5g-based remote real-time control method, system and device for port equipment and medium

Abstract

The application discloses a 5G-based port equipment remote real-time control method, system, device and medium, and particularly relates to the technical field of port automation control, and comprises the following steps: a global synchronization time reference is established through a 5G time-sensitive network, and a global control cycle is divided into three non-overlapping time slots of state uploading, decision scheduling and instruction execution; channel state information is extracted based on a port equipment uplink sounding reference signal, a device motion state is solved to generate control sensing information; a control instruction sequence with an absolute execution timestamp is generated, communication resources are allocated and transmitted according to instruction priority and channel quality difference; the device accurately executes the instruction according to the timestamp, and simultaneously predicts a device motion trend based on the channel state information to actively correct the instruction to realize predictive anti-collision. The application improves the synchronization accuracy, instruction execution certainty and operation safety of port remote control, and realizes multi-device high-precision collaborative operation.

Description

Technical Field

[0001] This invention relates to the field of port automation control technology, and more specifically, to a method, system, device, and medium for remote real-time control of port equipment based on 5G. Background Technology

[0002] In port automation operations, remote real-time control is a key technology for achieving efficient collaboration among equipment such as quay cranes, yard cranes, automated guided vehicles (AGVs), and unmanned container trucks. Currently, remote control systems for port equipment typically use wired communication, Wi-Fi, or 4G for data transmission and command issuance. Remote control systems based on 4G LTE usually employ a periodic uplink status reporting and downlink command issuance mechanism. Time synchronization between the base station and the equipment is achieved via NTP or PTP. Control commands are issued immediately after generation, and the equipment executes them immediately upon receipt, lacking precise planning of execution timing.

[0003] Existing port equipment control methods can meet control requirements, but they still have some shortcomings in actual use: Insufficient time synchronization accuracy: The traditional PTP or NTP synchronization method used in the existing system usually has a clock synchronization error between microseconds and milliseconds in the complex electromagnetic environment of the port. It is difficult to meet the requirements of high-precision collaborative operation of multiple devices within a millisecond control cycle, which can easily lead to disordered command execution timing and affect operational safety and efficiency. Uncertain command execution time: The existing system adopts the "receive and execute" command issuance mode. Due to the jitter of air interface transmission delay, there is a deviation in the actual time when different devices receive and execute commands within the same cycle. It is impossible to realize that multiple devices can perform actions under a strictly synchronized time reference, which limits the accuracy of port equipment collaborative operation. Equipment status perception relies on external sensors: Existing systems rely on external sensors such as GPS, lidar, and millimeter-wave radar to acquire equipment status such as position and speed. This not only increases the system hardware cost, but also makes it easy to lose signals or reduce perception accuracy in complex port environments (such as container obstruction, multipath reflection, and weather effects), affecting the reliability of control. Lack of a joint scheduling mechanism for communication and control: In the existing system, the issuance of control commands and the allocation of air interface resources are independent of each other. There is no differentiated scheduling based on command priority and channel quality. This can lead to transmission delays or failures for high real-time commands (such as emergency braking) when the channel quality is poor, which poses a security risk. Collision avoidance mechanisms rely on the current state and lack predictive capabilities: Existing systems typically determine collision avoidance strategies based on the current equipment position, triggering emergency braking when the detected distance is less than a safety threshold. This approach lacks the ability to predict future movement trends and cannot proactively correct control commands before a collision risk occurs, easily leading to frequent emergency braking, impacting operational efficiency and equipment lifespan.

[0004] In summary, existing remote control systems for port equipment have significant shortcomings in terms of time synchronization accuracy, deterministic command execution timing, equipment status perception methods, coordinated scheduling of communication and control, and collision risk prediction capabilities. There is an urgent need for a remote real-time control method that can achieve high-precision time synchronization, deterministic command execution, joint optimization of communication and control, and has motion trend prediction and active collision avoidance capabilities. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method, system, equipment and medium for remote real-time control of port equipment based on 5G, and solves the problems mentioned in the background art through the following solutions.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a 5G-based remote real-time control method for port equipment, the method comprising: S1: Establish a global synchronization time reference between the base station and port equipment through a 5G time-sensitive network, and set a fixed global control cycle. Divide each global control cycle into non-overlapping status upload time slots, decision scheduling time slots, and instruction execution time slots. S2: During the state upload time slot, each port device sends an uplink detection reference signal to the base station. The base station extracts the channel state information of the detection reference signal and sends it to the edge computing server. The edge computing server calculates the relative position, relative speed and relative acceleration of the port device based on the channel state information to obtain control sensing information. S3: Within the decision scheduling time slot, the edge motion planner generates a sequence of control instructions containing absolute execution timestamps based on the control perception information and the control instructions from the remote control center. The absolute execution timestamps are the start times of the instruction execution time slots pointing to the future global control cycle. S4: The control and communication joint scheduler in the base station allocates physical resource blocks for the control command sequence according to the command priority and the current air interface channel quality, and completes the transmission to the port equipment within the time window of the absolute execution timestamp; S5: After receiving the control command sequence, the port equipment executes the commands in the control command sequence according to the start time of the absolute execution timestamp; S6: Based on the channel state information obtained within the global control cycle, predict the motion trend of each port device within the future time window. When it is predicted that the minimum distance between any two port devices is less than the preset safety threshold, the edge motion planner automatically corrects the control instruction sequence that has not yet been executed, generates correction instructions including deceleration, stopping or trajectory fine-tuning, and issues the correction instructions for execution according to S3 to S5.

[0007] Preferably, the 5G-based remote real-time control system for port equipment is characterized by comprising: Global time synchronization and cycle division module: Deployed in 5G base stations and edge computing servers, it establishes a global synchronization time reference between port base stations and equipment through 5G time-sensitive network, and sets a global control cycle. The global control cycle is divided into three time slots, namely the status upload slot, the decision scheduling slot, and the instruction execution slot. Sensing and State Calculation Module: In the state upload slot of the global control cycle, it receives the uplink detection reference signal sent by the port equipment, extracts the channel state information, and calculates the relative position, relative speed and relative acceleration of each port equipment based on the channel state information to generate control sensing information; Edge motion planning module: Deployed on the edge computing server, within the decision scheduling slot of the global control cycle, it receives the control perception information and control instructions from the remote control center, and generates a sequence of control instructions carrying an absolute execution timestamp, which points to the start time of the instruction execution time slot of the future global control cycle; Control and communication joint scheduling module: Deployed in the 5G base station, within the decision scheduling slot, it allocates physical resource blocks according to the priority of the instructions in the control instruction sequence and the current channel quality, and performs downlink transmission to the port equipment within the time window of the absolute execution timestamp; Device-side timed execution module: Receives the control instruction sequence, uses a local hardware timer to trigger the execution of the instructions in the control instruction sequence based on the start time of the absolute execution timestamp, and feeds back the execution information to the edge computing server; Predictive Collision Avoidance Correction Module: Deployed on an edge computing server, it predicts the movement trend of port equipment within a future time window based on the acquired channel state information. When it predicts that the minimum distance between any two port equipment is less than a preset safety threshold, it automatically corrects the unexecuted control command sequence and generates a correction command.

[0008] Preferably, an electronic device includes: a processor, a communication bus, a user interface, a network interface, and a memory. The memory stores an executable program, which, when executed by the processor, causes the electronic device to execute the computer program of the aforementioned 5G-based remote real-time control method for port equipment.

