intervention mode, possesses an intelligent anomaly handling mechanism for process failure-preserving console jump-over and flexible transfer of equipment failures, achieves cross-batch network unification through switch stacking, SDN, or 5G private networks, and enables unified access for heterogeneous PLC devices through an industrial protocol adaptation layer. This invention increases operator console utilization from 45% to 82% and reduces equipment waiting time by 63%, making it suitable for the automated construction of smart port container terminals.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of port automation control technology, specifically to a remote centralized control and dynamic resource scheduling system and method for automated rail-mounted gantry cranes in container terminals. Background Technology

[0002] Automated rail-mounted gantry cranes are core equipment for intelligent operations at container terminals. Currently, remote control of automated rail-mounted gantry cranes typically employs a one-to-one or fixed-zone mode, meaning one remote control console controls a specific gantry crane. In terminals built in multiple phases (such as Phase I, Phase II, and Phase III), equipment from different phases usually operates in independent computer room and network environments, and uses its own independent automation software systems.

[0003] The aforementioned prior art has the following technical defects: First, resource utilization is low. The control panel is fixedly tied to the rail-mounted crane, resulting in fragmented control panel resources that cannot be flexibly allocated among various devices. This leads to significant uneven workloads and a serious waste of equipment resources.

[0004] Second, system silos. Due to incompatible network architectures and software versions, multiple batches of equipment form isolated information islands, preventing the control console from controlling different rail-mounted gantry cranes across batches and regions, thus hindering the improvement of the overall operational efficiency of the terminal.

[0005] Third, it has poor scalability. As the number of devices increases, the fixed pairing mode requires a proportional increase in operating stations, resulting in a waste of hardware resources and an inability to flexibly respond to scheduling needs during sudden failures or peak production periods.

[0006] Fourth, the fault recovery capability is weak. Once a certain control panel malfunctions, the rail-mounted crane it controls cannot be taken over by other control panels, which seriously affects the continuity of production.

[0007] Furthermore, port automation remote control scenarios present unique technical constraints that differ from those in the general IT scheduling field, making it impossible to directly port and apply load balancing and resource scheduling solutions from the general IT industry. (I) Ultra-low latency and deterministic constraints. The end-to-end latency from the issuance of control commands by remote operators to the response of the rail-mounted crane actuator must be controlled within a strict range. Remote control of a single rail-mounted crane requires network latency of less than 30 milliseconds; the allocation response time of the remote control console should be less than 200 milliseconds; in ultra-long-range control scenarios, the end-to-end latency must be controlled within 6 milliseconds. Once the latency exceeds the above thresholds, there will be a time difference between the screen seen by the operator and the actual equipment status, which may lead to the spreader colliding with containers, equipment damage, or even safety accidents. General IT industry load balancing solutions (such as web server request distribution) allow response latency of seconds or even longer, which does not meet the deterministic requirements of industrial real-time control.

[0008] (II) Strict Synchronization of Video Streams and Control Commands. Automated rail-mounted gantry cranes typically deploy 5 to 20 high-definition surveillance cameras. The remote control console needs to simultaneously receive multiple real-time video streams and PLC control data. Video streams and PLC commands must be strictly aligned on the timeline—the operator makes control decisions based on the real-time video feed, and the control commands are transmitted via the network to the equipment's PLC for execution. If the video stream and commands are misaligned (for example, the operator's view lags behind the actual state by 500 milliseconds), the operator's commands based on the lagging image will affect the already changed equipment state, potentially causing minor issues like misalignment of the grab / release box or even serious equipment collisions. Common scheduling solutions in the IT industry typically only focus on data transmission bandwidth and latency, without addressing the time synchronization constraints between video streams and control commands.

[0009] (III) Heterogeneity of Industrial Communication Protocols. Different manufacturers' rail-mounted gantry cranes use completely different industrial communication protocols. Siemens PLCs primarily use the Profinet / S7 protocol, while PLCs from brands like ABB and Rockwell use the EtherNet / IP / CIP protocol. The Siemens S7 protocol and AB's CIP protocol are incompatible at the application layer; even if both run on Ethernet, they cannot communicate directly. Replacing all PLC master station boards directly would be costly in terms of hardware and have an unacceptable timeframe. Existing general scheduling solutions in the IT industry lack the capability to adapt to these industrial protocols.

[0010] (iv) Physical constraints on operational safety and anti-sway control. During operation, the spreader and container of the rail-mounted gantry crane will sway due to the movement of the trolley and the crane itself. A stable real-time closed loop must be established between the issuance of remote operation commands and the actual actions of the equipment. Any uncertain scheduling delay may cause the electronic anti-sway system to fail, resulting in deviations in container handling.

[0011] In summary, no solution has been found in the existing technology that can simultaneously and systematically solve the above-mentioned technical defects. Although some literature involves console resource scheduling, none of them can simultaneously solve the problems of multi-batch network heterogeneity compatibility, video stream and control command synchronization, and abnormal handling order preservation. A new technical solution is urgently needed to improve this. Summary of the Invention

[0012] The purpose of this invention is to provide an automated N-to-N control system and method for rail-mounted cranes based on a resource pooling architecture, in order to solve the technical problems existing in the prior art, such as low utilization of operator console resources, system silos, poor scalability, and weak fault recovery capability.

[0013] This invention proposes an "N-to-N" (many-to-many) dynamic control system that breaks the original physical and logical isolation between batches, incorporates all operating consoles and all rail-mounted gantry cranes into a unified control resource pool, and realizes logical decoupling and dynamic binding between operating consoles and rail-mounted gantry cranes through a core scheduling engine.

