Intelligent forklift based on 5g network remote scheduling management system

Abstract

This invention discloses a remote scheduling and management system for intelligent forklifts based on a 5G network, relating to the field of intelligent remote scheduling and control technology for industrial forklifts. The system includes: a scheduling center subsystem, used to receive multi-source heterogeneous data from forklifts in real time and generate a globally optimal scheduling path set, which is broken down into atomic task units that can be executed in parallel; a remote control terminal cluster containing multiple heterogeneous operating terminals, each equipped with a holographic projection interactive interface and a force feedback joystick, receiving atomic task units and panoramic video streams from the forklifts, while simultaneously feeding back control commands and tactile pressure waveforms to the forklifts; an intelligent forklift subsystem, executing commands and sensing the environment; a collaborative interlocking verification gateway, verifying device digital certificates and automatically establishing spatiotemporal trajectory mutual exclusion locks, allocating path access rights; and a self-healing safety takeover unit, monitoring task execution deviations and triggering a safety recovery mechanism. This invention significantly improves the automation level, operational efficiency, and safety of warehousing and logistics.

Description

Technical Field

[0001] This invention relates to the field of intelligent remote scheduling and control technology for industrial forklifts, and in particular to a remote scheduling and management system for intelligent forklifts based on 5G networks. Background Technology

[0002] In the warehousing and logistics sector, forklifts, as core material handling equipment, directly determine the overall operational efficiency through their scheduling and management efficiency and operational safety. Currently, forklift scheduling and management generally face a series of severe challenges. In the scheduling process, over-reliance on the manual experience of management personnel makes it difficult to obtain real-time and comprehensive information on the specific location status, task execution progress, and battery consumption of all forklifts. This results in delays in issuing scheduling instructions and a lack of a basis for global optimization, leading to both idle and unevenly utilized forklifts, and ultimately requiring further improvement in overall equipment utilization.

[0003] In terms of the working environment, warehouse conditions are complex and constantly changing. Traditional scheduling methods struggle to respond dynamically to unexpected situations such as the insertion of temporary tasks, sudden path blockages, or equipment malfunctions. This often leads to task delays and disrupts on-site operations. The mixing of people and vehicles is common in warehouses; however, there is a lack of effective real-time monitoring and intervention methods for forklift movement and speed. Coupled with limited driver visibility and weak safety awareness, the risk of collisions remains high, posing a significant safety hazard.

[0004] From a communication system perspective, existing dispatching systems based on conventional wireless communication have significant limitations. They suffer from narrow transmission bandwidth, high communication latency, and insufficient connection stability, making it difficult to support the real-time acquisition of massive amounts of forklift operation data, high-concurrency processing, and millisecond-level command issuance. This renders the system unable to meet the stringent requirements of real-time information transparency and rapid response for large-scale fleet collaborative operations and refined management. Summary of the Invention

[0005] In view of this, the present invention proposes a remote scheduling and management system for intelligent forklifts based on 5G networks to solve the problems of low scheduling efficiency and poor safety of warehouse logistics forklifts in the existing technology.

[0006] The specific technical solution of this invention is as follows: The intelligent forklift remote dispatch and management system based on a 5G network includes: The scheduling center subsystem, deployed in a cloud server cluster, is used to receive multi-source heterogeneous data from forklifts in real time and generate a globally optimal scheduling path set through a pre-trained spatiotemporal conflict prediction model. This path set is broken down into atomic task units that can be executed in parallel. The remote control terminal cluster contains multiple heterogeneous operation terminals. Each terminal is equipped with a holographic projection interactive interface and a force feedback joystick. It receives atomic task units and forklift panoramic video streams through ultra-low latency slices of the 5G network, and simultaneously feeds back control commands and tactile pressure waveforms to the forklift. The intelligent forklift subsystem is embedded in the forklift control unit, executing commands and sensing the environment; The collaborative interlocking verification gateway is deployed at the network edge node to verify device digital certificates and automatically establish a spatiotemporal trajectory mutual exclusion lock, and allocate path access rights through the blockchain consensus mechanism. Self-healing safety takeover units are distributed throughout the system to monitor task execution deviations and trigger safety recovery mechanisms.