[0009] Preferably, a computer-readable storage medium is characterized in that a computer program is stored on the computer-readable storage medium, and the computer program is executed by a processor as the computer program of the above-mentioned 5G-based remote real-time control method for port equipment.

[0010] The technical effects and advantages of this invention are as follows: 1. This invention establishes a global synchronization time reference in S1 based on 5G time-sensitive network and IEEE1588v2PTP protocol. The base station has a built-in clock source and periodically broadcasts time synchronization signals, so that the synchronization error between the local clock of the port equipment and the base station clock does not exceed ±100ns. At the same time, the global slot counter and periodic start pulse are used to achieve periodic alignment, thus solving the clock deviation problem when multiple devices work together. 2. In S3, the edge motion planner generates a sequence of control instructions carrying an absolute execution timestamp, and binds the execution time to the start time of the instruction execution time slot of a future global control cycle. In S5, the port equipment uses a hardware timer locked to the global clock to precisely trigger the interrupt execution instruction when the absolute execution timestamp arrives, thereby improving the accuracy of collaborative operations. 3. In S4 of this invention, the joint control and communication scheduler within the base station determines the priority of control commands ( Based on the current air interface channel quality, Physical Resource Blocks (PRBs) are allocated differently for commands of different priorities: at least 4 consecutive PRBs are reserved for Level 0 urgent commands, and high real-time commands are allocated accordingly. The system generates one PRB and calculates the transmission deadline based on the absolute execution timestamp, reserving resources in advance and completing the transmission within the time window, thereby improving the reliability and timeliness of high real-time command transmission. 4. In S6 of this invention, the edge computing server utilizes the device status and CSI features stored in historical cycles to output the trajectory sequence of each device within the future time window through lightweight motion prediction models such as Kalman filtering and LSTM. When the minimum distance between any two devices is predicted to be less than the safety threshold, it automatically generates correction instructions for low-priority devices to decelerate, stop, or fine-tune their trajectories, thereby improving work efficiency and device safety. Attached Figure Description

[0011] Figure 1 This is a schematic diagram of the overall method structure of the present invention.

[0012] Figure 2 This is a schematic diagram of the establishment and period division of the global synchronization time base in S1 of the present invention.

[0013] Figure 3 This is a schematic diagram of the port equipment execution command structure of S5 of the present invention.

[0014] Figure 4 This is a schematic diagram of the overall system structure of the present invention.

[0015] Figure 5 This is a schematic diagram of the electronic device of the present invention. Detailed Implementation

[0016] 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.

[0017] Please see Figures 1-3 As shown, this embodiment of the invention provides a remote real-time control method for port equipment based on 5G, the method comprising: S1: Establish a global synchronization time reference between the base station and port equipment through a 5G time-sensitive network, and set a fixed global control cycle. Divide each global control cycle into non-overlapping status upload time slots, decision scheduling time slots, and instruction execution time slots. This embodiment establishes a global synchronization time reference through a 5G Time Sensitive Network (TSN) and sets a fixed global control period. Each The time slot is divided into three consecutive slots, specifically including: S101: Deploy a high-precision time synchronization network Within the port operation area, deploy one or more base stations supporting 5G Time Sensitive Network (TSN) functionality. Each base station integrates a high-precision clock source, which preferentially uses: A boundary clock or transparent clock based on the IEEE 1588v2PTP precision time protocol is synchronized with the upstream master clock via optical fiber. Directly receiving timing signals from the Global Navigation Satellite System (GNSS) ensures that the deviation between the base station clock and UTC time is less than [a certain value]. ; The base station periodically sends time synchronization signals to all port equipment (quay cranes, yard cranes, AGVs, unmanned container trucks, etc.) within its coverage area via 5G broadcast signaling. Each port device's 5G communication module automatically parses the broadcast signaling and adjusts its local clock to synchronize with the base station clock. After initial synchronization, the clock error between all devices does not exceed [a certain value]. It meets the sub-microsecond synchronization requirements of port equipment under millisecond-level control cycles.

[0018] S102: Define the global control cycle In edge computing servers, a fixed global control cycle is pre-configured. Global control cycle The value is determined based on the maximum speed of the port equipment, the control accuracy requirements, and the 5G air interface frame structure, and is preferably... ; Global control cycle The value is kept as an integer multiple of the 5G air interface wireless frame period to facilitate time slot alignment. The edge server broadcasts the value via the 5G downlink broadcast channel. The values ​​and the start boundary time of each cycle are transmitted to the port equipment and control center.

[0019] S103: Global control cycle Divided into three consecutive time slots Each global control cycle The timeline is divided into three non-overlapping and consecutive time slots, namely: the state upload slot. Decision scheduling slot Instruction execution slot ; The length allocation of the time slot is preset based on the data volume, computing latency, and transmission requirements of the port equipment, and can be dynamically adjusted. hour: Length is It is used for all port equipment to send sensor data and detection reference signals uplink; Length is It is used for edge computing, collision avoidance prediction, and resource scheduling decisions; Length is It is used for the execution of downlink transmission control commands.

[0020] The switching between the three time slots is controlled by a global slot counter jointly maintained by the base station and the edge server. The counter is based on... The cycle increments, and each port device can automatically identify the time slot to which the current moment belongs within the local synchronized clock.

[0021] S104: Functional allocation and synchronization mechanism of time slots Status upload slot During this time slot, all port equipment is permitted to perform uplink transmissions, sending their sensor data and 5G probe reference signals. Packing, in After initiation, the data is sent to the base station; the base station utilizes... The received data is processed. All uplink transmissions must be terminated at the end to avoid interfering with the next time slot; Decision scheduling slot During this time slot, base stations and edge computing servers perform core computing tasks, including: From the received probe reference signal Extracting channel state information Solve the motion state of the equipment; Run the collision avoidance prediction algorithm to generate correction instructions; The joint control and communication scheduler allocates the next time slot based on command priority and channel quality. Downlink transmission allocation physical resource blocks .

[0022] Within this channel, port equipment does not transmit any information upstream, but it can receive some broadcast information. At the end, all decision results must generate control frames to be sent out.

[0023] Instruction Execution Slot The first half of this time slot is used by the base station to transmit a sequence of control commands carrying absolute execution timestamps to each device; the second half is executed locally by each device based on the timestamp. The device in the instruction execution slot Upon receiving an instruction, the device does not execute it immediately. Instead, it stores the instruction in a local buffer and parses the absolute execution timestamp. The hardware timer on the device is configured to generate an interrupt at a future time indicated by the absolute execution timestamp and outputs the instruction to the servo driver.

[0024] S105: Synchronization Maintenance and Drift Prevention of Time Slots To address potential clock drift during prolonged operation, at the start of each global control cycle, the base station broadcasts a cycle start pulse. Upon receiving the pulse, the device aligns its local cycle counter with the pulse; simultaneously, the device updates its status upload slot. It carries the actual execution timestamp of the previous cycle locally, which is used by the edge server for closed-loop calibration.