[0014] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: An automated N-to-N control system for rail-mounted gantry cranes based on a resource pooling architecture includes: The remote operation terminal layer contains N remote operation consoles, where N is a positive integer greater than 1; The core network layer is used to interconnect multiple batches and regions of device networks to build a unified control and communication network. The service scheduling layer includes the ROCS server and the BMS system. The ROCS server has a built-in dynamic task allocation engine, a status monitoring module, and an exception handling module. The equipment execution layer contains M automated rail-mounted gantry cranes, where M is a positive integer greater than 1; The N remote control consoles and M automated rail-mounted gantry cranes register with the ROCS server, forming a console resource pool and a gantry crane resource pool, respectively. The ROCS server, through the dynamic task allocation engine, dynamically establishes temporary logical mapping relationships between the consoles in the console resource pool and the rail-mounted gantry cranes in the gantry crane resource pool based on the work instructions issued by the BMS system, thereby achieving dynamic control between the N consoles and the M gantry cranes. After the work is completed, the temporary logical mapping relationship is disconnected, and the consoles and gantry cranes return to their respective resource pools. The temporary logical mapping relationship is stored in the ROCS server in the form of a mapping relationship table, which records at least the console identifier, the gantry crane identifier, and the binding timestamp information.

[0015] Furthermore, the remote control console supports at least two operating modes: system-assisted mode and manual intervention mode, which can be switched automatically or manually depending on the task type.

[0016] Furthermore, the exception handling module is configured to: when a process failure occurs, force the current task to jump to the original operating console to maintain the context continuity of exception handling; when a device failure occurs, jump the current task to any idle operating console to achieve flexible fault transfer.

[0017] Furthermore, the dynamic task allocation engine employs at least one of the weighted round-robin algorithm, the shortest queue algorithm, or the least connection algorithm to dynamically select the operating console based on the console's idle state, current load, physical distance, and / or operator skill level factor; and when there is a preset binding strategy for the target rail-mounted crane, the load balancing algorithm is ignored, and the task is directly allocated to the operating console specified by the strategy.

[0018] Furthermore, the ROCS server also includes an industrial protocol adaptation layer, which is used to realize bidirectional conversion between at least two different industrial communication protocols, so as to achieve unified access and data interaction of heterogeneous PLC devices from different manufacturers.

[0019] Furthermore, the industrial protocol adaptation layer enables bidirectional conversion between the Profinet protocol and the EtherNet / IP protocol, achieving unified access and data interaction between Siemens series PLCs and ABB / Rockwell series PLCs.

[0020] Furthermore, the status monitoring module of the ROCS server monitors the online status of each operator station through a heartbeat mechanism: each operator station sends a heartbeat packet to the ROCS server at a preset time period, and if no heartbeat packet is received for a preset number of consecutive times, it is determined to be an offline fault; the preset time period is 100ms, and the preset number of times is 3 consecutive times.

[0021] Furthermore, the ROCS server also includes a video stream and control command time synchronization module, used to push the real-time video stream and PLC status data of the target rail crane to the target operator console. The video stream data packets and PLC command data packets both carry a unified timestamp. The time synchronization module performs time alignment verification on the video stream and PLC data to ensure that the time deviation between the video image and the actual status of the equipment does not exceed a preset delay threshold. The ROCS server also includes a manual takeover priority management module, used to query the current binding status of the target rail crane when the operator initiates a manual takeover request. If the target rail crane is in an idle state, takeover is allowed and the equipment is marked as "under manual takeover" to prevent the dynamic task allocation engine from automatically scheduling the equipment again. If the target rail crane is already occupied, the takeover request is rejected and a status prompt is returned.

[0022] This invention also proposes an automated N-to-N control method for rail-mounted gantry cranes based on a resource pooling architecture, comprising the following steps: S10, Network Unification and Resource Registration: A unified control and communication network is constructed by connecting multiple batches of physical networks through the core aggregation switch; N remote control consoles and M automated rail gantry cranes register with the ROCS server to form a console resource pool and a rail gantry crane resource pool, where N and M are both positive integers greater than 1; S20, Real-time Status Monitoring: The ROCS server monitors the occupancy status of each control panel and the task status, position, and operation mode of each rail-mounted crane in real time. S30, Dynamic Task Allocation: When the BMS system issues a work instruction, the dynamic task allocation engine of the ROCS server traverses the console resource pool and selects a target console based on at least one factor among idle status, current load, and physical distance, and establishes a temporary logical mapping relationship between the target console and the target rail gantry. S35, Abnormal Switchboard Handling: When a device failure occurs, the ROCS server will switch the current job task to any available workstation; when a process failure occurs, the ROCS server will force the current job task to switch back to the original workstation to ensure that the abnormal handling context is not lost. S40, Command forwarding and control: The ROCS server pushes the real-time video stream and PLC data of the target rail crane to the target control console, and forwards the operator's control commands to the target rail crane for execution; S50, Task completion and resource release: After the operation is completed, the temporary logical mapping relationship is disconnected, and the target control panel and the target rail crane return to their respective resource pools.

[0023] Furthermore, in step S35, the fault type is distinguished in the following ways: when the heartbeat of the control panel is interrupted and no heartbeat packet is received for a preset number of consecutive times, it is determined to be a device fault; when the abnormal task status code reported by the rail gantry crane belongs to the preset set of process fault codes, it is determined to be a process fault.

[0024] Furthermore, in step S40, the ROCS server uses the RTP / RTSP protocol to transmit video streams and the industrial Ethernet protocol to transmit PLC control commands; both the video stream data packets and the PLC command data packets carry a unified timestamp. The ROCS server performs time alignment verification on the operator console side to ensure that the time deviation between the video image and the actual state of the equipment does not exceed a preset delay threshold; the preset delay threshold is 100 milliseconds.

[0025] Beneficial effects Compared with the prior art, the present invention has the following beneficial effects: (1) Resource utilization has been greatly improved. The fixed binding relationship between the control panel and the rail-mounted gantry crane has been broken, enabling flexible scheduling across batches and regions. Actual test data shows that 4 control panels can efficiently control 10 rail-mounted gantry cranes, the average utilization rate of the control panels has increased from 45% to 82%, and the equipment waiting time has been reduced by 63%.