[0007] Specifically, the multi-source heterogeneous data received by the dispatch center subsystem includes the three-dimensional pose information, load status, battery power remaining, environmental point cloud map, and warehouse cargo distribution heat map of the networked intelligent forklift; the three-dimensional pose information covers position coordinates and angle attitude data, the load status collects the load weight of the forks and the center of gravity position in real time, the battery power remaining reflects the energy consumption status, the environmental point cloud map generates a dynamic obstacle topology structure through sensors, and the warehouse cargo distribution heat map identifies areas with dense cargo.

[0008] Specifically, the intelligent forklift subsystem includes an environment adaptive perception unit, which uses solid-state LiDAR and a depth vision camera to construct a dynamic obstacle topology map in real time and identify non-preset moving targets. The solid-state LiDAR provides anti-interference distance information, and the depth vision camera captures texture and depth data. The two are fused to generate a spatial model.

[0009] Specifically, the intelligent forklift subsystem includes a dynamic arbitration unit for control, which dynamically switches control between remote commands and local decisions based on atomic task units issued by the dispatch center. When network communication latency exceeds a preset safety threshold or a sudden obstacle is detected, the dynamic arbitration unit for control automatically takes over control and executes an emergency braking strategy; otherwise, the planned task is executed first.

[0010] Specifically, the intelligent forklift subsystem includes an energy consumption optimization execution unit that dynamically adjusts the hydraulic lifting speed of the forklift based on the load status and path slope, so that the actual energy consumption conforms to the preset energy consumption envelope.

[0011] Specifically, the self-healing safety takeover unit activates a triple safety mechanism when the execution deviation of the atomic task exceeds the fault tolerance threshold: immediately sending a quantum alarm signal to the dispatch center, activating the local reinforcement learning model to generate an emergency collision avoidance path, and broadcasting control commands to freeze signals and suspend the operation of associated equipment until manual authorization is granted to lift the restrictions.

[0012] Specifically, the dispatch center subsystem uses a spatiotemporal alignment algorithm to fuse multi-source heterogeneous data. The spatiotemporal alignment algorithm matches forklift pose, environmental point cloud, and cargo heat map to a unified spatiotemporal coordinate system and inputs them into a pre-trained spatiotemporal conflict prediction model to predict path intersection conflicts and resource contention bottlenecks. The model dynamically generates path sets based on historical data reasoning.

[0013] Specifically, the scheduling center subsystem uses an optimization algorithm to break down the global scheduling path into atomic task units that can be executed in parallel. Each unit represents an indivisible forklift operation instruction, which is transmitted to the terminal unit using 5G network slicing technology.

[0014] Specifically, the heterogeneous operating terminals in the remote control terminal cluster display a panoramic immersive video stream of the forklift through a holographic projection interactive interface. The force feedback joystick encodes the operator's force waveform into a packet loss-resistant data packet and feeds it back via the 5G network, forming a two-way information flow of video downlink and control uplink.

[0015] Specifically, the collaborative interlocking verification gateway establishes a spatiotemporal trajectory mutual exclusion lock when multiple forklifts enter the collaborative operation radius. Based on the forklift task priority and predicted arrival order, it uses a blockchain consensus mechanism to allocate the unique path passage right in the conflict area. Unauthorized forklifts are controlled to pause until unlocked.

[0016] The beneficial effects of this invention are as follows: 1. By fusing multi-source heterogeneous data through a spatiotemporal alignment algorithm, the spatiotemporal conflict prediction model is input to dynamically analyze the forklift status and environment, predict potential conflicts, resource contention and efficiency bottlenecks, and improve the safety and efficiency of warehousing operations.

[0017] 2. Utilize optimization algorithms to generate a globally optimal scheduling path set and decompose it into atomic task units. Use 5G network slicing technology to ensure timely and reliable transmission of instructions, thereby achieving precise scheduling and efficient execution.