[0025] S2: During the state upload time slot, each port device sends an uplink detection reference signal to the base station. The base station extracts the channel state information of the detection reference signal and sends it to the edge computing server. The edge computing server calculates the relative position, relative speed and relative acceleration of the port device based on the channel state information to obtain control sensing information. This embodiment, within the framework of the global synchronization time base and periodic time slots established in S1, utilizes the state upload slots in each global control cycle. This enables physical layer sensing data interaction between port equipment and base stations, and calculates the relative motion state of the equipment for subsequent real-time control.

[0026] S201: Uplink detection reference signal Configuration and sending During the system initialization phase, the base station central control unit configures a dedicated [device / system] for each port device (quay crane, yard crane, AGV, etc.). Send parameters, including: Sequence: using The sequence has good autocorrelation and low cross-correlation, and supports simultaneous transmission by multiple devices with minimal mutual interference. Sending period: Set to match the global control period Same, that is, each Periodic Within the slot, each device sends an uplink probe reference signal once. ; Frequency domain position: Different cyclic shifts are allocated according to the device ID to achieve code division multiplexing. When the number of port devices does not exceed 32, each device can occupy a different cyclic shift. Temporal location: in The slot is further divided into several micro-time slots, and each device sends an uplink probe reference signal at the start of a specified micro-time slot. According to the above configuration, the 5G modules of the port equipment are in each Periodic Within the slot, automatically generated and sent. The transmission power is adaptively adjusted according to the distance between the device and the base station.

[0027] S202: Channel state information on the base station side extract The receiver of the 5G base station is in The slot continuously monitors the uplink channel, and when it receives data from various port devices... Then, the physical layer of the base station performs the following operations: Channel estimation: utilizing locally stored The reference sequence and the received signal are processed to obtain the channel frequency response on each subcarrier, specifically using the minimum mean square error (MMSE) estimation algorithm. Matrix construction: Organizing the estimation results into Matrix, one set for each device Parameters, including: Amplitude: Signal attenuation of each subcarrier; Phase: The phase offset of each subcarrier; Multipath delay: The power delay spectrum is obtained by inverse Fourier transform, and the delay difference of the main path is extracted; Doppler frequency shift: for consecutive multiple Periodic The phase sequence is differentially calculated to obtain the frequency shift caused by the movement of the equipment; Angle of arrival (AoA): The direction of arrival of a signal is estimated by using the phase difference of a multi-antenna array.

[0028] Timestamp: Each All measurements are appended with a global timestamp at the time of receipt, which is consistent with the global clock reference of S1; The base station will extract the timestamped data mentioned above. The raw data is transmitted in real time to the edge computing server deployed in the port data center via the 5G backhaul network, and the transmission latency should be controlled within [specific parameters]. Within, to ensure in the same Periodic decision scheduling slot The solution is completed within the time limit.

[0029] S203: Edge computing server calculates relative position, relative velocity, and relative acceleration. Edge computing servers in each Periodic Initial stage of the tank, that is, upon receiving all equipment for this cycle. After the data is collected, the following calculation process is executed: S2031: Single device absolute position estimation For a single device, triangulation is performed using the CSI angle of arrival (AoA) of multiple base stations: If at least three 5G base stations are deployed in the port, the angle of arrival (AoA) and multipath latency from the device to each base station are extracted respectively. By solving the hyperbolic equations, the three-dimensional coordinates of the equipment in the port coordinate system are obtained. Coordinates, utilizing the high temporal resolution brought by 5G's large bandwidth, can achieve positioning accuracy down to the centimeter level; At least three 5G macro base stations are deployed in the port operation area in an equilateral triangle coverage layout, with a base station spacing of 150 to 300 meters and an antenna height of more than 5 meters above the top of the containers in the yard; each base station is equipped with an 8-channel uniform linear array, supporting joint estimation of horizontal angle of arrival and pitch angle of arrival, and the global coordinates of the base station are pre-calibrated; The hyperbola system of equations requires further explanation. Let the first equation be... The global coordinates of each base station are The coordinates of the device under test are The signal arrived at the The horizontal angle of arrival of each base station is The pitch angle reached is Then it satisfies the hyperbolic equation system:

[0030] By combining the above hyperbolic equations of at least three base stations, the initial solution of the three-dimensional coordinates of the equipment is solved using the weighted least squares method, with the weights determined by the proportion of the main path energy of the received signal from each base station. Multi-base station data fusion employs a federated filtering architecture for error elimination. Each base station independently outputs a set of location estimation results, and the main filter performs weighted fusion based on the estimated covariance of each base station. The fusion formula is as follows:

[0031] in The number of base stations participating in the integration. For the first Location estimates for each base station, This is to estimate the covariance matrix.

[0032] To address the multipath effects and non-line-of-sight errors caused by container obstruction at the port, a two-step suppression mechanism is employed to achieve centimeter-level positioning: Multipath component elimination: The main path component is extracted based on the power delay spectrum. Multipath components with energy lower than 10dB of the main path are eliminated. Only the arrival angle of the main path is used in the positioning calculation to eliminate the angular deviation caused by the reflection path. Extended Kalman Filter Smoothing: An extended Kalman filter is introduced to recursively smooth the position solution results. Using the AoA positioning results as observations, non-line-of-sight abrupt errors are suppressed through prediction and update iteration. After the above processing, the static positioning accuracy can reach 5cm, and the dynamic positioning accuracy is better than 10cm, meeting the centimeter-level positioning requirements.

[0033] S2032: Calculation of relative positions between equipment For any two port equipment and The edge servers calculate their relative position vectors based on their respective absolute positions: Using two devices to the same base station antenna The phase difference is calculated differentially to eliminate common errors and improve the accuracy of relative position.

[0034] S2033: Relative velocity estimate Using the Doppler frequency shift calculation device to calculate the radial velocity relative to the base station, let... The Doppler frequency shift measured periodically is Then the radial velocity of the device relative to the base station for:

[0035] in Represents the speed of light, The carrier frequency is represented by the relative speed between devices, which is obtained by differentiating the relative positions over two consecutive cycles.

[0036] To suppress noise, a Kalman filter is used to smooth the velocity estimate.

[0037] S2034: Relative Acceleration estimate Acceleration is obtained by differentiating and filtering a continuous multi-period velocity sequence, using a three-point forward difference method. , , Velocity values ​​at three historical moments:

[0038] The calculation formula is based on the acquired historical velocity data. The calculation result is output after being filtered by a first-order low-pass filter and is used for short-period motion trend prediction.

[0039] S204: Generate control sensing information The edge computing server integrates the above calculation results into control perception information, which is stored in the form of a data structure and includes the following fields: Equipment identification ( ), timestamp, absolute position Absolute speed Absolute acceleration The relative position, relative velocity, and relative acceleration of the relative equipment; This control perception information is output to the edge motion planner in S3 as a basis for generating control commands; on the other hand, it is stored in the history cache for use by S6. Used in time series forecasting models.