[0026] (2) Eliminate information silos. By using a unified core network layer and industrial protocol adaptation layer, the physical and logical isolation between multiple batches, multiple vendors and multiple versions of equipment is broken down, enabling collaborative operation of equipment across different periods and completely solving the problem of system silos.

[0027] (3) High reliability and fault self-healing capability. When a fault occurs at the operating station, the system will automatically detect it within 300ms and switch the task to another idle operating station without interrupting production; the intelligent sequence preservation mechanism for abnormal station jump ensures that the processing context of abnormal processes is not lost, and the fault transfer time is only about 8 seconds.

[0028] (4) Flexible expansion. New devices only need to be connected to the unified control communication network and registered with the ROCS server to join the resource pool and participate in scheduling. There is no need to reconfigure complex physical connections, which greatly reduces the cost of expansion.

[0029] (5) Reduce labor costs. Under the same workload (240 instructions / 8 hours), the number of operators can be reduced from 4 to 3, saving about 25% of labor costs.

[0030] (6) Sufficient quantitative verification. The system completed 1,000 fault injection tests under simulated port strong electromagnetic interference and vibration environment, with a fault judgment misjudgment rate and a fault judgment miss rate of 0%; the SHA-256 integrity verification failure rate was 0% during a continuous 12-month operation cycle, and the automatic trial recovery success rate reached 100%; the task allocation response time under the fully domestic software and hardware environment was basically the same as that under the imported environment (175ms→182ms, a difference of about 4%), providing sufficient empirical support for the technical effects of system decoupling, high reliability and independent controllability. Attached Figure Description

[0031] Figure 1 This is a topological diagram of the N-to-N control system in an embodiment of the present invention.

[0032] Figure 2 This is a flowchart of the N-to-N control method in an embodiment of the present invention.

[0033] Figure 1In the diagram: 100 is the remote operation terminal layer; 101, 102...10N are remote operation consoles; 200 is the core network layer; 201 is the core stacking switch; 300 is the service scheduling layer; 301 is the ROCS server; 302 is the BMS system; 400 is the equipment execution layer; 401, 402...40M are automated rail-mounted gantry cranes.

[0034] Figure 2 In the diagram: S10 is for network unification and resource registration; S20 is for real-time status monitoring; S30 is for dynamic task allocation; S35 is for handling abnormal platform jumps; S40 is for command forwarding and control; and S50 is for task termination and resource release. Detailed Implementation

[0035] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. The following embodiments are merely illustrative of the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention.

[0036] Terminology Definition The terms used in this invention are defined as follows: ROS: System-assisted mode, in which the system assists in completing most of the operation process, and the operator only intervenes at key nodes; MROS: Manual intervention mode, where the operator controls the entire process manually; ROCS: Remote Operation Control System, the core scheduling server in this invention; BMS: Yard Management System, responsible for issuing work instructions; OCR: Optical Character Recognition; PLC: Programmable Logic Controller, the underlying control unit of a rail-mounted gantry crane.

[0037] Example 1: Dynamic scheduling and automatic release under normal operating mode This embodiment describes how the system achieves dynamic matching and automatic release between the control panel and the rail-mounted gantry crane during normal production operations at the dock.

[0038] Scenario setting: The N-to-N control system of the present invention is deployed in a container terminal. The resource pool contains 4 remote control consoles (ROS#01 to ROS#04) and 10 automated rail-mounted gantry cranes (ARMG#01 to ARMG#10). The configuration ratio of control consoles to rail-mounted gantry cranes is 4:10.

[0039] The implementation steps are as follows: Step 1: Resource Registration. After all consoles and rail-mounted gantry cranes are started, they register with the ROCS server through the core stacking switch. The ROCS server records the IP address, current status (initially all are idle), and capability parameter set of each device, thus completing the initialization of the console resource pool and rail-mounted gantry crane resource pool.

[0040] Step 2, Task Trigger. The BMS system issues a stacking operation instruction for ARMG#05. The ROCS server status monitoring module confirms that ARMG#05 is in a remotely controllable state, and the instruction enters the allocation queue.

[0041] Step 3, Dynamic Allocation. In this invention, the storage method for the temporary logical mapping relationship is not limited to a hash table. It can also employ relational database tables, memory cache key-value pairs, or object mappings—data structures known in the art—as long as they enable fast querying and updating between the console ID and the rail crane ID. The dynamic task allocation engine traverses the console resource pool and selects ROS#02 using a weighted round-robin algorithm. The ROCS server establishes a temporary logical mapping: {Channel ID: ROS#02, Rail Crane ID: ARMG#05, Binding Time: T0}, and writes it to the memory database in key-value pair format. The status of ROS#02 and ARMG#05 is updated to "Job in Progress".

[0042] In the actual allocation process, the dynamic task allocation engine does not rely solely on the load balancing algorithm. Instead, it sequentially checks whether each operator console meets the task allocation conditions: ① The operator console's automatic task push switch is enabled; ② The operator console can control the equipment type of the target rail-mounted crane; ③ The operator console can accept the task type (ROS / MROS); ④ The operator console is idle (task status: no task, task completed, task terminated, task canceled, or task abnormally completed); ⑤ The operator console has not pressed the emergency stop button; ⑥ A user is logged into the operator console and has remote operation permissions. One operator console corresponds to one logged-in user, and user permissions are verified by the Unified User Management System (UMS); ⑦ The operator console is not connected to other rail-mounted cranes and is not in device browsing mode; ⑧ The operator console has control permissions for the target rail-mounted crane. Among the candidate operator consoles that meet all the above conditions, if a priority operator console is configured, the new task is preferentially allocated to that priority operator console; if no priority operator console is configured or the priority operator console is unavailable, the task is allocated according to the operator console's idle time from longest to shortest, thus ensuring that the task is allocated to the most suitable operator console.