[0018] 3. The remote control terminal is equipped with a holographic projection interactive interface and a force feedback joystick. Combined with the 5G network, it enables downlink high-definition panoramic video and uplink precise control commands and tactile feedback, supporting remote transparent operation.

[0019] 4. The collaborative interlocking verification gateway ensures communication security through identity authentication and uses spatiotemporal collaborative interlocking and distributed coordination protocols to allocate right-of-way in conflict areas to prevent multiple forklift collisions.

[0020] 5. The self-healing safety takeover unit monitors the execution deviation of atomic tasks in real time. Upon triggering, it generates an alarm signal and activates the reinforcement learning collision avoidance model to generate an emergency path. At the same time, it broadcasts control commands to freeze the signal, ensuring the safe recovery of the system. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a schematic diagram of the data flow of the intelligent forklift remote dispatch and management system based on a 5G network according to the present invention. Detailed Implementation

[0023] To make the technical problems to be solved, the technical solutions, and the beneficial effects of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.

[0024] This invention proposes a remote dispatch and management system for intelligent forklifts based on a 5G network. The core objective is to leverage the ultra-low latency, high reliability, and high bandwidth of the 5G network to construct an intelligent forklift management system integrating intelligent dispatching, remote precise control, autonomous safety decision-making, multi-device collaborative interlocking, and fault self-healing. This significantly improves the automation level, operational efficiency, and safety of warehousing and logistics. The entire management system consists of a tightly coupled dispatch center subsystem, a remote control terminal cluster, an intelligent forklift subsystem, a collaborative interlocking verification gateway, and a self-healing safety takeover unit. These components interact and transmit commands at high speed via the 5G network. Figure 1 As shown.

[0025] The dispatch center subsystem is the decision-making hub of the entire system. It is physically deployed on a cloud server cluster consisting of multiple high-performance computing nodes, which typically employ a distributed architecture and are equipped with dedicated hardware acceleration cards to meet real-time computing demands. This subsystem continuously receives multi-source heterogeneous data streams from all networked intelligent forklifts in real time; the fusion and processing of these data streams is its primary function. Specifically, the multi-source heterogeneous data includes, but is not limited to: centimeter-level precision three-dimensional pose information generated by forklift navigation systems such as GNSS, IMU, and wheel speedometers, including position X, Y, and Z coordinates, as well as pitch, roll, and yaw angles; load status data measured in real time by strain sensor arrays integrated into the fork root, including weight and center of gravity position estimation; battery percentage, health status, and estimated range information from the battery management system; environmental point cloud map data collected in real time and preprocessed by solid-state LiDAR and depth vision cameras on the forklift, which typically undergoes noise reduction and key feature extraction; and warehouse cargo distribution heatmaps periodically updated by the warehouse management system or analyzed by wide-angle surveillance cameras deployed on the warehouse roof, which use a coordinate grid to identify the cargo stacking density and type preference in different areas.

[0026] Spatiotemporal alignment algorithms map data from different sources, frequencies, and precisions to a unified timestamp and spatial coordinate system. For example, Kalman filtering or more advanced factor graph optimization methods are used to process pose information and environmental point clouds, and heatmaps are combined to perform probabilistic inference of cargo positions. The fused comprehensive situational data is input into a pre-trained spatiotemporal conflict prediction model, such as a spatiotemporal graph convolutional network or a long short-term memory network with attention mechanisms. Its training data comes from historical operation records, simulated scenarios, and physical constraints. This model dynamically analyzes the current position, speed, planned path (if any), load, warehouse environment (including static and dynamic obstacle predictions), cargo distribution indicated by the heatmap, and operational demand points of all forklifts, predicting potential path intersection conflicts, resource contention such as narrow passages and charging station congestion, and operational efficiency bottlenecks within the next few seconds to tens of seconds.