[0040] S3: Within the decision scheduling time slot, the edge motion planner generates a sequence of control instructions containing absolute execution timestamps based on the control perception information and the control instructions from the remote control center. The absolute execution timestamps are the start times of the instruction execution time slots pointing to the future global control cycle. The edge computing server obtains control and sensing information (including absolute / relative position, speed, acceleration, and timestamp) for each port device in S2, within the same global control cycle. Decision scheduling slot Inside, the edge motion planner generates a sequence of control instructions with absolute execution timestamps based on control perception information and manipulation instructions from the remote control center, which are intended for future execution. S301: Input Acquisition for Edge Motion Planner Deploy the edge motion planner on each edge computing server. Periodic The start time of the slot, i.e. After the slot is finished, the planner input is: Current cycle control sensing information: from S2, including status information of all devices: position, speed, acceleration and timestamp; Historical state cache: past Periodic device state sequences are used for motion prediction and trajectory smoothing; Remote control commands: Issued in real-time by the operator at the remote control center via a touchscreen interface and transmitted to the edge motion planner via 5G. Control commands include: Target location, such as the absolute position of the spreader Speed ​​setting, acceleration limit, emergency stop; each remote command carries a timestamp indicating the time it was issued.

[0041] S302: Determine the execution time window for control commands To enable remote real-time control, the edge motion planner does not assign "immediate execution" to instructions; instead, it binds the execution of instructions to instruction execution slots in future global control cycles. Specifically, it includes: Set a forward period number This indicates the period from the current cycle. After the tank is finished, after A complete After the cycle, the generated instruction sequence is executed; Preferably Time, that is, the instruction in the next cycle. Periodic Execution within the slot, from End of Sufficient time is allowed between the start and end of the transmission for downlink transmission and equipment preprocessing.

[0042] Absolute execution timestamp The calculation formula is:

[0043] in For the current global control cycle The serial number, For the first Instruction execution slots per cycle The starting time; The absolute execution timestamp uses a 64-bit unsigned integer encoding, fully aligned with the global synchronization time base. The encoding structure from high to low is as follows: High 32 bits: Second-level UTC timestamp, based on IEEE 1588v2 PTP synchronization benchmark; The middle 20 bits: microsecond-level time offset, with a precision of... ; The lower 12 bits: 5G air interface timing index, containing a 10-bit frame number and a 2-bit timeslot number, used to align with the air interface frame structure; The encoding and parsing of timestamps use the same benchmark across the entire link to ensure that the base station, edge server, and port equipment represent the same moment in a consistent manner.

[0044] For continuous motion instructions that require execution over multiple cycles, a series of instruction sequences can be generated, each corresponding to a specific cycle. Multiple cycles Start time.

[0045] S303: Generate discrete-time control command sequence The edge motion planner translates remote control commands into discrete control quantities aligned with the global control cycle: State prediction: Using current control sensing information and devices, predicting future execution moments. The equipment is in open-loop condition. The prediction model uses a uniform velocity or uniform acceleration model, and the prediction step size is [value missing]. cycle; Error calculation: interpreting remote control commands as target states. This includes the target position and velocity, and calculating the deviation between the predicted state and the target. Control law calculation: Model predictive control is used to determine the control input sequence required for the device to approach the target within a future time window. The sequence is in discrete time form, with a sampling interval of . .

[0046] Generate an instruction sequence data structure, where each control instruction contains the following fields: Equipment ID: Target port equipment; Command types: speed control, position control, torque control, emergency braking, etc. Control parameters: such as desired velocity vector Desired position, desired acceleration limit; Absolute execution timestamp : pointing to a certain future Periodic The start time of the tank; Effective duration: The duration during which the instruction remains in effect. Number of cycles; Priority: 0 (urgent) to 7 (normal), automatically mapped by remote command type. Emergency stop is mapped to 0, safety avoidance is mapped to 1, and normal operation is mapped to 7; redundancy check, CRC is used for receiver verification.

[0047] S304: Precise alignment and collision prevention of timestamps To ensure multiple devices are in the same When instructions are executed concurrently at the start of the slot, the edge motion planner must follow these rules: Only one valid instruction can be executed on the same device at any given time; if there are multiple instructions, they are merged or overwritten according to priority. Instructions from different devices can share the same absolute execution timestamp, allowing multiple devices to execute synchronously.

[0048] The absolute execution timestamp is greater than the current time of the protection interval. include: Remaining Slot duration, next Periodic and Slot duration, maximum downlink transmission delay; In this embodiment, the following settings are provided: This is used to ensure that the data is delivered on time even if retransmissions occur.

[0049] S305: Packing and caching of instruction sequences The generated control command sequence is grouped by device and packaged into a control frame, the header of which includes: Monotonically increasing frame number, current period The timestamp generated within the slot contains the number of devices and the number of device instructions; The edge motion planner temporarily stores the control frames in the edge server's transmission queue and sends them to the S4 control and communication co-schedule for downlink resource allocation.

[0050] S4: The control and communication joint scheduler in the base station allocates physical resource blocks for the control command sequence according to the command priority and the current air interface channel quality, and completes the transmission to the port equipment within the time window of the absolute execution timestamp; In S3, the edge motion planner generates a sequence of control instructions carrying absolute execution timestamps and packages them into control frames. In this embodiment, the decision scheduling slots are within the same global control cycle. Inside, a joint control and communication scheduler is deployed on the base station side to allocate physical layer resources through control frames and ensure that downlink transmission is completed within the time window specified by the absolute execution timestamp.

[0051] S401: Scheduler Input Information Acquisition The joint scheduler in each Periodic Within the slot, control frames to be sent are received from the edge motion planner, and the following real-time information is obtained from the base station MAC layer: Downlink channel quality indicator, rank indicator, and precoding matrix indicator for each device; Distribution of currently available Physical Resource Blocks (PRBs); The buffer status of each channel and the usage of the hybrid automatic retransmission request process.

[0052] S402: Priority-based differentiated scheduling strategy The joint scheduler parses the priority field of each instruction in the control frame. 0 represents the highest urgency and 7 represents normal operation. Different Physical Resource Blocks (PRBs) are allocated to instructions of different priorities: Priority 0 (Emergency Braking / Safe Stop): The scheduler reserves at least 4 consecutive physical resource blocks (PRBs) for the instruction in the next available transmission time interval, so that it can be transmitted successfully even under poor channel conditions. Priority (Safety Avoidance / High Real-Time Control): distribute Each Physical Resource Block (PRB) allows retransmission, but the total transmission delay must be less than the protection interval before the absolute execution timestamp. Priority (Normal operation / status query): The remaining Physical Resource Blocks (PRBs) are allocated according to the conventional scheduling algorithm, with a maximum of two retransmissions allowed, and the process must be completed before the absolute execution timestamp.

[0053] S403: Time Window Constraints and Resource Reservation The joint scheduler uses the absolute execution timestamp of each instruction. and the current moment Calculate the available transmission time window:

[0054] in For the device receiving and processing time, the scheduler ensures that within Downlink transmission is completed within the specified range, and the scheduler employs a time-aware pre-allocation mechanism: For each control frame, the scheduler... Calculate the deadline for completing the transmission. At the deadline Previously, the scheduler reserved the required PRB resources for each device in advance, so that the PRBs would not be occupied even if the channel was idle at the current time. The location and quantity of the reserved PRBs were determined by both priority and channel quality. If the channel quality suddenly deteriorates after reserving resources, the scheduler can dynamically adjust to other better physical resource blocks (PRBs) within the same time window.

[0055] S404: Transmission Execution and Status Feedback The base station MAC layer encapsulates the control frame into a transport block according to the scheduling decision and sends it to the target port equipment through the downlink shared channel; after the equipment successfully receives it, it replies with an ACK at the physical layer; if there is a margin in the time window, the scheduler triggers HARQ retransmission; once it is confirmed that the equipment has received it correctly, the scheduler marks the instruction as "delivered"; If at the deadline If no physical layer ACK response is received before the scheduled time, the scheduler immediately reports the exception to the edge motion planner.