[0043] Step 4, Job Execution. The ROCS server pushes the real-time video stream and PLC status data from the ARMG#05 to the ROS#02 operating interface. The video stream is transmitted using the RTP / RTSP protocol, and the PLC data is transmitted using the Profinet industrial Ethernet protocol. Both the video stream data packets and the PLC instruction data packets carry a unified timestamp. The ROCS server performs time alignment verification on the operator's side to ensure that the time deviation between the video image seen by the operator and the actual status of the equipment does not exceed 100 milliseconds. The operator uses the system-assisted mode to control the ARMG#05 to complete the stacking operation. The control commands are forwarded by the ROCS server to the PLC execution layer of the ARMG#05.

[0044] During the above-mentioned operation, the ROCS server also sends the binding signal between the operator console and the rail-mounted gantry crane, as well as real-time equipment information, to the video display system (CCTV). The CCTV automatically selects the screen switching scheme based on the binding information and equipment status, pushing multiple high-definition monitoring images of the target rail-mounted gantry crane to the target operator console, thereby realizing automatic linkage between video images and task binding.

[0045] Step 5, Automatic Release. After the job is completed, ARMG#05 sends a task completion signal to the ROCS server. The ROCS server automatically disconnects the logical mapping, and ROS#02 and ARMG#05 are reset to idle and return to their respective resource pools, waiting for the next allocation.

[0046] Technical effect In this embodiment, the four control consoles efficiently processed a total of 240 work instructions from 10 rail-mounted gantry cranes within 8 hours. The average utilization rate of the control consoles increased from 45% in the fixed pairing mode to 82%, and the average waiting time of the equipment was reduced by 63%.

[0047] Example 2: Failover and Abnormal Tripping in High Reliability Mode This embodiment describes the system's fault self-healing and abnormal order preservation mechanism when a sudden hardware failure occurs on the control panel or when the equipment experiences a process abnormality.

[0048] Scenario setting: The ROS#01 control panel is controlling the rail-mounted gantry crane ARMG#03 to perform a critical ship loading operation. The current logical mapping is {ROS#01 ↔ ARMG#03}.

[0049] Fault Scenario A (Equipment Failure Trip): ROS#01 suddenly interrupts its heartbeat signal due to a damaged monitor. The ROCS server's status monitoring module monitors the online status of each operator station through a heartbeat mechanism—each operator station sends a heartbeat packet to the ROCS server every 100ms. When the ROCS server fails to receive a heartbeat packet from an operator station three times consecutively, it determines that the operator station is in an offline fault state. In this scenario, the status monitoring module detects the loss of the ROS#01 heartbeat within 300ms and determines it as an equipment failure. The exception handling module triggers the equipment failure trip logic, queries the operator station resource pool, and finds that ROS#04 is in an idle state. The system automatically updates the binding relationship to {ROS#04 ↔ ARMG#03}. After the operator logs in to ROS#04, the interface automatically restores the real-time video and control of ARMG#03 within approximately 8 seconds, and the loading operation continues uninterrupted.

[0050] Fault Scenario B (Process Fault Preservation Jump): During truck operations, the OCR recognition result of ARMG#08 does not match the task order, triggering a process fault. The ROCS server's exception handling module distinguishes fault types as follows: when the operator's heartbeat is interrupted and no heartbeat packet is received for three consecutive times, it is determined to be an equipment fault; when the task status exception code reported by the rail gantry crane belongs to the preset process fault code set, it is determined to be a process fault. In this scenario, the exception code reported by ARMG#08 belongs to the process fault code set, triggering the preservation jump logic, which forcibly jumps the task to the original operator's station that initiated the operation. The operator at the original operator's station switches to manual intervention mode, verifies the OCR-captured image, manually enters the correct license plate number, and submits confirmation. The exception is cleared, and the process continues normally.

[0051] During task execution, the ROCS server maintains a complete task state machine: task generated but not assigned, assigned but not received, received and executing, task completed, task terminated (no longer assigned after termination), task canceled (cancelled but still assignable), and task abnormally completed. Task types include box grabbing tasks, box placing tasks, general fault tasks, severe fault tasks, trolley movement tasks, and equipment browsing tasks. When the rail-mounted gantry crane's single-machine automatic control system (ACCS) is in fault mode, the ROCS server generates a fault task (MROS task). If a semi-automatic operation task (ROS task) currently exists for the equipment, the ROS task is canceled first, and then the MROS task is generated. When the ACCS is in semi-automatic operation mode, the ROCS server generates a box grabbing task or a box placing task based on the equipment's open / closed status. Task status and task type information are reported to the ROCS server along with the equipment status as part of the exception code. The exception handling module uses this information to accurately distinguish between equipment faults and process faults, providing a basis for judgment for sequence-keeping jumpers and flexible transfers.

[0052] Technical effect The equipment failure transfer process took only about 8 seconds and did not cause any work interruption; the sequence-keeping jumper mechanism ensures that the same operator can continuously handle the same process exception, avoiding the loss of exception handling context information and greatly improving the integrity of the work.

[0053] Example 3: Inter-period collaboration and designated binding in hybrid batch mode This embodiment describes how the system can be compatible with heterogeneous equipment from multiple phases and support specific binding requirements in the scenario of old dock renovation.

[0054] Scenario setting: The first phase of a dock uses an old version of the Siemens PLC system (Profinet protocol) for its rail-mounted gantry cranes, while the second phase uses a new version of the ABB PLC system (EtherNet / IP protocol). The networks of the two phases are connected and unified through SDN software-defined networking technology.

[0055] The implementation steps are as follows: Step 1: Heterogeneous Device Registration. The ROCS server identifies and registers all rail-mounted gantry cranes in Phase I and Phase II through the industrial protocol adaptation layer. This protocol adaptation layer enables bidirectional conversion between the Profinet and EtherNet / IP protocols—Siemens PLC Profinet control commands are converted to EtherNet / IP commands via the adaptation layer and sent to ABB devices; conversely, ABB devices' EtherNet / IP status data is converted to Profinet data via the adaptation layer and reported to the ROCS server. This achieves unified access and management of devices from different manufacturers and using different protocols, forming a unified rail-mounted gantry crane resource pool.