[0027] Based on this prediction, the model dynamically generates a globally optimal or near-optimal set of scheduling paths by combining optimization algorithms such as improved genetic algorithms, tabu search, or reinforcement learning-based strategies. This global optimum is relative to a weighted combination of preset evaluation metrics such as minimum total job completion time, minimum total energy consumption, and minimum conflict risk. The generated path set is not directly issued to the forklift, but is broken down into a set of finer-grained, logically independent, and parallel-executable atomic task units. Each atomic task unit represents the smallest indivisible job instruction, such as "move from coordinate A to coordinate B", "raise the forks to height H at coordinate C", "lower the pallet at coordinate D", or "go to charging station numbered E". The path instruction decomposition algorithm must consider the physical constraints of the forklift itself, such as minimum turning radius, load state, environmental structure, and dependencies between tasks, to ensure that the execution sequence of atomic task units is logically correct and physically feasible.

[0028] The dispatch center utilizes the core network control plane function of the 5G network and network slicing technology to allocate dedicated slice channels with ultra-low latency and ultra-high reliability for the transmission of dispatch instructions, ensuring that atomic task units can be sent to designated remote control terminals or intelligent forklift subsystems in a timely and reliable manner.

[0029] The remote control terminal cluster consists of multiple heterogeneous operating terminals with potentially different physical forms and performance parameters. These terminals include, but are not limited to, fixed workstations, mobile tablets, or augmented reality glasses combined with dedicated handheld controllers. The most critical feature of each terminal is its holographic projection interface and force feedback joystick. The holographic projection interface typically utilizes light field display technology or projection technology based on special transparent screens to project a stereoscopic image of the controlled forklift and its surrounding environment into the space in front of the operator. This image has depth information and can be viewed from all angles, providing an immersive experience far exceeding that of traditional two-dimensional screens. The image data originates from atomic task unit information such as target points and path instructions issued by the dispatch center, as well as panoramic immersive video streams transmitted via ultra-low latency slices from the 5G network, collected and stitched in real time by the forklift's environmental perception unit. This video stream integrates data from multiple fisheye cameras and depth cameras, and undergoes distortion correction, stitching fusion, and low-latency encoding (such as H.265 with Low Latency mode), typically maintaining end-to-end latency within tens of milliseconds.

[0030] The operator sends control commands by manipulating a force feedback joystick while observing a holographic image. The joystick integrates a high-precision servo motor and position sensor, accurately capturing the operator's intentions such as pushing angle, speed, and force, achieving millimeter-level precision. Furthermore, based on feedback from forklift sensors or the dispatch center—such as bumps caused by obstacles, resistance when the forks contact the cargo, and warnings when approaching physical limits—it generates corresponding counterforce or vibration waveforms that are applied to the operator's hands. While the operator sends control commands, such as fine-tuning the fork position or precise steering, the terminal simultaneously transmits the original control vectors (direction, speed, acceleration) captured by the joystick, along with the tactile pressure waveform applied by the operator, to the corresponding smart forklift. This data is then encapsulated into data packets using specific anti-interference and anti-packet loss coding techniques, such as forward error correction coding combined with data packet redundancy, and sent back to the corresponding smart forklift via the same 5G ultra-low latency slice channel.

[0031] This two-way high-fidelity information flow, namely the downlink of high-definition panoramic video and the uplink of precise control commands and tactile feedback signals, is the key to achieving remote transparent operation, requiring the 5G network to provide ultimate network performance guarantees.

[0032] The intelligent forklift subsystem, serving as the execution end, is embedded in the core control unit of each intelligent forklift. It is a highly integrated hardware and software system. This system comprises three core units that work together to ensure the intelligent, safe, and efficient operation of the forklift in complex environments.