[0056] S405: Multi-device concurrent transmission optimization To conserve Physical Resource Blocks (PRBs), the joint scheduler performs joint optimization of instructions for multiple devices within the same control frame: Devices with similar channel conditions are assigned to adjacent Physical Resource Blocks (PRBs) to reduce resource fragmentation; For high-priority devices, frequency division multiplexing is used to avoid queuing delays; Under the premise of meeting the delay deadline of each control frame, the average physical resource block (PRB) occupancy is calculated per... The period is the decision moment, the current channel state is the input, and the output is a resource allocation vector.

[0057] S5: After receiving the control command sequence, the port equipment executes the commands in the control command sequence according to the start time of the absolute execution timestamp; The control and communication joint scheduler transmits a sequence of control instructions carrying absolute execution timestamps to the target port equipment. After receiving the instructions, the port equipment uses a local synchronization clock and a hardware timer to trigger the execution of the instructions within the absolute execution timestamp, which is used for the synchronous control of multiple equipment actions.

[0058] S501: Command Reception and Integrity Verification Each port equipment is in the command execution slot Internally listens to the physical downlink shared channel. Upon receiving its own control frame, it performs the following operations: CRC check: Verifies the integrity of the transport block. If the check fails, a HARQ retransmission request is triggered. Unpacking and parsing: The control frame is disassembled according to the protocol, and the fields of each control instruction are extracted, including device ID, instruction type, control parameters, absolute execution timestamp, effective duration, priority, and sequence number; Sequence number check: Compare the instruction sequence number with the local cache. If a jump or duplicate is found, a retransmission request is sent to the edge motion planner.

[0059] S502: Hardware Timer Configuration The port equipment uses a 32-bit automatic reload timer, with a clock source of 100MHz local high-frequency clock phase-locked with the global clock. The timestamp mapping process is as follows: Device resolution of absolute execution timestamp To convert the target count value to a local timer: first calculate the global time corresponding to the current timer count value, then calculate the current local clock. and The difference ,Will Convert the timer clock cycles into a number of times and write it directly into the hardware compare-match register; When the timer counter increments to the matching value, the hardware automatically generates an interrupt trigger signal, and the interrupt response delay does not exceed one clock cycle. like If the value is positive and less than the preset maximum look-ahead time, the instruction along with its parameters is stored in the timer execution queue, and the hardware timer is configured. An interrupt is triggered afterward; if If the value is negative, it means the execution time has passed, so the instruction will be executed immediately and a time deviation alarm will be reported. The 5G module on the device side extracts the frame boundary pulse from the air interface synchronization signal block (SSB) and outputs it to the external calibration pin of the hardware timer. The timer integrates a digital phase-locked loop (PLL) to achieve phase locking using the air interface frame boundary pulse as a reference signal. The phase detector compares the phase difference between the local timer period pulse and the air interface frame pulse, and dynamically adjusts the frequency division coefficient of the local clock through the loop filter. In the locked state, the phase deviation between the local timer and the 5G air interface synchronization signal is less than 50ns; When the air interface signal is briefly lost, the timer enters hold mode, relying on the local temperature-controlled crystal oscillator to maintain clock accuracy, and the drift does not exceed 100ns within 1 second.

[0060] S503: Timing Alignment and Conflict Handling of Multiple Instructions A device may have the same absolute execution timestamp Upon receiving multiple instructions, the device must process them according to the following rules: For multiple instructions with the same absolute execution timestamp, a hardware pre-judgment and single-cycle output mechanism is employed to meet microsecond-level response requirements. After the instruction is received and verified, the software layer sorts them by priority in advance and writes the control parameters of the highest priority instruction into the hardware output buffer register in advance; if it is a compatible instruction, the parameters are pre-merged and written into the same register. When a hardware timer interrupt is triggered, the hardware state machine directly outputs the parameters of the cache register to the executor interface, without the CPU's involvement in instruction arbitration. The total delay from interrupt triggering to instruction output does not exceed [a certain value]. ; Only non-periodic exception instructions are handled by CPU interrupts; all regular control instructions are triggered with zero latency through a hardware preloading mechanism.

[0061] S504: Precise Triggering and Execution of Instructions When the hardware timer count reaches When an interrupt signal is generated, the device control unit responds to the interrupt by immediately reading the instruction from the timed execution queue and outputting it to the actuator in the following manner: Speed ​​control command: Vector the desired velocity. Converted into analog signals or bus commands for servo drivers to drive the device motor in an open-loop or closed-loop manner; Position control command: Generates a position ramp curve, which smoothly moves the user to the target position during the command duration; Emergency stop command: Regardless of the current state of the equipment, immediately cut off the power supply and apply mechanical braking, and upload the status feedback.

[0062] After receiving the instruction, the actuator completes the response within one control cycle. The device caches the execution completion status and actual position and speed information, and uploads it to the slot in the next status cycle. Internal data is then sent back to the edge server.

[0063] S505: Execution Feedback and Closed-Loop Monitoring To ensure control reliability, the device generates an execution confirmation message after executing the command, including: Instruction sequence number, absolute timestamp of actual execution The status of the equipment after execution; The message is uploaded to the status slot in the next global control cycle. Inside, along with the sensor data, it is sent to the motion planner at the base station edge for closed-loop correction and anomaly detection; If the edge server is in two consecutive If no confirmation of instruction execution is received within the cycle, the device is determined to be offline or the instruction is lost.

[0064] S6: Based on the channel state information obtained within the global control cycle, predict the motion trend of each port device within the future time window. When it is predicted that the minimum distance between any two port devices is less than the preset safety threshold, the edge motion planner automatically corrects the control instruction sequence that has not yet been executed, generates correction instructions including deceleration, stopping or trajectory fine-tuning, and issues the correction instructions for execution according to S3 to S5.

[0065] The edge computing server obtains historical channel status information for each port device. Data, S3 to S5 are used for the generation and execution of control instructions; this embodiment uses... The time-series information of the data predicts the future movement trend of the device and automatically corrects the sequence of control commands that have not been executed before a collision risk occurs.

[0066] S601: Accumulation and preprocessing of time series data Edge computing servers in each global control cycle After completion, the device status (position, velocity, acceleration) obtained from the periodic calculations will be... Features (Doppler frequency shift, phase change rate) are stored in the historical data cache, and the cache retains past data. Data for each period; Preprocessing includes: denoising pairs The phase sequence is low-pass filtered, time alignment is performed to interpolate data from different devices to a unified time grid, and normalization is performed to transform physical quantities such as position and velocity into the port's global coordinate system.

[0067] S602: Motion Trend Prediction Model The edge computing server runs a lightweight motion prediction model, with past data as input. Equipment status for each cycle and Features, output as the future prediction time window The sequence of equipment trajectories within; The lightweight motion prediction model uses: Kalman filter prediction: based on a uniform or uniformly accelerated model, utilizing... Doppler frequency shift is used as an observation to recursively predict future states; Long Short-Term Memory Network (LSTM): Train a lightweight LSTM model offline, input a sequence of historical trajectories, and output future trajectory points; The lightweight LSTM model needs further explanation. The lightweight LSTM is a single-layer, unidirectional, minimally simplistic structure, with the following specific parameters: Input layer: Input dimension 16, including 16 features such as device 3D position, 3D velocity, 3D acceleration, CSI Doppler frequency shift, CSI principal diameter amplitude, and time index. The input sequence length is 8 historical control cycles. LSTM hidden layer: 32 hidden units, retaining the forget gate, input gate and output gate three-gate structure, and pruning redundant fully connected branches; Output layer: Single-layer fully connected, with 6 output dimensions, corresponding to the predicted 3D position and 3D velocity of the device for the next 4 control cycles.