[0056] Furthermore, the ROCS server communicates with the automatic control system (ACCS) of the rail-mounted gantry crane and the remote control system (RCCS) of the operator console via OPC to read and write point information, achieving synchronous sharing of real-time equipment status and operator console status. The ROCS server interacts with the video display system and data display system via TCP communication. In addition to bidirectional conversion between Profinet and EtherNet / IP protocols, the industrial protocol adaptation layer also supports access conversion for communication protocols such as OPC UA / OPC DA and Modbus TCP, to ensure compatibility with heterogeneous PLC devices from more manufacturers and further expand the range of devices eligible for unified access.

[0057] Step 2, specify the binding configuration. The administrator configures the specified binding policy in the ROCS server: for ARMG#01, force binding to console ROS#03.

[0058] Step 3, Dynamic Allocation Trigger. The BMS issues a job instruction for ARMG#01. The ROCS dynamic task allocation engine detects that ARMG#01 has a specified binding strategy, ignores the load balancing algorithm, and directly assigns the task to ROS#03, establishing a temporary logical mapping {ROS#03 ↔ ARMG#01}.

[0059] Step 4, Cross-Phase Collaborative Operation. After the operator of ROS#03 completes the unloading operation of ARMG#01, the resources are returned to the pool. The ROCS server continues to dynamically allocate operating consoles to other rail-mounted gantry cranes in Phase 1 or Phase 2 according to the subsequent tasks issued by the BMS, realizing unified scheduling and cross-phase collaborative operation between the old equipment in Phase 1 and the new equipment in Phase 2.

[0060] Technical effect This embodiment demonstrates that the present invention is fully compatible with rail-mounted gantry cranes of different brands and batches. Through a unified resource pooling architecture and dynamic scheduling mechanism, it achieves logically unified scheduling of multiple batches of equipment, completely eliminating the information silo problem caused by hardware generation differences and network isolation.

[0061] Example 4: Weighted Allocation of Operator Skill Levels This embodiment describes how to perform weighted allocation based on the operator's skill level during dynamic task assignment to achieve optimized allocation of human resources.

[0062] Scenario Setting: A dock resource pool contains 4 remote control consoles, each corresponding to 4 operators. The operator of ROS#01 is a senior operator (skill level 9 / 10), skilled in handling complex operations; the operator of ROS#02 is an intermediate operator (skill level 6 / 10); and the operators of ROS#03 and ROS#04 are junior operators (skill level 4 / 10).

[0063] The implementation steps are as follows: Step 1: Skill Level Registration. When each operator station registers with the ROCS server, in addition to submitting its IP address and device status, it also submits the corresponding operator skill level parameters. The ROCS server maintains a skill level field for each operator station in the operator station resource pool.

[0064] Step 2, Task Complexity Assessment. When the BMS system issues a work instruction, it also issues a task complexity label. The "Loading" operation is marked as high complexity (requires skill level ≥ 8), the "Stacking" operation is marked as medium complexity (requires skill level ≥ 5), and the "Shifting" operation is marked as low complexity (requires skill level ≥ 3).

[0065] Step 3, Weighted Allocation Calculation. When selecting an operator console, the dynamic task allocation engine calculates a weighted score: Allocation Score = Idle State Coefficient × Load Coefficient × Skill Level Matching Coefficient. The skill level matching coefficient is defined as follows: when the operator's skill level ≥ the task complexity requirement, the matching coefficient is 1.0; when the operator's skill level < the task complexity requirement, the matching coefficient is 0.

[0066] Step 4, Assign Execution. The BMS issues a highly complex "loading" operation instruction. The dynamic task assignment engine traverses the console resource pool, excluding ROS#02 (skill level 6 / 10<8), ROS#03, and ROS#04 (skill level 4 / 10<8). Only ROS#01 meets the conditions, and the task is assigned to ROS#01.

[0067] Technical effect This embodiment demonstrates that the present invention not only supports automatic allocation based on load balancing, but also supports refined allocation based on operator skill level, ensuring that complex tasks are performed by highly skilled operators, reducing operational risks and improving operational quality.

[0068] Example 5: Priority Mechanism for Manual Takeover This embodiment describes how the system handles priority conflicts between a takeover request and automatic scheduling when an operator initiates a manual takeover request.

[0069] Scenario setting: A dock resource pool contains 3 remote control consoles (ROS#01 to ROS#03) and 5 automated rail-mounted gantry cranes (ARMG#01 to ARMG#05). Currently, ROS#01 is controlling ARMG#03, ROS#02 is controlling ARMG#05, and ROS#03 is idle.

[0070] It should be noted that each console corresponds to one logged-in user, and user permissions are managed by the Unified User Management System (UMS). The system verifies whether the user has remote operation permissions and whether they are the only logged-in console. Consoles without logged-in users will not participate in task allocation. Before initiating manual takeover, the console can also observe the real-time status of the target device through the device browsing function; at any given time, only one console is allowed to browse a device, and while browsing, that console will not receive automatic task allocation from other consoles.

[0071] The implementation steps are as follows: Step 1: Manual Takeover Initiation. The operator of ROS#03 discovers abnormal jitter in ARMG#02 and decides to initiate a manual takeover request. The operator clicks the "Manual Takeover" button on the ROS#03 interface and selects the target device ARMG#02.

[0072] Step 2, Takeover Request Processing. Upon receiving the takeover request, the ROCS server's manual takeover priority management module queries the current binding status of ARMG#02 in the in-memory database. The query shows that ARMG#02 is currently idle. The manual takeover priority management module allows the takeover, establishes a new temporary logical mapping {ROS#03 ↔ ARMG#02}, and updates the status of ARMG#02 in the in-memory database to "Manual Takeover in Progress".

[0073] Step 3: Prevent automatic scheduling conflicts. While ARMG#02 is in the "Manual Takeover" state, the ROCS server's dynamic task allocation engine skips devices in the "Manual Takeover" state when traversing the rail gantry resource pool, ensuring that ARMG#02 is not automatically assigned to other operators and avoiding resource conflicts.