[0033] The environment adaptive perception unit is the forklift's perception unit, primarily relying on a combination of solid-state LiDAR and multi-view depth vision cameras. Solid-state LiDAR, such as MEMS or FLASH solutions, provides high-precision, ambient light-resistant distance information, scanning to form a point cloud of the surrounding environment. Depth vision cameras, such as binocular stereo vision or structured light / time-of-flight solutions, provide rich texture information and high-precision close-range depth maps. This data from both types of sensors is spatiotemporally aligned and fused, and through point cloud registration and target detection and tracking algorithms, such as a deep learning-based 3D target detection network, a dynamic obstacle topology map centered on the forklift's own coordinate system is constructed in real time. This map not only includes static environmental structures such as shelves and walls, but more importantly, it can identify and track non-preset moving targets in real time, such as suddenly appearing workers, other mobile devices like AGVs, and fallen objects. This topology map is stored in a graph data structure, where nodes represent obstacles or key feature points, edges represent spatial relationships or reachability probabilities, and the motion state of each target is continuously updated, including position, speed, and trajectory prediction.

[0034] The dynamic arbitration unit for control is the decision-making unit of the forklift. Its core function is to dynamically and in real-time arbitrate between receiving remote control commands and activating the local obstacle avoidance decision tree, based on the atomic task units issued by the dispatch center. The arbitration logic is based on preset multi-level rules and real-time status monitoring. The primary rule is network status monitoring: the system continuously measures the end-to-end latency of communication with the dispatch center and remote control terminals. When the latency exceeds a preset safety threshold (e.g., a threshold set to 100 milliseconds to ensure safe operation), or when excessive network jitter or a surge in packet loss is detected, the arbitration unit immediately determines that the remote control link is unreliable and automatically triggers the control takeover process. Secondly, when the environmental perception unit detects a sudden obstacle, such as a moving object rapidly approaching the forklift's path, and local calculations predict that the obstacle will pose a collision risk within the time window when remote control commands may not take effect in time due to network latency, the arbitration unit will also immediately take over control. Taking over control means that the forklift's local control system will block or ignore currently received or about to receive remote commands, i.e., direct control commands from the dispatch center or operator. Once takeover occurs, the system immediately executes a preset emergency braking strategy. This strategy is not a simple emergency braking, but a sequential operation: First, the optimal deceleration curve is calculated based on the current speed, load, and ground friction coefficient for braking; simultaneously, the local real-time path planner is immediately activated, such as using a dynamic window method or a fast randomized expanding tree algorithm, to attempt to plan a safe local path around sudden obstacles during braking, if space permits and it is safe; if safe avoidance is not possible, full braking is applied to a complete stop, and an alarm is triggered via the vehicle's audible and visual alarm system. In non-emergency situations, the arbitration unit will prioritize executing instructions from the atomic task units of the dispatch center, or safely execute instructions from remote operators that have undergone local risk assessment.

[0035] The energy optimization execution unit focuses on improving operational efficiency and economy. Based on path and slope information provided by the environmental adaptive perception unit (i.e., by combining map information or visual estimation of slope via inertial measurement unit) and real-time load data provided by the load status unit, it dynamically calculates the optimal execution strategy. The goal is to make the forklift's actual energy consumption curve as close as possible to a preset energy consumption envelope, which defines a reference curve for the theoretically lowest energy consumption under specific load and slope conditions. The core optimization target is the hydraulic lifting system. The unit calculates and dynamically adjusts the hydraulic pump's flow and pressure output curves based on the target lifting height, current load weight, and slope information, generating an optimized hydraulic lifting speed curve. For example, it promotes rapid lifting under no-load or light-load conditions, smooth and uniform lifting under heavy load conditions, and adjusts the lifting speed on sloped sections to avoid hydraulic system overload or energy waste. Simultaneously, the unit also suggests or directly controls the forklift's acceleration and constant-speed driving strategies during movement to minimize peak motor power consumption and maximize regenerative braking energy recovery efficiency.

[0036] As a critical component of network infrastructure, the collaborative interlocking verification gateway is typically deployed on 5G network edge computing nodes near warehouse areas to minimize communication latency. This gateway undertakes two core security responsibilities.