[0068] Physical model driven: For devices with known dynamic models (such as AGVs), the current control command sequence is used in conjunction with... Feedback is used for closed-loop prediction to correct model errors.

[0069] The model is based on each The period is the step size, and the output device is in the future. Predicted position at the end of each cycle and speed ,in , , ; The latency of the model inference deployed on the DSP hardware acceleration unit of the edge computing server is calculated as follows: The computation cost of a single forward inference operation is approximately 0.05 MOPs, and the execution time on an 8-bit integer DSP with a 1 GHz clock frequency is approximately 0.3 ms. If FPGA hardware pipeline acceleration is used, the inference latency can be reduced to less than 0.15 ms. The superimposed data preprocessing includes normalization and tensor arrangement, which takes about 0.2ms. The total time for a single complete inference is 0.5ms, which is much less than the 2ms decision scheduling time slot. Parallel inference of all port equipment can be completed in a single cycle.

[0070] In the 2ms decision scheduling time slot The predicted collision avoidance process timing is as follows: From 0 to 0.3ms: Complete CSI data preprocessing and device status calculation for this cycle; From the first 0.3 to 0.8 ms: Complete LSTM trajectory prediction inference for all devices; From the beginning of 0.8 to 1.2 ms: Collision risk calculation and correction command generation completed; From the beginning of 1.2 to 2.0 ms: Complete the packaging of control commands and the allocation of air interface resources.

[0071] The total time for the entire process does not exceed 2ms, which fully meets the time limit requirements of the decision scheduling time slot.

[0072] S603: Collision Risk Assessment For each pair of port equipment and Edge servers in each Periodic decision scheduling slot The following assessments will be performed internally: Calculate the minimum distance to predict: Traverse the future Calculate at each time point. and take the minimum value. ; Get proximity time: Find the time reached Predicted collision time ; Comparison with safety threshold: Preset safety threshold The settings are based on the equipment size and braking capacity. If so, it is determined that there is a risk of collision.

[0073] S604: Modification instruction generation Once a collision risk is detected, the edge motion planner immediately initiates proactive correction, modifying any unexecuted control command sequences. The correction strategy is determined based on device priority and risk level. Priority arbitration: Equipment is dynamically assigned priorities during operation. When a collision is predicted, high-priority equipment retains its original instructions, while low-priority equipment executes avoidance instructions. These avoidance instructions include: Deceleration: Reduce speed to a safe level, causing the equipment to stop or delay its arrival at the conflict zone before approaching it; Stop: If deceleration is still insufficient to ensure safety, a speed stop command is generated; Track fine-tuning: For devices with changeable paths (such as AGVs), generate new waypoints.

[0074] S6041: Correction instruction sequence generation: The planner predicts the collision time. Based on the currently unexecuted instruction sequence, determine the time range that needs correction, and generate a new control instruction sequence carrying an absolute execution timestamp for the low-priority device according to S3, for example, from the future. Periodic The slot begins deceleration; the priority of the correction instruction is set to... It needs to be higher than normal operation but lower than emergency stop.

[0075] S605: Issuance and execution of correction instructions The correction instruction sequence is allocated and transmitted through the joint scheduler of S4, and then executed by the S5 device at the specified absolute execution timestamp. To ensure the timeliness of the correction, the scheduler executes the correction instructions at the current time. The distribution will be completed within the specified period.

[0076] S606: Exception Handling and Monitoring If the collision risk can no longer be avoided by corrective commands, the edge motion planner directly generates an emergency stop command (priority 0) and sends it to the relevant port equipment, and sends an alarm to the remote control center; After the correction command is executed, the actual distance change of the device needs to be continuously monitored in subsequent cycles; if the distance continues to shrink, iterative corrections should be made until the risk is eliminated.

[0077] Based on the above method, this application also discloses a 5G-based remote real-time control system for port equipment, referencing... Figure 4 The 5G-based remote real-time control system for port equipment shown includes: Global time synchronization and cycle division module: Deployed in 5G base stations and edge computing servers, it establishes a global synchronization time reference between port base stations and equipment through 5G time-sensitive network, and sets a global control cycle. The global control cycle is divided into three time slots, namely the status upload slot, the decision scheduling slot, and the instruction execution slot. Sensing and State Calculation Module: In the state upload slot of the global control cycle, it receives the uplink detection reference signal sent by the port equipment, extracts the channel state information, and calculates the relative position, relative speed and relative acceleration of each port equipment based on the channel state information to generate control sensing information; Edge motion planning module: Deployed on the edge computing server, within the decision scheduling slot of the global control cycle, it receives the control perception information and control instructions from the remote control center, and generates a sequence of control instructions carrying an absolute execution timestamp, which points to the start time of the instruction execution time slot of the future global control cycle; Control and communication joint scheduling module: Deployed in the 5G base station, within the decision scheduling slot, it allocates physical resource blocks according to the priority of the instructions in the control instruction sequence and the current channel quality, and performs downlink transmission to the port equipment within the time window of the absolute execution timestamp; Device-side timed execution module: Receives the control instruction sequence, uses a local hardware timer to trigger the execution of the instructions in the control instruction sequence based on the start time of the absolute execution timestamp, and feeds back the execution information to the edge computing server; Predictive Collision Avoidance Correction Module: Deployed on an edge computing server, it predicts the movement trend of port equipment within a future time window based on the acquired channel state information. When it predicts that the minimum distance between any two port equipment is less than a preset safety threshold, it automatically corrects the unexecuted control command sequence and generates a correction command.