[0074] Step 4, Manual Takeover and Release. After the operator completes manual control of ARMG#02 on ROS#03, they click the "Release" button. The ROCS server disconnects the logical mapping, and the status of ARMG#02 is reset from "Manual Takeover" to "Idle," returning it to the rail-mounted gantry resource pool, awaiting the next dynamic allocation.

[0075] Conflict scenario: If the operator of ROS#03 attempts to take over ARMG#03 (which is currently occupied by ROS#01), the manual takeover priority management module will find that the binding status of ARMG#03 is "in operation", reject the takeover request, and return the message "Target device is occupied and cannot be taken over" on the operation interface of ROS#03.

[0076] Technical effect: This embodiment demonstrates that the manual takeover priority mechanism of the present invention can effectively handle the operator's proactive takeover needs, while preventing resource conflicts between automatic scheduling and manual takeover through status marking, thus ensuring the timeliness of manual intervention and the orderliness of system scheduling.

[0077] Example 6: System Performance and Reliability Verification This embodiment aims to verify the technical effectiveness and reliability of the system described in this invention by using quantitative data.

[0078] 1. Verification of ROCS server resource consumption and decoupling effect To verify the technical effectiveness of system decoupling, a comparative test was conducted under a co-deployment scheme with business nodes. The resource consumption of the ROCS server under the same workload of 8 hours and processing 240 instructions is shown in the table below:

[0079] Test results show that the present invention significantly reduces the resource overhead of the core scheduling node through logical decoupling and resource pooling, avoids competing for resources with the business system, and ensures the stability and real-time performance of the control system.

[0080] 2. Verification of the accuracy of heartbeat detection and fault classification Fault injection tests were conducted under simulated port electromagnetic interference and vibration conditions to verify the accuracy of the status monitoring and anomaly handling modules. The results of a total of N=1000 fault injection tests are as follows: Control panel equipment failure (heartbeat interruption): After 500 injections, the system accurately identified the equipment failure within 300ms (three consecutive 100ms heartbeat losss), with a false positive rate of 0% and a false negative rate of 0%.

[0081] Process failure (OCR anomalies, etc.): After 500 injections, the system can accurately determine the process failure based on the preset set of anomaly codes and trigger the order-keeping jumper logic, with a false positive rate of 0% and a false negative rate of 0%.

[0082] The fault type differentiation adopts a dual mechanism of "heartbeat detection + service status detection": when the operator's heartbeat is interrupted, it is determined to be a device fault, triggering a flexible switchover; when the operator's heartbeat is normal but the task status abnormal code reported by the rail gantry belongs to the preset process fault code set, it is determined to be a process fault, triggering a sequence-preserving switchover. The dual mechanism, together with the preset abnormal code set, achieves accurate differentiation between device faults and process faults, avoiding false fault judgments and incorrect switching.

[0083] 3. Data integrity and automatic recovery mechanism verification SHA-256 Integrity Verification: Over a 12-month operating cycle, the system performed SHA-256 verification on critical control commands and status data, verifying over 1 billion data packets with a 0% failure rate, demonstrating high data transmission integrity.

[0084] Automatic trial recovery test: Under simulated instantaneous failure scenarios such as network outages, the system automatically triggered 1200 trial recovery operations within 12 months, with a success rate of 100%, effectively avoiding unnecessary task interruptions and switching.

[0085] 4. Sliding window strategy and storage optimization To optimize the storage and retrieval of status data, this invention employs a sliding window strategy to record the binding status between the control panel and the rail-mounted crane. The effects compared to the traditional fixed window strategy are as follows:

[0086] 5. Three-tiered response strategy and switching overhead The server where ROCS is located employs a three-level handling strategy (application-level retry, virtual machine-level restart, and backup / recovery-level switchover) in its anomaly handling module. Based on 12 months of real-world testing data, the triggering behavior at each level is as follows: Application-level retries: 10 times in 12 months. The vast majority of transient exceptions are successfully handled at this level and do not require escalation.

[0087] Virtual machine-level reboots: occurred 3 times within 12 months.

[0088] Backup and recovery level switch: This did not occur in actual operation over 12 months; its availability was only verified in drill scenarios.

[0089] The three-level handling state transition logic is as follows: Under normal operating conditions, if an application process is detected to be abnormal or unresponsive, the system enters the first-level handling phase and automatically restarts the application service. Each level of handling has a maximum number of retries (3 times), and the system will only escalate to the next level if the threshold is exceeded. If the first-level handling is successful and the health check passes, the system directly returns to the normal operating state without escalating to the second level. If the first-level handling fails (the system remains abnormal after 3 consecutive restarts) or an operating system crash is detected, the system escalates to the second-level handling phase and automatically restarts the virtual machine. If the second-level handling is successful (the operating system starts normally, the network is connected, and basic services are ready), the internal business applications are automatically restored and the system returns to the normal operating state without escalating to the third level. If the second-level handling fails (the system fails to restart 3 times) or the system data is abnormal, the system escalates to the third-level handling phase, performs a backup and recovery level switch, restores key configurations and mapping relationships from backup data, and returns to the normal operating state after business verification is passed. Through the "success without escalation" state transition design, most transient anomalies are digested at the application level, effectively reducing the overall switching overhead.

[0090] Data shows that the strategy of "not upgrading after a successful attempt after setting a maximum number of retries" effectively reduces the overall switching overhead, and most anomalies (about 77%) are resolved at the application level, ensuring the stable operation of the system.

[0091] 6. Verification of Layered Switchover Time and RTO / RPO Metrics To achieve the goals of overall handover time ≤ 10 minutes, RPO ≤ 5 minutes, RTO ≤ 10 minutes, and success rate ≥ 99.99%, we conducted a statistical analysis of the time consumption of each stage of the layered handover, and the results are as follows:

[0092] Actual test data shows that the total switching time at each level is far below the theoretical upper limit of 10 minutes, providing a solid guarantee for achieving high reliability and business continuity.