[0037] First, there's authentication and communication security. All remote control terminals and smart forklifts attempting to connect to the system must submit their unique digital certificate, issued by the system's certificate authority, to the gateway before establishing a communication session. The gateway performs rigorous verification using a pre-configured root certificate and certificate revocation list. Only after successful verification is the device allowed to access the network and participate in system communication. All subsequent instructions and data transmitted between the dispatch center, terminals, and forklifts are encrypted and protected for integrity using a session key negotiated with the gateway, such as employing the Chinese national cryptographic standard SM series or TLS protocol.

[0038] Secondly, there is spatiotemporal collaborative interlocking. The gateway continuously monitors the real-time 3D pose information of all intelligent forklifts, which is forwarded by the dispatch center or reported directly by the forklifts. When the system detects that two or more forklifts are about to enter a preset collaborative operation radius, which is dynamically adjustable based on the forklift size, speed, and operation type (e.g., set to 5 meters), the collaborative interlocking mechanism is automatically triggered. The gateway establishes a temporary, virtual spatiotemporal trajectory mutual exclusion lock for potential conflict areas where these forklifts are about to overlap, such as a narrow aisle or in front of the same shelf. The essence of this lock is a distributed coordination protocol. The gateway, acting as the coordinator, uses blockchain technology to implement a consensus mechanism. Specifically, the gateway predicts the order in which the forklifts arrive at the conflict point based on their task priorities (e.g., priority for emergency tasks), preset traffic rules (e.g., driving on the right), current speed, and position. Through a lightweight Byzantine fault-tolerant consensus algorithm, it quickly reaches consensus among the relevant forklift nodes, clearly allocating a unique path right of way for the conflict area within a specific time window. Forklifts granted right-of-way can continue along their original or arbitrated path, while those denied right-of-way are instructed by the gateway through the dispatch center to slow down or pause at a safe location outside the conflict zone. This instruction is contained within the arbitrated atomic task unit and remains in effect until the conflict is resolved, right-of-way is released, or it is reassigned. This mechanism ensures that even when remote operators have limited visibility or local obstacle avoidance decisions focus solely on the safety of individual forklifts, collisions will not occur during the operation of multiple forklifts in the physical space.

[0039] The self-healing safety takeover unit is the last automated line of defense for the system in response to severe anomalies. This system-level monitoring function may be distributed across the dispatch center, collaborative gateway, and intelligent forklift subsystem, working collaboratively. Its core trigger condition is the detection of atomic task execution deviations exceeding the fault tolerance threshold.

[0040] Execution deviation here refers to the significant difference between the actual state of the intelligent forklift (position, speed, fork height, orientation, etc.) and the expected state planned by the currently executing atomic task unit. This difference may stem from various reasons: environmental perception errors, such as failing to identify obstacles and causing path deviation; control actuator malfunctions, such as hydraulic system failure leading to inability to lift; severe external interference, such as the forklift being hit by an external force; extreme network latency causing control commands to fail; or serious errors by the remote operator.

[0041] The fault tolerance threshold is a preset boundary value for different task types and status parameters, such as a position deviation exceeding 50 cm, a speed deviation exceeding 20% ​​of the set value, and a fork height deviation exceeding 10 cm. Real-time monitoring is typically based on comparing local sensor data from the forklift with the expected task status issued by the dispatch center. Once the real-time monitoring system determines that the deviation exceeds the threshold, the self-healing safety takeover unit immediately activates. The self-healing safety takeover unit first generates a high-priority alarm signal. This alarm signal is sent to the dispatch center via the 5G network using a dedicated high-priority slice. The alarm signal includes detailed information such as the faulty forklift identifier, deviation type, degree, location, and current sensor snapshot.