[0078] Please see Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 5 As shown, the electronic device 1000 includes: at least one processor 1001, at least one communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to realize the connection and communication between the above components, and to ensure the stable and low-latency transmission of time synchronization signals, control commands, sensing data, scheduling information and equipment status feedback data. The user interface 1003 may include a display screen, a touch interaction unit, and a joystick / handle control unit. Optionally, it may also include a standard wired debugging interface and a wireless Bluetooth interaction interface. The user interface 1003 is used to realize human-machine interaction for remote port operators, configuration of global control cycle and time slot parameters, real-time preview of 5G time synchronization status and port equipment operating status, visualization of equipment position / velocity / acceleration calculation results, monitoring of the operating status of edge motion planning and anti-collision prediction algorithms, issuance of remote control commands and real-time display of execution results, port equipment alarm information prompts, and generation and export of operation logs and inspection reports. The network interface 1004 may optionally include a 5G core network interface, a standard industrial Ethernet interface, a port fieldbus interface, an IEEE 1588v2PTP high-precision time synchronization interface, a GNSS signal receiving interface, a port MES / ERP production management system interface, and a wireless private network interface. The network interface 1004 is used to realize bidirectional communication between electronic equipment and 5G base stations, port operation equipment (quay cranes, yard cranes, AGVs, unmanned trucks, etc.), remote control centers, edge computing nodes, port scheduling management systems, and cloud management platforms. It can complete global time synchronization signal transmission, real-time interaction between uplink SRS signals and downlink CSI data, issuance of control command sequences, feedback of port equipment execution status, distributed training and parameter synchronization of motion prediction models, work data archiving, remote scheduling command reception, and safety alarm information reporting. The processor 1001 may include one or more processing cores. The processor 1001 is connected to various functional units within the electronic device through various interfaces and lines. By running or executing instructions, programs, code sets, or instruction sets stored in the memory 1005, and calling data stored in the memory 1005, the processor 1001 executes all the functions and data processing steps of the 5G-based remote real-time control method for port equipment described in this invention. This includes: establishing a global synchronization time base based on a 5G time-sensitive network, setting the global control cycle and dividing the time slots, receiving uplink SRS signals and extracting CSI features, calculating the relative position / relative speed / relative acceleration of port equipment and generating control perception information, edge motion planning and generating control instruction sequences with absolute execution timestamps, joint scheduling of control and communication based on priority and channel quality, closed-loop management of timed execution of port equipment instructions, prediction of equipment motion trends based on CSI time-series data, collision risk assessment, and automatic correction of unexecuted control instruction sequences, etc., for the entire process.

[0079] Optionally, the processor 1001 can be one or more combinations of a central processing unit (CPU), a graphics processing unit (GPU), a field-programmable gate array (FPGA), and a digital signal processor (DSP). Among them, the CPU mainly handles general computing tasks such as operating system, full-process control logic, user interaction program, and data preprocessing; the GPU is used to handle computationally intensive tasks such as parallel computing for multi-device motion trajectory prediction, forward and backward propagation calculation of LSTM lightweight prediction model, and parallel processing of large-scale port equipment status data; the FPGA can be used to implement hardware-level processing of 5G time synchronization signals, real-time parsing of SRS signals and CSI data, precise timing control of global control cycle and time slot, and hardware acceleration execution of IEEE1588v2PTP protocol; the DSP is responsible for specialized computing tasks such as multi-device triangulation, Kalman filter prediction, Doppler frequency shift and equipment speed calculation, and digital signal processing of channel state information.

[0080] The memory 1005 may include random access memory (RAM), which includes a non-transitory computer-readable medium. The memory 1005 can be used to store instructions, programs, code, code sets, or instruction sets, and includes a program storage area and a data storage area. The program storage area can store instructions for implementing an operating system, instructions for implementing at least one function, and instruction programs for implementing all the steps of the above-described method embodiments of the present invention. The storage data area can store various types of data involved in the implementation of the method of this invention, including but not limited to: global synchronization time base configuration parameters, global control cycle and time slot division parameters, port equipment SRS sequence configuration parameters, CSI historical time series data, calculated equipment position / velocity / acceleration control perception information, control commands issued by the remote control center, generated control command sequences with absolute execution timestamps, physical resource block (PRB) scheduling and allocation data, port equipment command execution status feedback data, motion trend prediction model parameters and trajectory prediction results, preset safety thresholds and anti-collision correction command data, port equipment basic information database, operation logs, alarm records and historical operation reports, etc.

[0081] In the electronic device disclosed in this embodiment, the processor 1001 can call the executable program of the 5G-based port equipment remote real-time control method stored in the memory 1005. When the executable program is executed by one or more processors 1001, the electronic device performs all the steps of the 5G-based port equipment remote real-time control method as described in any of the above embodiments.

[0082] An electronic device readable storage medium storing instructions that, when executed by one or more processors, cause the electronic device to perform one or more of the methods described in the above embodiments: It should be noted that the computer-readable storage medium may include, but is not limited to, USB flash drives, portable hard drives, magnetic disks, optical disks, solid-state drives (SSDs), embedded read-only memory, random access memory, and various non-volatile memories, etc., any medium capable of storing program code. When the computer program product runs on hardware carriers such as electronic devices, port edge industrial control computers, 5G base station edge computing nodes, port dispatch center servers, or cloud management platforms, it can fully execute the entire process of global time synchronization, equipment status perception, edge motion planning, joint communication scheduling, instruction timing execution, and predictive collision avoidance correction in the embodiments of the present invention. This enables highly reliable, low-latency, and highly synchronized remote real-time control of port equipment, providing technical support for improving the efficiency of port automation operations and ensuring safety management across all scenarios.

[0083] 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. 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 5G-based remote real-time control method for port equipment, characterized in that, include: S1: Establish a global synchronization time reference between the base station and port equipment through a 5G time-sensitive network, and set a fixed global control cycle. Divide each global control cycle into non-overlapping status upload time slots, decision scheduling time slots, and instruction execution time slots. S2: During the state upload time slot, each port device sends an uplink detection reference signal to the base station. The base station extracts the channel state information of the detection reference signal and sends it to the edge computing server. The edge computing server calculates the relative position, relative speed and relative acceleration of the port device based on the channel state information to obtain control sensing information. S3: Within the decision scheduling time slot, the edge motion planner generates a sequence of control instructions containing absolute execution timestamps based on the control perception information and the control instructions from the remote control center. The absolute execution timestamps are the start times of the instruction execution time slots pointing to the future global control cycle. S4: The control and communication joint scheduler in the base station allocates physical resource blocks for the control command sequence according to the command priority and the current air interface channel quality, and completes the transmission to the port equipment within the time window of the absolute execution timestamp; S5: After receiving the control command sequence, the port equipment executes the commands in the control command sequence according to the start time of the absolute execution timestamp; S6: Based on the channel state information obtained within the global control cycle, predict the motion trend of each port device within the future time window. When it is predicted that the minimum distance between any two port devices is less than the preset safety threshold, the edge motion planner automatically corrects the control instruction sequence that has not yet been executed, generates correction instructions including deceleration, stopping or trajectory fine-tuning, and issues the correction instructions for execution according to S3 to S5.

2. The 5G-based remote real-time control method for port equipment according to claim 1, characterized in that, S1 specifically includes: Establish a global time reference: Deploy 5G time-sensitive network base stations in the port operation area. The base stations have built-in clock sources, which are synchronized with the master clock based on the IEEE 1588v2PTP precision time protocol. The base station periodically sends time synchronization signals to port equipment within its coverage area via 5G broadcast signaling. The port equipment parses the time synchronization signals and adjusts its local clock to ensure that the synchronization error between the port equipment's local clock and the base station's clock does not exceed ±100ns. Global control cycle: Global control cycle The lengths of the three time slots are allocated as follows: 1ms for status upload, 2ms for decision scheduling, and 2ms for instruction execution. The switching of the three time slots is controlled by a global slot counter jointly maintained by the base station and the edge computing server. At the beginning of each global control cycle, the base station broadcasts a periodic pulse, and the port equipment receives the pulse and aligns its local period counter with the pulse.