[0093] The switching success rate of ≥99.99% comes from testing and statistics at three levels: (1) Test environment verification: In a test environment with the same configuration as the production environment, more than 10,000 switching tests were performed for various scenarios such as console equipment failure, process failure, ROCS server failure, core network failure, and console batch failure. Among them, 9,999 switching tests were successful and 1 failed (due to data inconsistency caused by the manual removal of storage media in the test environment, which has been repaired by adding a data consistency verification mechanism before switching). The switching success rate of the test environment was 99.99%; (2) Simulation environment stress test: In the simulation environment The system simulated the peak port workload (10 rail-mounted gantry cranes operating at full capacity, 240 commands / 8 hours, and full video stream load), and performed more than 5,000 fault injection and switching tests, covering various fault types such as CPU exhaustion, memory exhaustion, network interruption, heartbeat interruption, and ROCS server power failure. It successfully switched 4,999 times, with a switching success rate of 99.98% in the simulation environment. (3) Actual operation statistics: During the 12-month pilot operation of this system at a large container terminal, a total of 13 abnormal handling and automatic switching events occurred, all of which were successfully handled without business interruption, and the actual operation switching success rate was 100%. Based on the comprehensive test environment, simulation environment, and actual operation statistics, the switching success rate of this system is no less than 99.99%.

[0094]

[0095] The above measured data shows that the overall switching time under various fault scenarios is far better than the 10-minute RTO, and the RPO is 0, which verifies the feasibility of quantitative indicators such as overall switching time ≤ 10 minutes, RPO ≤ 5 minutes, RTO ≤ 10 minutes, and switching success rate ≥ 99.99%.

[0096] Example 7: Field Adaptation in Industrial Control Scenarios for Port Rail-Mounted Cranes This embodiment further illustrates the field adaptation data and security measures of the present invention in a fully domestically produced software and hardware environment, in order to support its technical effectiveness in specific application scenarios.

[0097] 1. Performance Comparison of Full-Stack Domestic Environment This invention has been deployed at a large container terminal, with its core service scheduling layer running in a fully domestically produced environment (Kunpeng 920 CPU + Kylin V10 operating system + DM database). In comparative tests with the original imported hardware environment (Intel Xeon CPU + Windows Server + Oracle database), the performance is as follows:

[0098] Tests show that the present invention suffers minimal performance loss under domestic production conditions, fully meets the real-time requirements of port production operations, and has good independent controllability.

[0099] 2. List of Security Measures for Level 3 Compatibility To meet the requirements of Level 3 Cybersecurity Classification Protection, this invention adopts the following specific security measures in system design and deployment: Network communication security: The core network layer uses SDN technology for logical isolation, dividing the network into production control area, security management area, and external interface area. Cross-area access is strictly controlled through industrial firewalls, opening only necessary communication ports and protocols.

[0100] Identity Authentication and Access Control: Both the ROCS server and remote console employ two-factor authentication based on digital certificates. The system implements a separation of powers (system administrator, security officer, and security auditor), and assigns operational permissions to different roles according to the principle of least privilege.

[0101] Security Audit: The system has a built-in security audit module that fully records user logins, logouts, critical operations (such as manual takeover and parameter modifications), and system anomalies. Audit logs employ tamper-proof technology and are backed up regularly, with a retention period of no less than 6 months.

[0102] Intrusion prevention: Deploy an industrial intrusion detection system at the core network layer to monitor and block malicious attacks targeting the PLC protocol and ROCS service in real time.

[0103] Data integrity and confidentiality: Critical control commands and status data are signed and encrypted using national cryptographic algorithms during transmission to ensure data integrity and confidentiality.

[0104] Through the above measures, the system successfully passed the Level 3 Information Security Protection Assessment, proving its security and reliability in industrial control scenarios.

[0105] During the 12-month actual operation period of this system, a total of 13 abnormal handling and automatic switching events occurred, including 10 application-level retries and 3 virtual machine-level restarts. In actual operation, no backup and recovery-level switching was triggered, all of which were successfully handled and there was no business interruption. The actual switching success rate was 100%. This statistic corroborates the three-level handling data in Examples 6-5, further verifying the high reliability of this system in port production scenarios.

[0106] Key comparative experimental data Under the same dock conditions, with the same workload (240 orders / 8 hours) and the same equipment configuration (10 rail-mounted gantry cranes), data was collected for 7 consecutive days and the average value was taken. The comparison test results between the present invention and the existing fixed zoning mode are as follows:

[0107] Comparison of the performance of three scheduling algorithms (under the same conditions):

[0108] Actual measured recovery time of faulty trolley:

[0109] Other alternative implementation schemes (1) Network architecture alternatives: In addition to using switch stacking technology, the core network layer can also adopt a dual-active data center architecture, SDN software-defined network technology, or 5G uRLLC private network to achieve wireless control.

[0110] (2) Alternative scheduling algorithms: In addition to allocation based on idle state, dynamic task allocation can also be weighted based on the energy consumption state of the rail gantry, or the shortest queue algorithm or the minimum number of connections algorithm can be used.

[0111] (3) Hardware configuration alternatives: The number of operating consoles N and the number of rail gantry cranes M can be adjusted according to the actual scale of the wharf. The ratio of N to M is not fixed, and this invention does not impose any restrictions on it.

[0112] (4) ROCS deployment alternative: ROCS servers can be deployed in a dual-machine hot standby or cluster mode. When the main server fails, the standby server automatically takes over the scheduling service to ensure that the scheduling is not interrupted.

[0113] Industrial applicability This invention is applicable to automated rail-mounted gantry control systems for various container terminals, and is particularly suitable for smart port construction scenarios with requirements for multiple batches, multiple devices, high concurrency, and high reliability. This invention has been successfully deployed and verified at a large container terminal, demonstrating excellent operational performance and significant industrial practical value.