[0042] Almost simultaneously, during the alarm transmission, the self-healing safety takeover unit immediately activates the forklift's local reinforcement learning collision avoidance model. This model, typically trained extensively offline in a simulation environment to model various sudden failures and hazardous scenarios, can generate an emergency path in a very short time based on the forklift's real-time sensor data (from the environmental adaptive perception unit), its own state (speed, orientation, load), and the local environmental map. The sole objective of this path is to allow the forklift to escape the current hazardous state or enter a stable safe state in the fastest and safest way, such as braking, pulling over, avoiding obstacles, and entering open space. After the model generates the emergency path, the forklift control system attempts to execute it autonomously. Upon triggering the safety takeover, the unit immediately broadcasts a control command freeze signal to all remote control terminals and smart forklifts directly associated with the triggered alarm forklift, such as those within the collaborative operation radius, interacting with the same goods, and sharing the same path segment, through the collaborative gateway and dispatch center. This signal has the highest priority. Forced freeze means suspending or ignoring any control commands currently being received or about to be received by all associated devices, whether they are atomic tasks from the dispatch center, real-time control commands from remote operators, or coordination signals from other forklifts. The entire system enters a controlled safety freeze state to prevent accidents from escalating due to command conflicts or misoperation. This freeze state will continue until authorized personnel with higher privileges, such as the dispatch center administrator or safety supervisor, verify the alarm information and confirm the situation through other cameras or on-site personnel, and then issue a release command through a dedicated safety channel. After release, the system typically enters a safety recovery procedure, including task reassignment and manual forklift reset.

[0043] The beneficial effects of this invention are as follows: 1. By fusing multi-source heterogeneous data such as forklift pose, load status, environmental point cloud, and cargo distribution through a spatiotemporal alignment algorithm, the data is input into a spatiotemporal conflict prediction model to dynamically analyze the forklift movement status and warehouse environment, predict future path intersection conflicts, resource contention, and operational efficiency bottlenecks, thereby significantly improving the safety and overall efficiency of warehousing operations.

[0044] 2. An optimization algorithm is used to generate a globally optimal scheduling path set, which is then decomposed into atomic task units based on the physical constraints of forklifts. Network slicing technology of the fifth-generation mobile communication network is used to ensure the timeliness and reliability of instruction transmission, thereby achieving precise scheduling and efficient execution of forklift tasks within the warehouse.

[0045] 3. The remote control terminal is equipped with a holographic projection interactive interface and a force feedback joystick. Combined with the fifth-generation mobile communication network, it realizes the downlink of panoramic video and the uplink of control commands and tactile feedback, forming a two-way high-fidelity information flow, thereby supporting operators to perform highly immersive remote transparent operation.

[0046] 4. The collaborative interlocking verification gateway ensures system communication security through an identity authentication mechanism and dynamically allocates access rights in conflict areas using spatiotemporal collaborative interlocking and distributed coordination protocols, effectively preventing the risk of collisions when multiple forklifts are operating in the warehouse.

[0047] 5. The self-healing safety takeover unit monitors the execution deviation of atomic tasks in real time. Upon triggering, it generates an alarm signal and activates the reinforcement learning collision avoidance model to generate an emergency path. At the same time, it broadcasts control commands to freeze signals, enabling the system to quickly and safely recover in abnormal situations.

[0048] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A remote scheduling management system for intelligent forklifts based on a 5G network, characterized in that, include: The scheduling center subsystem, deployed in a cloud server cluster, is used to receive multi-source heterogeneous data from forklifts in real time and generate a globally optimal scheduling path set through a pre-trained spatiotemporal conflict prediction model. This path set is broken down into atomic task units that can be executed in parallel. The remote control terminal cluster contains multiple heterogeneous operation terminals. Each terminal is equipped with a holographic projection interactive interface and a force feedback joystick. It receives atomic task units and forklift panoramic video streams through ultra-low latency slices of the 5G network, and simultaneously feeds back control commands and tactile pressure waveforms to the forklift. The intelligent forklift subsystem is embedded in the forklift control unit, executing commands and sensing the environment; The collaborative interlocking verification gateway is deployed at the network edge node to verify device digital certificates and automatically establish a spatiotemporal trajectory mutual exclusion lock, and allocate path access rights through the blockchain consensus mechanism. Self-healing safety takeover units are distributed throughout the system to monitor task execution deviations and trigger safety recovery mechanisms.