3. The 5G-based remote real-time control method for port equipment according to claim 1, characterized in that, S2 specifically includes: S201: During the system initialization phase, the base station configures dedicated uplink sounding reference signal (SRS) transmission parameters for each port device. The transmission period is consistent with the global control period, and the frequency domain position is allocated with different cyclic shifts according to the device ID to achieve code division multiplexing. S202: After receiving the uplink sounding reference signal sent by the port equipment, the base station performs channel estimation through the minimum mean square error (MMSE) estimation algorithm, extracts the channel state information (CSI) including signal amplitude, phase, multipath delay, Doppler shift, and angle of arrival, and marks each set of CSI parameters with a timestamp consistent with the global synchronization time reference. S203: The edge computing server performs triangulation based on the base station's CSI angle of arrival, calculates the three-dimensional absolute coordinates of the port equipment in the port coordinate system, and then calculates the relative position between the equipment based on the absolute coordinates; it calculates the radial velocity of the equipment relative to the base station through the Doppler frequency shift of the CSI, and obtains the relative velocity between the equipment by combining the relative position difference of continuous periods; it obtains the relative acceleration between the equipment by calculating the velocity sequence difference of continuous multi-periods. S204: Integrate the equipment identification, timestamp, absolute position, absolute speed, absolute acceleration of the port equipment with the relative position, relative speed, and relative acceleration of the equipment to generate control sensing information.

4. The 5G-based remote real-time control method for port equipment according to claim 1, characterized in that, S3 specifically includes: S301: At the start of the decision scheduling time slot of each global control cycle, the edge motion planner obtains the control perception information of the current cycle, historical state cache data, and time-stamped control commands issued by the remote control center. S302: Set the number of look-ahead cycles N, bind the execution time of the control instruction to the start time of the instruction execution time slot of the global control cycle after the current cycle, and generate an absolute execution timestamp; S303: Based on the control perception information, predict the state of the device at the absolute execution timestamp, predict the deviation between the state and the remote control command, generate a control input sequence through model predictive control, and package it into a control command sequence; S304: The control command sequence includes device ID, command type, control parameters, absolute execution timestamp, and priority. The priority is divided into 0 to 7 levels, with level 0 being the highest emergency priority, corresponding to emergency braking commands, and level 7 being the normal priority, corresponding to normal operation commands.

5. The 5G-based remote real-time control method for port equipment according to claim 1, characterized in that, S4 specifically includes: S401: The joint control and communication scheduler acquires the sequence of control commands to be issued, the channel quality information of the port equipment, and the distribution information of the currently available physical resource blocks (PRBs) within the decision scheduling time slot. S402: Perform differentiated resource allocation based on the priority of the control command sequence: For level 0 emergency commands, reserve at least 4 consecutive physical resource blocks; for level 1 to 3 high real-time control commands, allocate 2 to 4 physical resource blocks; for level 4 to 7 normal commands, allocate the remaining available physical resource blocks. S403: Calculate the transmission deadline based on the absolute execution timestamp of the control command sequence, reserve physical resource blocks for the corresponding port equipment before the deadline, and complete the transmission of the control command sequence within the time window of the absolute execution timestamp; S404: If no ACK signal is received from the port equipment before the transmission deadline, immediately report the transmission anomaly to the edge motion planner.

6. The 5G-based remote real-time control method for port equipment according to claim 1, characterized in that, S5 specifically includes: S501: After receiving the control command sequence, the port equipment first verifies the integrity of the transmission through CRC check. If the check fails, a retransmission request is triggered. If the check succeeds, the absolute execution timestamp, priority and control parameters in the control command sequence are parsed. S502: The port equipment calculates the time difference between the current local clock and the absolute execution timestamp by using a hardware timer locked to the global synchronization time base. If the time difference is positive and less than the preset maximum look-ahead time, the control instruction is stored in the timed execution queue and the hardware timer is configured to trigger an interrupt at the absolute execution timestamp. If the time difference is negative, the instruction is executed immediately and a time deviation alarm is reported. S503: When the hardware timer triggers an interrupt, the port equipment reads control instructions from the timer execution queue and outputs them to the actuator. For multiple instructions with the same absolute execution timestamp, they are executed according to priority. S504: After the instruction is executed, the port equipment generates an instruction sequence number and an actual execution timestamp, and transmits it back within the status upload time slot of the global control cycle.

7. The 5G-based remote real-time control method for port equipment according to claim 1, characterized in that, S6 specifically includes: S601: After each global control cycle, the edge computing server stores the calculated device status data and CSI features into the historical data cache, and performs noise reduction, time alignment and normalization preprocessing on the CSI time series data. S602: Using a lightweight motion prediction model, taking the equipment status and CSI features of historical cycles as input, outputs the trajectory sequence of each port equipment within the future prediction time window. The lightweight motion prediction model includes a Kalman filter model and a Long Short-Term Memory (LSTM) network model. S603: Iterate through the time points within the future prediction time window, calculate the predicted distance between any two port equipment, obtain the minimum predicted distance and the predicted collision time. If the minimum predicted distance is less than the preset safety threshold, it is determined that there is a collision risk. S604: After detecting a collision risk, arbitration is performed based on the equipment operation priority. Correction instructions such as deceleration, stopping, or trajectory fine-tuning are generated for low-priority equipment, and the issuance and execution of correction instructions are completed according to S3 to S5. S605: If the predicted collision risk cannot be avoided through correction instructions, a Level 0 emergency stop instruction will be generated and sent to the corresponding port equipment, and a safety alarm will be sent to the remote control center at the same time.

8. A 5G-based remote real-time control system for port equipment, characterized in that: include: Global time synchronization and cycle division module: Deployed in 5G base stations and edge computing servers, it establishes a global synchronization time reference between port base stations and equipment through 5G time-sensitive network, and sets a global control cycle. The global control cycle is divided into three time slots, namely the status upload slot, the decision scheduling slot, and the instruction execution slot. Sensing and State Calculation Module: In the state upload slot of the global control cycle, it receives the uplink detection reference signal sent by the port equipment, extracts the channel state information, and calculates the relative position, relative speed and relative acceleration of each port equipment based on the channel state information to generate control sensing information; Edge motion planning module: Deployed on the edge computing server, within the decision scheduling slot of the global control cycle, it receives the control perception information and control instructions from the remote control center, and generates a sequence of control instructions carrying an absolute execution timestamp, which points to the start time of the instruction execution time slot of the future global control cycle; Control and communication joint scheduling module: Deployed in the 5G base station, within the decision scheduling slot, it allocates physical resource blocks according to the priority of the instructions in the control instruction sequence and the current channel quality, and performs downlink transmission to the port equipment within the time window of the absolute execution timestamp; Device-side timed execution module: Receives the control instruction sequence, uses a local hardware timer to trigger the execution of the instructions in the control instruction sequence based on the start time of the absolute execution timestamp, and feeds back the execution information to the edge computing server; Predictive Collision Avoidance Correction Module: Deployed on an edge computing server, it predicts the movement trend of port equipment within a future time window based on the acquired channel state information. When it predicts that the minimum distance between any two port equipment is less than a preset safety threshold, it automatically corrects the unexecuted control command sequence and generates a correction command.

9. An electronic device, characterized in that, include: This includes the processor, communication bus, user interface, network interface, and memory. The communication bus is used for communication between the processor, user interface, network interface, and memory. The user interface is used for human-computer interaction, parameter configuration, status monitoring, and command issuance; Network interfaces are used for bidirectional data communication between electronic devices and 5G base stations, port equipment, remote control centers, and edge computing nodes; The memory stores an executable program that, when executed by a processor, causes the electronic device to perform the 5G-based remote real-time control method for port equipment as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method of any one of claims 1 to 7.

Images