Claims

1. An automated N-to-N control system for rail-mounted gantry cranes based on a resource pooling architecture, characterized in that, include: The remote operation terminal layer contains N remote operation consoles, where N is a positive integer greater than 1; The core network layer is used to interconnect multiple batches and regions of device networks to build a unified control and communication network. The service scheduling layer includes the ROCS server and the BMS system. The ROCS server has a built-in dynamic task allocation engine, a status monitoring module, and an exception handling module. The equipment execution layer contains M automated rail-mounted gantry cranes, where M is a positive integer greater than 1; The N remote control consoles and M automated rail-mounted gantry cranes are registered with the ROCS server to form a console resource pool and a rail-mounted gantry crane resource pool, respectively. The ROCS server uses the dynamic task allocation engine to dynamically establish a temporary logical mapping relationship between the consoles in the console resource pool and the rail-mounted gantry cranes in the rail-mounted gantry crane resource pool according to the work instructions issued by the BMS system, thereby realizing dynamic control between the N consoles and the M rail-mounted gantry cranes. After the operation is completed, the temporary logical mapping relationship is disconnected, and the control panel and the rail gantry return to their respective resource pools. The temporary logical mapping relationship is stored in the ROCS server in the form of a mapping relationship table, which records at least the control panel identifier, the rail gantry identifier, and the binding timestamp information.

2. The system according to claim 1, characterized in that, The remote control console supports at least two operating modes: system-assisted mode and manual intervention mode. The two operating modes can be switched automatically or manually depending on the task type.

3. The system according to claim 1, characterized in that, The exception handling module is configured to: when a process failure occurs, force the current task to jump back to the original operation console in order to maintain the context continuity of exception handling; When equipment failure occurs, the current task will be switched to any available workstation to achieve flexible fault transfer.

4. The system according to claim 1, characterized in that, The dynamic task allocation engine employs at least one of the weighted round-robin algorithm, the shortest queue algorithm, or the least connection algorithm to dynamically select the operating console based on the console's idle state, current load, physical distance, and / or operator skill level factor. Furthermore, when a preset binding strategy for the target rail-mounted crane exists, the load balancing algorithm is ignored, and the task is directly allocated to the operating console specified by the strategy.

5. The system according to claim 1, characterized in that, The ROCS server also includes an industrial protocol adaptation layer, which is used to realize bidirectional conversion between at least two different industrial communication protocols, so as to realize unified access and data interaction of heterogeneous PLC devices from different manufacturers; the industrial protocol adaptation layer realizes bidirectional conversion between Profinet protocol and EtherNet / IP protocol, so as to realize unified access and data interaction between Siemens series PLCs and ABB / Rockwell series PLCs.

6. The system according to claim 1, characterized in that, The status monitoring module of the ROCS server monitors the online status of each operator station through a heartbeat mechanism: each operator station sends a heartbeat packet to the ROCS server at a preset time period, and if no heartbeat packet is received for a preset number of consecutive times, it is determined to be an offline fault; the preset time period is 100ms, and the preset number of times is 3 consecutive times; the ROCS server also includes a video stream and control command time synchronization module, which is used to push the real-time video stream and PLC status data of the target rail crane to the target operator station. The video stream data packet and the PLC command data packet both carry a unified timestamp. The time synchronization module performs time alignment verification on the video stream and PLC data to ensure that the time deviation between the video image and the actual status of the equipment does not exceed a preset delay threshold.

7. The system according to claim 1, characterized in that, The ROCS server also includes a manual takeover priority management module, which is used to query the current binding status of the target rail crane when the operator initiates a manual takeover request. If the target rail crane is in an idle state, takeover is allowed and the device is marked as "under manual takeover" to prevent the dynamic task allocation engine from automatically scheduling the device again. If the target rail crane is already occupied, the takeover request is rejected and a status prompt is returned.

8. A method for N-to-N control of an automated rail-mounted crane based on a resource pooling architecture, characterized in that, Includes the following steps: S10, Network Unification and Resource Registration: A unified control and communication network is constructed by connecting multiple batches of physical networks through the core aggregation switch; N remote control consoles and M automated rail gantry cranes register with the ROCS server to form a console resource pool and a rail gantry crane resource pool, where N and M are both positive integers greater than 1; S20, Real-time Status Monitoring: The ROCS server monitors the occupancy status of each control panel and the task status, position, and operation mode of each rail-mounted crane in real time. S30, Dynamic Task Allocation: When the BMS system issues a work instruction, the dynamic task allocation engine of the ROCS server traverses the console resource pool and selects a target console based on at least one factor among idle status, current load, and physical distance, and establishes a temporary logical mapping relationship between the target console and the target rail gantry. S35, Abnormal Switchboard Handling: When a device failure occurs, the ROCS server will switch the current job to any available workstation. When a process failure occurs, the ROCS server will forcibly redirect the current job to the original console to ensure that the exception handling context is not lost. S40, Command forwarding and control: The ROCS server pushes the real-time video stream and PLC data of the target rail crane to the target control console, and forwards the operator's control commands to the target rail crane for execution; S50, Task completion and resource release: After the operation is completed, the temporary logical mapping relationship is disconnected, and the target control panel and the target rail crane return to their respective resource pools.

9. The method according to claim 8, characterized in that, In step S35, the fault type is distinguished in the following ways: when the heartbeat of the control panel is interrupted and no heartbeat packet is received for a preset number of consecutive times, it is determined to be a device fault; when the abnormal task status code reported by the rail gantry crane belongs to the preset set of process fault codes, it is determined to be a process fault.

10. The method according to claim 8, characterized in that, In step S40, the ROCS server uses the RTP / RTSP protocol to transmit video streams and the industrial Ethernet protocol to transmit PLC control commands. Both the video stream data packets and the PLC instruction data packets carry a unified timestamp. The ROCS server performs time alignment verification on the operator console side for the video stream and PLC data to ensure that the time deviation between the video image and the actual state of the device does not exceed a preset delay threshold.