2. The intelligent forklift remote scheduling management system based on a 5G network according to claim 1, characterized in that, The dispatch center subsystem receives multi-source heterogeneous data including the three-dimensional pose information, load status, battery power remaining, environmental point cloud map, and warehouse cargo distribution heat map of the networked intelligent forklift; the three-dimensional pose information covers position coordinates and angle attitude data, the load status collects the load weight and center of gravity position of the forks in real time, the battery power remaining reflects the energy consumption status, the environmental point cloud map generates a dynamic obstacle topology structure through sensors, and the warehouse cargo distribution heat map identifies areas with dense cargo.

3. The intelligent forklift remote scheduling management system based on a 5G network according to claim 1, characterized in that, The intelligent forklift subsystem includes an environment adaptive perception unit, which uses solid-state LiDAR and a depth vision camera to construct a dynamic obstacle topology map in real time and identify non-preset moving targets. The solid-state LiDAR provides anti-interference distance information, and the depth vision camera captures texture and depth data. The two are fused to generate a spatial model.

4. The intelligent forklift remote dispatch and management system based on a 5G network as described in claim 1, characterized in that, The intelligent forklift subsystem includes a dynamic arbitration unit for control, which dynamically switches control between remote commands and local decisions based on atomic task units issued by the scheduling center. When network communication latency exceeds a preset safety threshold or a sudden obstacle is detected, the dynamic arbitration unit for control automatically takes over control and executes an emergency braking strategy; otherwise, the planned task is executed first.

5. The intelligent forklift remote dispatch and management system based on a 5G network as described in claim 1, characterized in that, The intelligent forklift subsystem includes an energy consumption optimization execution unit that dynamically adjusts the hydraulic lifting speed of the forklift based on the load status and path slope, so that the actual energy consumption conforms to the preset energy consumption envelope.

6. The intelligent forklift remote dispatch and management system based on a 5G network as described in claim 1, characterized in that, The self-healing safety takeover unit activates a triple safety mechanism when the execution deviation of the atomic task exceeds the fault tolerance threshold: immediately sends a quantum alarm signal to the scheduling center, activates the local reinforcement learning model to generate an emergency collision avoidance path, and broadcasts control commands to freeze the signal and suspend the operation of associated equipment until manual authorization is granted to lift the restrictions.

7. The intelligent forklift remote dispatch and management system based on a 5G network as described in claim 2, characterized in that, The dispatch center subsystem uses a spatiotemporal alignment algorithm to fuse multi-source heterogeneous data. The spatiotemporal alignment algorithm matches forklift pose, environmental point cloud and cargo heat map to a unified spatiotemporal coordinate system and inputs it into a pre-trained spatiotemporal conflict prediction model to predict path intersection conflicts and resource contention bottlenecks. The model dynamically generates path sets based on historical data reasoning.

8. The intelligent forklift remote dispatch and management system based on a 5G network as described in claim 1, characterized in that, The scheduling center subsystem uses an optimization algorithm to break down the global scheduling path into atomic task units that can be executed in parallel. Each unit represents an indivisible forklift operation instruction, which is transmitted to the terminal unit using 5G network slicing technology.

9. The intelligent forklift remote dispatch and management system based on a 5G network as described in claim 1, characterized in that, The heterogeneous operating terminals in the remote control terminal cluster display a panoramic immersive video stream of the forklift through a holographic projection interactive interface. The force feedback joystick encodes the operator's force waveform into a packet loss-resistant data packet and feeds it back via the 5G network, forming a two-way information stream of video downlink and control uplink.

10. The intelligent forklift remote dispatch and management system based on a 5G network as described in claim 1, characterized in that, The collaborative interlocking verification gateway establishes a spatiotemporal trajectory mutual exclusion lock when multiple forklifts enter the collaborative operation radius. Based on the forklift task priority and predicted arrival order, it uses a blockchain consensus mechanism to allocate the unique path passage right in the conflict area. Unauthorized forklifts are controlled to suspend until unlocked.

Images